commit fbf14517e1ee1dbd72cc6a02eb49ca1ccdb09e27 Author: ModelHub XC Date: Sat May 9 07:46:50 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_v00.07 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..27048c6 --- /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_v00.07 +tags: +- generated_from_trainer +- open-r1 +- trl +- sft +licence: license +--- + +# Model Card for Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.07 + +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_v00.07", 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/7pvi1oij) + + + +This model was trained with SFT. + +### Framework versions + +- TRL: 1.0.0.dev0 +- Transformers: 4.57.6 +- Pytorch: 2.9.0 +- Datasets: 4.8.3 +- 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..bda08f7 --- /dev/null +++ b/all_results.json @@ -0,0 +1,9 @@ +{ + "ewc_loss": 0.0002346038818359375, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.9308701857364808, + "train_runtime": 37091.194, + "train_samples": 145693, + "train_samples_per_second": 11.784, + "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..20e01e0 --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c79144068720a7075fc11a24449afdb023ebbc2e6d12c549334cf21f53eeb46 +size 4976698672 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..816ddc8 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c203b24149f39137e1ca63a24063411d3dc7bf008d4b42607b9574e2ddf9183 +size 4999802720 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..4c5e6f9 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:affbf4eebad4f415dc7cf6ab49fb2770664096ff217e77445408064af5716ed9 +size 4915916176 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..2708eee --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:900e35676d1fdc1efbac19a858e839edd68c8240e237f2258ff8e9f2e5a80a3b +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..bda08f7 --- /dev/null +++ b/train_results.json @@ -0,0 +1,9 @@ +{ + "ewc_loss": 0.0002346038818359375, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.9308701857364808, + "train_runtime": 37091.194, + "train_samples": 145693, + "train_samples_per_second": 11.784, + "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..7692b97 --- /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": 6.514246940612793, + "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": 6.408624172210693, + "learning_rate": 3.6603221083455343e-10, + "loss": 1.4062, + "mean_token_accuracy": 0.628243088722229, + "num_tokens": 58744.0, + "step": 2 + }, + { + "epoch": 0.000329453107840984, + "ewc_loss": 5.488773498207866e-19, + "grad_norm": 7.283684730529785, + "learning_rate": 7.320644216691069e-10, + "loss": 1.4743, + "mean_token_accuracy": 0.6152434945106506, + "num_tokens": 84293.0, + "step": 3 + }, + { + "epoch": 0.0004392708104546453, + "ewc_loss": 2.5912431922403556e-17, + "grad_norm": 7.036529541015625, + "learning_rate": 1.0980966325036603e-09, + "loss": 1.5504, + "mean_token_accuracy": 0.5984386205673218, + "num_tokens": 113639.0, + "step": 4 + }, + { + "epoch": 0.0005490885130683066, + "ewc_loss": 5.93275428784068e-16, + "grad_norm": 7.08705472946167, + "learning_rate": 1.4641288433382137e-09, + "loss": 1.515, + "mean_token_accuracy": 0.6070650815963745, + "num_tokens": 139615.0, + "step": 5 + }, + { + "epoch": 0.000658906215681968, + "ewc_loss": 7.008282842946301e-16, + "grad_norm": 6.856829643249512, + "learning_rate": 1.8301610541727671e-09, + "loss": 1.6125, + "mean_token_accuracy": 0.5813599228858948, + "num_tokens": 168855.0, + "step": 6 + }, + { + "epoch": 0.0007687239182956292, + "ewc_loss": 2.192690473634684e-15, + "grad_norm": 7.646552562713623, + "learning_rate": 2.1961932650073206e-09, + "loss": 1.5227, + "mean_token_accuracy": 0.602959156036377, + "num_tokens": 192650.0, + "step": 7 + }, + { + "epoch": 0.0008785416209092906, + "ewc_loss": 1.0935696792557792e-14, + "grad_norm": 6.324676513671875, + "learning_rate": 2.562225475841874e-09, + "loss": 1.5189, + "mean_token_accuracy": 0.6046813726425171, + "num_tokens": 225812.0, + "step": 8 + }, + { + "epoch": 0.000988359323522952, + "ewc_loss": 1.9206858326015208e-14, + "grad_norm": 8.854625701904297, + "learning_rate": 2.9282576866764274e-09, + "loss": 1.5178, + "mean_token_accuracy": 0.6043404936790466, + "num_tokens": 245223.0, + "step": 9 + }, + { + "epoch": 0.0010981770261366132, + "ewc_loss": 2.3869795029440866e-14, + "grad_norm": 6.881178855895996, + "learning_rate": 3.294289897510981e-09, + "loss": 1.5665, + "mean_token_accuracy": 0.5958327054977417, + "num_tokens": 273619.0, + "step": 10 + }, + { + "epoch": 0.0012079947287502745, + "ewc_loss": 7.37188088351104e-14, + "grad_norm": 7.198832035064697, + "learning_rate": 3.6603221083455343e-09, + "loss": 1.3787, + "mean_token_accuracy": 0.641814112663269, + "num_tokens": 297418.0, + "step": 11 + }, + { + "epoch": 0.001317812431363936, + "ewc_loss": 9.903189379656396e-14, + "grad_norm": 6.053097724914551, + "learning_rate": 4.026354319180088e-09, + "loss": 1.3924, + "mean_token_accuracy": 0.6356325149536133, + "num_tokens": 328966.0, + "step": 12 + }, + { + "epoch": 0.0014276301339775973, + "ewc_loss": 1.3322676295501878e-13, + "grad_norm": 7.581663608551025, + "learning_rate": 4.392386530014641e-09, + "loss": 1.3888, + "mean_token_accuracy": 0.6391720771789551, + "num_tokens": 353003.0, + "step": 13 + }, + { + "epoch": 0.0015374478365912585, + "ewc_loss": 1.6431300764452317e-13, + "grad_norm": 7.223958969116211, + "learning_rate": 4.758418740849194e-09, + "loss": 1.4648, + "mean_token_accuracy": 0.6214761734008789, + "num_tokens": 377957.0, + "step": 14 + }, + { + "epoch": 0.0016472655392049198, + "ewc_loss": 5.968558980384842e-13, + "grad_norm": 8.122052192687988, + "learning_rate": 5.124450951683748e-09, + "loss": 1.4742, + "mean_token_accuracy": 0.614383339881897, + "num_tokens": 398957.0, + "step": 15 + }, + { + "epoch": 0.0017570832418185812, + "ewc_loss": 7.318590178329032e-13, + "grad_norm": 7.1998515129089355, + "learning_rate": 5.490483162518302e-09, + "loss": 1.5626, + "mean_token_accuracy": 0.5938377976417542, + "num_tokens": 424867.0, + "step": 16 + }, + { + "epoch": 0.0018669009444322424, + "ewc_loss": 8.348877145181177e-13, + "grad_norm": 6.99540376663208, + "learning_rate": 5.856515373352855e-09, + "loss": 1.4468, + "mean_token_accuracy": 0.6201783418655396, + "num_tokens": 450753.0, + "step": 17 + }, + { + "epoch": 0.001976718647045904, + "ewc_loss": 1.0089706847793423e-12, + "grad_norm": 6.828135967254639, + "learning_rate": 6.222547584187408e-09, + "loss": 1.4524, + "mean_token_accuracy": 0.6178641319274902, + "num_tokens": 478721.0, + "step": 18 + }, + { + "epoch": 0.002086536349659565, + "ewc_loss": 1.1581846592889633e-12, + "grad_norm": 8.602494239807129, + "learning_rate": 6.588579795021962e-09, + "loss": 1.5917, + "mean_token_accuracy": 0.5922543406486511, + "num_tokens": 499326.0, + "step": 19 + }, + { + "epoch": 0.0021963540522732263, + "ewc_loss": 3.552713678800501e-12, + "grad_norm": 7.63787317276001, + "learning_rate": 6.954612005856515e-09, + "loss": 1.4422, + "mean_token_accuracy": 0.6256654858589172, + "num_tokens": 522720.0, + "step": 20 + }, + { + "epoch": 0.002306171754886888, + "ewc_loss": 5.059064278611913e-12, + "grad_norm": 6.3194580078125, + "learning_rate": 7.3206442166910685e-09, + "loss": 1.4217, + "mean_token_accuracy": 0.6313837170600891, + "num_tokens": 551422.0, + "step": 21 + }, + { + "epoch": 0.002415989457500549, + "ewc_loss": 6.0254023992456496e-12, + "grad_norm": 8.070073127746582, + "learning_rate": 7.686676427525622e-09, + "loss": 1.5622, + "mean_token_accuracy": 0.601646900177002, + "num_tokens": 572912.0, + "step": 22 + }, + { + "epoch": 0.0025258071601142102, + "ewc_loss": 6.480149750132114e-12, + "grad_norm": 6.0002665519714355, + "learning_rate": 8.052708638360176e-09, + "loss": 1.4115, + "mean_token_accuracy": 0.627722978591919, + "num_tokens": 604172.0, + "step": 23 + }, + { + "epoch": 0.002635624862727872, + "ewc_loss": 7.44648787076585e-12, + "grad_norm": 7.643189907073975, + "learning_rate": 8.418740849194728e-09, + "loss": 1.4466, + "mean_token_accuracy": 0.6210081577301025, + "num_tokens": 627461.0, + "step": 24 + }, + { + "epoch": 0.002745442565341533, + "ewc_loss": 8.185452315956354e-12, + "grad_norm": 6.385133266448975, + "learning_rate": 8.784773060029282e-09, + "loss": 1.4485, + "mean_token_accuracy": 0.6269435882568359, + "num_tokens": 657516.0, + "step": 25 + }, + { + "epoch": 0.0028552602679551946, + "ewc_loss": 9.038103598868474e-12, + "grad_norm": 6.294010162353516, + "learning_rate": 9.150805270863836e-09, + "loss": 1.5439, + "mean_token_accuracy": 0.595317006111145, + "num_tokens": 691937.0, + "step": 26 + }, + { + "epoch": 0.0029650779705688557, + "ewc_loss": 1.3301360013429075e-11, + "grad_norm": 8.508988380432129, + "learning_rate": 9.516837481698389e-09, + "loss": 1.4644, + "mean_token_accuracy": 0.6258860230445862, + "num_tokens": 711900.0, + "step": 27 + }, + { + "epoch": 0.003074895673182517, + "ewc_loss": 3.205968823749572e-11, + "grad_norm": 7.260406970977783, + "learning_rate": 9.882869692532943e-09, + "loss": 1.4916, + "mean_token_accuracy": 0.6119690537452698, + "num_tokens": 737085.0, + "step": 28 + }, + { + "epoch": 0.0031847133757961785, + "ewc_loss": 4.001776687800884e-11, + "grad_norm": 8.06863021850586, + "learning_rate": 1.0248901903367497e-08, + "loss": 1.5751, + "mean_token_accuracy": 0.603097677230835, + "num_tokens": 759817.0, + "step": 29 + }, + { + "epoch": 0.0032945310784098397, + "ewc_loss": 4.4565240386873484e-11, + "grad_norm": 7.577227592468262, + "learning_rate": 1.0614934114202049e-08, + "loss": 1.4183, + "mean_token_accuracy": 0.6257381439208984, + "num_tokens": 783353.0, + "step": 30 + }, + { + "epoch": 0.003404348781023501, + "ewc_loss": 4.865796654485166e-11, + "grad_norm": 8.367369651794434, + "learning_rate": 1.0980966325036603e-08, + "loss": 1.4706, + "mean_token_accuracy": 0.6188283562660217, + "num_tokens": 804197.0, + "step": 31 + }, + { + "epoch": 0.0035141664836371624, + "ewc_loss": 5.1159076974727213e-11, + "grad_norm": 7.104915142059326, + "learning_rate": 1.1346998535871157e-08, + "loss": 1.5124, + "mean_token_accuracy": 0.6063626408576965, + "num_tokens": 830796.0, + "step": 32 + }, + { + "epoch": 0.0036239841862508236, + "ewc_loss": 5.4797055781818926e-11, + "grad_norm": 6.585354328155518, + "learning_rate": 1.171303074670571e-08, + "loss": 1.4162, + "mean_token_accuracy": 0.6407464742660522, + "num_tokens": 858092.0, + "step": 33 + }, + { + "epoch": 0.0037338018888644848, + "ewc_loss": 6.139089236967266e-11, + "grad_norm": 7.315120697021484, + "learning_rate": 1.2079062957540264e-08, + "loss": 1.455, + "mean_token_accuracy": 0.6134889125823975, + "num_tokens": 882474.0, + "step": 34 + }, + { + "epoch": 0.0038436195914781464, + "ewc_loss": 6.548361852765083e-11, + "grad_norm": 8.506824493408203, + "learning_rate": 1.2445095168374816e-08, + "loss": 1.4566, + "mean_token_accuracy": 0.6249269247055054, + "num_tokens": 902361.0, + "step": 35 + }, + { + "epoch": 0.003953437294091808, + "ewc_loss": 7.048583938740194e-11, + "grad_norm": 7.3786797523498535, + "learning_rate": 1.281112737920937e-08, + "loss": 1.5077, + "mean_token_accuracy": 0.6065844297409058, + "num_tokens": 927171.0, + "step": 36 + }, + { + "epoch": 0.004063254996705469, + "ewc_loss": 7.412381819449365e-11, + "grad_norm": 7.09689474105835, + "learning_rate": 1.3177159590043924e-08, + "loss": 1.4723, + "mean_token_accuracy": 0.6123377084732056, + "num_tokens": 952048.0, + "step": 37 + }, + { + "epoch": 0.00417307269931913, + "ewc_loss": 1.6916601452976465e-10, + "grad_norm": 6.875712871551514, + "learning_rate": 1.3543191800878477e-08, + "loss": 1.5742, + "mean_token_accuracy": 0.596790075302124, + "num_tokens": 981040.0, + "step": 38 + }, + { + "epoch": 0.004282890401932792, + "ewc_loss": 2.510205376893282e-10, + "grad_norm": 6.9186110496521, + "learning_rate": 1.390922401171303e-08, + "loss": 1.4836, + "mean_token_accuracy": 0.6060539484024048, + "num_tokens": 1008756.0, + "step": 39 + }, + { + "epoch": 0.004392708104546453, + "ewc_loss": 2.9467628337442875e-10, + "grad_norm": 7.0992255210876465, + "learning_rate": 1.4275256222547585e-08, + "loss": 1.4987, + "mean_token_accuracy": 0.6016491651535034, + "num_tokens": 1035140.0, + "step": 40 + }, + { + "epoch": 0.004502525807160114, + "ewc_loss": 3.2014213502407074e-10, + "grad_norm": 7.792208194732666, + "learning_rate": 1.4641288433382137e-08, + "loss": 1.525, + "mean_token_accuracy": 0.6021480560302734, + "num_tokens": 1058792.0, + "step": 41 + }, + { + "epoch": 0.004612343509773776, + "ewc_loss": 3.4015101846307516e-10, + "grad_norm": 5.665067195892334, + "learning_rate": 1.500732064421669e-08, + "loss": 1.4627, + "mean_token_accuracy": 0.6146364212036133, + "num_tokens": 1096383.0, + "step": 42 + }, + { + "epoch": 0.0047221612123874365, + "ewc_loss": 3.583409124985337e-10, + "grad_norm": 6.812162399291992, + "learning_rate": 1.5373352855051243e-08, + "loss": 1.5745, + "mean_token_accuracy": 0.5920597314834595, + "num_tokens": 1125313.0, + "step": 43 + }, + { + "epoch": 0.004831978915001098, + "ewc_loss": 3.7471181713044643e-10, + "grad_norm": 6.3419413566589355, + "learning_rate": 1.57393850658858e-08, + "loss": 1.4455, + "mean_token_accuracy": 0.6222500801086426, + "num_tokens": 1155650.0, + "step": 44 + }, + { + "epoch": 0.00494179661761476, + "ewc_loss": 3.8744474295526743e-10, + "grad_norm": 7.605759143829346, + "learning_rate": 1.610541727672035e-08, + "loss": 1.4682, + "mean_token_accuracy": 0.6192141771316528, + "num_tokens": 1178843.0, + "step": 45 + }, + { + "epoch": 0.0050516143202284204, + "ewc_loss": 4.2382453102618456e-10, + "grad_norm": 6.931090831756592, + "learning_rate": 1.6471449487554904e-08, + "loss": 1.5244, + "mean_token_accuracy": 0.602024257183075, + "num_tokens": 1206271.0, + "step": 46 + }, + { + "epoch": 0.005161432022842082, + "ewc_loss": 4.656612873077393e-10, + "grad_norm": 5.864633560180664, + "learning_rate": 1.6837481698389456e-08, + "loss": 1.471, + "mean_token_accuracy": 0.6161503791809082, + "num_tokens": 1241401.0, + "step": 47 + }, + { + "epoch": 0.005271249725455744, + "ewc_loss": 4.838511813431978e-10, + "grad_norm": 7.2046098709106445, + "learning_rate": 1.7203513909224012e-08, + "loss": 1.4965, + "mean_token_accuracy": 0.6113574504852295, + "num_tokens": 1267455.0, + "step": 48 + }, + { + "epoch": 0.005381067428069404, + "ewc_loss": 5.020410753786564e-10, + "grad_norm": 6.811875343322754, + "learning_rate": 1.7569546120058564e-08, + "loss": 1.3767, + "mean_token_accuracy": 0.6368637084960938, + "num_tokens": 1293955.0, + "step": 49 + }, + { + "epoch": 0.005490885130683066, + "ewc_loss": 5.20230969414115e-10, + "grad_norm": 7.777338981628418, + "learning_rate": 1.7935578330893117e-08, + "loss": 1.4397, + "mean_token_accuracy": 0.6285771131515503, + "num_tokens": 1316419.0, + "step": 50 + }, + { + "epoch": 0.0056007028332967276, + "ewc_loss": 5.420588422566652e-10, + "grad_norm": 7.788139820098877, + "learning_rate": 1.8301610541727673e-08, + "loss": 1.5096, + "mean_token_accuracy": 0.6068319082260132, + "num_tokens": 1338607.0, + "step": 51 + }, + { + "epoch": 0.005710520535910389, + "ewc_loss": 5.748006515204906e-10, + "grad_norm": 6.895117282867432, + "learning_rate": 1.8667642752562225e-08, + "loss": 1.4852, + "mean_token_accuracy": 0.6092790365219116, + "num_tokens": 1366081.0, + "step": 52 + }, + { + "epoch": 0.00582033823852405, + "ewc_loss": 1.04046193882823e-09, + "grad_norm": 7.448916912078857, + "learning_rate": 1.9033674963396777e-08, + "loss": 1.4829, + "mean_token_accuracy": 0.614183783531189, + "num_tokens": 1390090.0, + "step": 53 + }, + { + "epoch": 0.0059301559411377115, + "ewc_loss": 1.6007106751203537e-09, + "grad_norm": 7.5142693519592285, + "learning_rate": 1.9399707174231333e-08, + "loss": 1.4921, + "mean_token_accuracy": 0.6113466024398804, + "num_tokens": 1413964.0, + "step": 54 + }, + { + "epoch": 0.006039973643751373, + "ewc_loss": 1.862645149230957e-09, + "grad_norm": 6.622805118560791, + "learning_rate": 1.9765739385065885e-08, + "loss": 1.5148, + "mean_token_accuracy": 0.6088438630104065, + "num_tokens": 1442849.0, + "step": 55 + }, + { + "epoch": 0.006149791346365034, + "ewc_loss": 2.0372681319713593e-09, + "grad_norm": 6.224918365478516, + "learning_rate": 2.0131771595900438e-08, + "loss": 1.5721, + "mean_token_accuracy": 0.5893316268920898, + "num_tokens": 1478493.0, + "step": 56 + }, + { + "epoch": 0.006259609048978695, + "ewc_loss": 2.2118911147117615e-09, + "grad_norm": 8.235153198242188, + "learning_rate": 2.0497803806734994e-08, + "loss": 1.5256, + "mean_token_accuracy": 0.6171998381614685, + "num_tokens": 1498811.0, + "step": 57 + }, + { + "epoch": 0.006369426751592357, + "ewc_loss": 2.3283064365386963e-09, + "grad_norm": 6.824991226196289, + "learning_rate": 2.0863836017569546e-08, + "loss": 1.4464, + "mean_token_accuracy": 0.6247538924217224, + "num_tokens": 1525932.0, + "step": 58 + }, + { + "epoch": 0.006479244454206018, + "ewc_loss": 2.444721758365631e-09, + "grad_norm": 7.397132873535156, + "learning_rate": 2.1229868228404098e-08, + "loss": 1.4909, + "mean_token_accuracy": 0.610171914100647, + "num_tokens": 1550438.0, + "step": 59 + }, + { + "epoch": 0.006589062156819679, + "ewc_loss": 2.546585164964199e-09, + "grad_norm": 8.694859504699707, + "learning_rate": 2.1595900439238654e-08, + "loss": 1.5011, + "mean_token_accuracy": 0.611994743347168, + "num_tokens": 1569961.0, + "step": 60 + }, + { + "epoch": 0.006698879859433341, + "ewc_loss": 2.6338966563344e-09, + "grad_norm": 6.060918807983398, + "learning_rate": 2.1961932650073206e-08, + "loss": 1.4962, + "mean_token_accuracy": 0.6061917543411255, + "num_tokens": 1604631.0, + "step": 61 + }, + { + "epoch": 0.006808697562047002, + "ewc_loss": 2.735760062932968e-09, + "grad_norm": 7.65239953994751, + "learning_rate": 2.232796486090776e-08, + "loss": 1.4847, + "mean_token_accuracy": 0.613245964050293, + "num_tokens": 1628098.0, + "step": 62 + }, + { + "epoch": 0.006918515264660663, + "ewc_loss": 2.8230715543031693e-09, + "grad_norm": 6.640145778656006, + "learning_rate": 2.2693997071742314e-08, + "loss": 1.4797, + "mean_token_accuracy": 0.6207623481750488, + "num_tokens": 1656281.0, + "step": 63 + }, + { + "epoch": 0.007028332967274325, + "ewc_loss": 2.9103830456733704e-09, + "grad_norm": 7.437846660614014, + "learning_rate": 2.3060029282576867e-08, + "loss": 1.5129, + "mean_token_accuracy": 0.6158215999603271, + "num_tokens": 1680754.0, + "step": 64 + }, + { + "epoch": 0.007138150669887986, + "ewc_loss": 3.1868694350123405e-09, + "grad_norm": 7.651536464691162, + "learning_rate": 2.342606149341142e-08, + "loss": 1.5122, + "mean_token_accuracy": 0.6087210178375244, + "num_tokens": 1704446.0, + "step": 65 + }, + { + "epoch": 0.007247968372501647, + "ewc_loss": 3.4051481634378433e-09, + "grad_norm": 6.914401054382324, + "learning_rate": 2.3792093704245972e-08, + "loss": 1.297, + "mean_token_accuracy": 0.6489487290382385, + "num_tokens": 1728711.0, + "step": 66 + }, + { + "epoch": 0.007357786075115309, + "ewc_loss": 3.5797711461782455e-09, + "grad_norm": 7.952259063720703, + "learning_rate": 2.4158125915080527e-08, + "loss": 1.4212, + "mean_token_accuracy": 0.6304378509521484, + "num_tokens": 1750331.0, + "step": 67 + }, + { + "epoch": 0.0074676037777289695, + "ewc_loss": 3.7107383832335472e-09, + "grad_norm": 6.344559192657471, + "learning_rate": 2.452415812591508e-08, + "loss": 1.4368, + "mean_token_accuracy": 0.6188677549362183, + "num_tokens": 1781737.0, + "step": 68 + }, + { + "epoch": 0.007577421480342631, + "ewc_loss": 3.725290298461914e-09, + "grad_norm": 6.546405792236328, + "learning_rate": 2.4890190336749632e-08, + "loss": 1.4815, + "mean_token_accuracy": 0.6124039888381958, + "num_tokens": 1811021.0, + "step": 69 + }, + { + "epoch": 0.007687239182956293, + "ewc_loss": 3.841705620288849e-09, + "grad_norm": 7.505778789520264, + "learning_rate": 2.5256222547584188e-08, + "loss": 1.5346, + "mean_token_accuracy": 0.5973970890045166, + "num_tokens": 1835564.0, + "step": 70 + }, + { + "epoch": 0.007797056885569954, + "ewc_loss": 3.958120942115784e-09, + "grad_norm": 10.623238563537598, + "learning_rate": 2.562225475841874e-08, + "loss": 1.5252, + "mean_token_accuracy": 0.6136423945426941, + "num_tokens": 1850951.0, + "step": 71 + }, + { + "epoch": 0.007906874588183616, + "ewc_loss": 4.016328603029251e-09, + "grad_norm": 7.096852779388428, + "learning_rate": 2.5988286969253293e-08, + "loss": 1.5371, + "mean_token_accuracy": 0.6032147407531738, + "num_tokens": 1877690.0, + "step": 72 + }, + { + "epoch": 0.008016692290797276, + "ewc_loss": 4.16184775531292e-09, + "grad_norm": 7.0273051261901855, + "learning_rate": 2.6354319180087848e-08, + "loss": 1.5774, + "mean_token_accuracy": 0.5918317437171936, + "num_tokens": 1904455.0, + "step": 73 + }, + { + "epoch": 0.008126509993410937, + "ewc_loss": 6.111804395914078e-09, + "grad_norm": 7.837868690490723, + "learning_rate": 2.67203513909224e-08, + "loss": 1.6077, + "mean_token_accuracy": 0.5873160362243652, + "num_tokens": 1928049.0, + "step": 74 + }, + { + "epoch": 0.008236327696024599, + "ewc_loss": 8.556526154279709e-09, + "grad_norm": 8.231500625610352, + "learning_rate": 2.7086383601756953e-08, + "loss": 1.4996, + "mean_token_accuracy": 0.6146513819694519, + "num_tokens": 1948422.0, + "step": 75 + }, + { + "epoch": 0.00834614539863826, + "ewc_loss": 1.0360963642597198e-08, + "grad_norm": 7.06660270690918, + "learning_rate": 2.745241581259151e-08, + "loss": 1.4683, + "mean_token_accuracy": 0.6183251142501831, + "num_tokens": 1974329.0, + "step": 76 + }, + { + "epoch": 0.008455963101251922, + "ewc_loss": 1.1408701539039612e-08, + "grad_norm": 6.838471412658691, + "learning_rate": 2.781844802342606e-08, + "loss": 1.4233, + "mean_token_accuracy": 0.630682647228241, + "num_tokens": 2000531.0, + "step": 77 + }, + { + "epoch": 0.008565780803865584, + "ewc_loss": 1.210719347000122e-08, + "grad_norm": 7.144686698913574, + "learning_rate": 2.8184480234260614e-08, + "loss": 1.4782, + "mean_token_accuracy": 0.6130644679069519, + "num_tokens": 2023823.0, + "step": 78 + }, + { + "epoch": 0.008675598506479244, + "ewc_loss": 1.2747477740049362e-08, + "grad_norm": 7.3415117263793945, + "learning_rate": 2.855051244509517e-08, + "loss": 1.5135, + "mean_token_accuracy": 0.6042744517326355, + "num_tokens": 2048734.0, + "step": 79 + }, + { + "epoch": 0.008785416209092905, + "ewc_loss": 1.3271346688270569e-08, + "grad_norm": 6.6862053871154785, + "learning_rate": 2.8916544655929722e-08, + "loss": 1.4141, + "mean_token_accuracy": 0.6267936825752258, + "num_tokens": 2075741.0, + "step": 80 + }, + { + "epoch": 0.008895233911706567, + "ewc_loss": 1.3620592653751373e-08, + "grad_norm": 7.426181793212891, + "learning_rate": 2.9282576866764274e-08, + "loss": 1.5167, + "mean_token_accuracy": 0.6025910377502441, + "num_tokens": 2100247.0, + "step": 81 + }, + { + "epoch": 0.009005051614320228, + "ewc_loss": 1.4202669262886047e-08, + "grad_norm": 7.726316452026367, + "learning_rate": 2.964860907759883e-08, + "loss": 1.5994, + "mean_token_accuracy": 0.6024360060691833, + "num_tokens": 2122939.0, + "step": 82 + }, + { + "epoch": 0.00911486931693389, + "ewc_loss": 1.4842953532934189e-08, + "grad_norm": 8.88399600982666, + "learning_rate": 3.001464128843338e-08, + "loss": 1.4993, + "mean_token_accuracy": 0.6035173535346985, + "num_tokens": 2141424.0, + "step": 83 + }, + { + "epoch": 0.009224687019547552, + "ewc_loss": 1.5133991837501526e-08, + "grad_norm": 6.257712364196777, + "learning_rate": 3.038067349926794e-08, + "loss": 1.398, + "mean_token_accuracy": 0.6270894408226013, + "num_tokens": 2171269.0, + "step": 84 + }, + { + "epoch": 0.009334504722161213, + "ewc_loss": 1.5599653124809265e-08, + "grad_norm": 6.799839019775391, + "learning_rate": 3.074670571010249e-08, + "loss": 1.3789, + "mean_token_accuracy": 0.6452237963676453, + "num_tokens": 2195892.0, + "step": 85 + }, + { + "epoch": 0.009444322424774873, + "ewc_loss": 1.594889909029007e-08, + "grad_norm": 6.112571716308594, + "learning_rate": 3.111273792093704e-08, + "loss": 1.4784, + "mean_token_accuracy": 0.6204155683517456, + "num_tokens": 2227992.0, + "step": 86 + }, + { + "epoch": 0.009554140127388535, + "ewc_loss": 1.618172973394394e-08, + "grad_norm": 6.566022872924805, + "learning_rate": 3.14787701317716e-08, + "loss": 1.5349, + "mean_token_accuracy": 0.6105773448944092, + "num_tokens": 2256995.0, + "step": 87 + }, + { + "epoch": 0.009663957830002196, + "ewc_loss": 1.6530975699424744e-08, + "grad_norm": 6.195683002471924, + "learning_rate": 3.184480234260615e-08, + "loss": 1.4484, + "mean_token_accuracy": 0.6225137710571289, + "num_tokens": 2286492.0, + "step": 88 + }, + { + "epoch": 0.009773775532615858, + "ewc_loss": 1.6880221664905548e-08, + "grad_norm": 7.737649440765381, + "learning_rate": 3.22108345534407e-08, + "loss": 1.5044, + "mean_token_accuracy": 0.607340395450592, + "num_tokens": 2309004.0, + "step": 89 + }, + { + "epoch": 0.00988359323522952, + "ewc_loss": 1.7695128917694092e-08, + "grad_norm": 6.357143402099609, + "learning_rate": 3.257686676427526e-08, + "loss": 1.4034, + "mean_token_accuracy": 0.620206356048584, + "num_tokens": 2336480.0, + "step": 90 + }, + { + "epoch": 0.009993410937843181, + "ewc_loss": 1.909211277961731e-08, + "grad_norm": 6.676301956176758, + "learning_rate": 3.294289897510981e-08, + "loss": 1.4477, + "mean_token_accuracy": 0.6153700947761536, + "num_tokens": 2363526.0, + "step": 91 + }, + { + "epoch": 0.010103228640456841, + "ewc_loss": 2.0139850676059723e-08, + "grad_norm": 6.237678050994873, + "learning_rate": 3.3308931185944364e-08, + "loss": 1.3987, + "mean_token_accuracy": 0.6277164220809937, + "num_tokens": 2393847.0, + "step": 92 + }, + { + "epoch": 0.010213046343070502, + "ewc_loss": 2.10711732506752e-08, + "grad_norm": 6.623597621917725, + "learning_rate": 3.367496339677891e-08, + "loss": 1.4982, + "mean_token_accuracy": 0.6058120727539062, + "num_tokens": 2421639.0, + "step": 93 + }, + { + "epoch": 0.010322864045684164, + "ewc_loss": 2.1653249859809875e-08, + "grad_norm": 7.0082011222839355, + "learning_rate": 3.404099560761347e-08, + "loss": 1.4509, + "mean_token_accuracy": 0.6228134632110596, + "num_tokens": 2445896.0, + "step": 94 + }, + { + "epoch": 0.010432681748297826, + "ewc_loss": 2.223532646894455e-08, + "grad_norm": 6.577927112579346, + "learning_rate": 3.4407027818448024e-08, + "loss": 1.5575, + "mean_token_accuracy": 0.6006249785423279, + "num_tokens": 2475814.0, + "step": 95 + }, + { + "epoch": 0.010542499450911487, + "ewc_loss": 2.2817403078079224e-08, + "grad_norm": 7.2118330001831055, + "learning_rate": 3.477306002928257e-08, + "loss": 1.3842, + "mean_token_accuracy": 0.6392303109169006, + "num_tokens": 2499512.0, + "step": 96 + }, + { + "epoch": 0.010652317153525149, + "ewc_loss": 2.3515895009040833e-08, + "grad_norm": 6.593734264373779, + "learning_rate": 3.513909224011713e-08, + "loss": 1.5601, + "mean_token_accuracy": 0.5943145751953125, + "num_tokens": 2528412.0, + "step": 97 + }, + { + "epoch": 0.010762134856138809, + "ewc_loss": 2.421438694000244e-08, + "grad_norm": 7.153746128082275, + "learning_rate": 3.5505124450951685e-08, + "loss": 1.5047, + "mean_token_accuracy": 0.6099463701248169, + "num_tokens": 2553855.0, + "step": 98 + }, + { + "epoch": 0.01087195255875247, + "ewc_loss": 2.4796463549137115e-08, + "grad_norm": 8.161630630493164, + "learning_rate": 3.5871156661786234e-08, + "loss": 1.5375, + "mean_token_accuracy": 0.6078559160232544, + "num_tokens": 2574996.0, + "step": 99 + }, + { + "epoch": 0.010981770261366132, + "ewc_loss": 2.561137080192566e-08, + "grad_norm": 6.7629075050354, + "learning_rate": 3.623718887262079e-08, + "loss": 1.4908, + "mean_token_accuracy": 0.6125961542129517, + "num_tokens": 2601684.0, + "step": 100 + }, + { + "epoch": 0.011091587963979794, + "ewc_loss": 2.6309862732887268e-08, + "grad_norm": 7.002525329589844, + "learning_rate": 3.6603221083455345e-08, + "loss": 1.5377, + "mean_token_accuracy": 0.6019004583358765, + "num_tokens": 2628612.0, + "step": 101 + }, + { + "epoch": 0.011201405666593455, + "ewc_loss": 2.7124769985675812e-08, + "grad_norm": 6.893762588500977, + "learning_rate": 3.6969253294289894e-08, + "loss": 1.4377, + "mean_token_accuracy": 0.6377737522125244, + "num_tokens": 2652776.0, + "step": 102 + }, + { + "epoch": 0.011311223369207117, + "ewc_loss": 2.9802322387695312e-08, + "grad_norm": 6.9422221183776855, + "learning_rate": 3.733528550512445e-08, + "loss": 1.4061, + "mean_token_accuracy": 0.6310571432113647, + "num_tokens": 2677807.0, + "step": 103 + }, + { + "epoch": 0.011421041071820778, + "ewc_loss": 3.3760443329811096e-08, + "grad_norm": 8.454118728637695, + "learning_rate": 3.7701317715959006e-08, + "loss": 1.554, + "mean_token_accuracy": 0.599973201751709, + "num_tokens": 2697989.0, + "step": 104 + }, + { + "epoch": 0.011530858774434438, + "ewc_loss": 3.934837877750397e-08, + "grad_norm": 7.041243553161621, + "learning_rate": 3.8067349926793555e-08, + "loss": 1.541, + "mean_token_accuracy": 0.6016915440559387, + "num_tokens": 2722341.0, + "step": 105 + }, + { + "epoch": 0.0116406764770481, + "ewc_loss": 4.44706529378891e-08, + "grad_norm": 6.141165733337402, + "learning_rate": 3.843338213762811e-08, + "loss": 1.5183, + "mean_token_accuracy": 0.6025673747062683, + "num_tokens": 2753675.0, + "step": 106 + }, + { + "epoch": 0.011750494179661761, + "ewc_loss": 4.819594323635101e-08, + "grad_norm": 6.791766166687012, + "learning_rate": 3.8799414348462666e-08, + "loss": 1.4657, + "mean_token_accuracy": 0.61378014087677, + "num_tokens": 2778974.0, + "step": 107 + }, + { + "epoch": 0.011860311882275423, + "ewc_loss": 5.029141902923584e-08, + "grad_norm": 6.2310590744018555, + "learning_rate": 3.9165446559297215e-08, + "loss": 1.5175, + "mean_token_accuracy": 0.6029950380325317, + "num_tokens": 2808967.0, + "step": 108 + }, + { + "epoch": 0.011970129584889085, + "ewc_loss": 5.192123353481293e-08, + "grad_norm": 6.640168190002441, + "learning_rate": 3.953147877013177e-08, + "loss": 1.5111, + "mean_token_accuracy": 0.6009050607681274, + "num_tokens": 2835393.0, + "step": 109 + }, + { + "epoch": 0.012079947287502746, + "ewc_loss": 5.3551048040390015e-08, + "grad_norm": 6.033178329467773, + "learning_rate": 3.9897510980966327e-08, + "loss": 1.4727, + "mean_token_accuracy": 0.61161208152771, + "num_tokens": 2863678.0, + "step": 110 + }, + { + "epoch": 0.012189764990116406, + "ewc_loss": 5.541369318962097e-08, + "grad_norm": 6.513078689575195, + "learning_rate": 4.0263543191800876e-08, + "loss": 1.4448, + "mean_token_accuracy": 0.6226648092269897, + "num_tokens": 2890376.0, + "step": 111 + }, + { + "epoch": 0.012299582692730068, + "ewc_loss": 5.681067705154419e-08, + "grad_norm": 6.381383419036865, + "learning_rate": 4.062957540263543e-08, + "loss": 1.4414, + "mean_token_accuracy": 0.6228087544441223, + "num_tokens": 2917128.0, + "step": 112 + }, + { + "epoch": 0.01240940039534373, + "ewc_loss": 5.820766091346741e-08, + "grad_norm": 6.920218467712402, + "learning_rate": 4.099560761346999e-08, + "loss": 1.4762, + "mean_token_accuracy": 0.6097208857536316, + "num_tokens": 2940812.0, + "step": 113 + }, + { + "epoch": 0.01251921809795739, + "ewc_loss": 5.9138983488082886e-08, + "grad_norm": 6.264307022094727, + "learning_rate": 4.1361639824304536e-08, + "loss": 1.5024, + "mean_token_accuracy": 0.6026222705841064, + "num_tokens": 2969881.0, + "step": 114 + }, + { + "epoch": 0.012629035800571052, + "ewc_loss": 6.007030606269836e-08, + "grad_norm": 6.696946620941162, + "learning_rate": 4.172767203513909e-08, + "loss": 1.5014, + "mean_token_accuracy": 0.6081157922744751, + "num_tokens": 2996049.0, + "step": 115 + }, + { + "epoch": 0.012738853503184714, + "ewc_loss": 6.146728992462158e-08, + "grad_norm": 5.753414154052734, + "learning_rate": 4.209370424597365e-08, + "loss": 1.3715, + "mean_token_accuracy": 0.6323637962341309, + "num_tokens": 3025829.0, + "step": 116 + }, + { + "epoch": 0.012848671205798374, + "ewc_loss": 6.472691893577576e-08, + "grad_norm": 6.819464206695557, + "learning_rate": 4.2459736456808197e-08, + "loss": 1.4888, + "mean_token_accuracy": 0.6100147366523743, + "num_tokens": 3051012.0, + "step": 117 + }, + { + "epoch": 0.012958488908412035, + "ewc_loss": 6.705522537231445e-08, + "grad_norm": 6.4530487060546875, + "learning_rate": 4.282576866764275e-08, + "loss": 1.5297, + "mean_token_accuracy": 0.6049206256866455, + "num_tokens": 3077750.0, + "step": 118 + }, + { + "epoch": 0.013068306611025697, + "ewc_loss": 6.845220923423767e-08, + "grad_norm": 6.6645684242248535, + "learning_rate": 4.319180087847731e-08, + "loss": 1.4794, + "mean_token_accuracy": 0.6083287596702576, + "num_tokens": 3103271.0, + "step": 119 + }, + { + "epoch": 0.013178124313639359, + "ewc_loss": 6.984919309616089e-08, + "grad_norm": 7.413395881652832, + "learning_rate": 4.355783308931186e-08, + "loss": 1.5068, + "mean_token_accuracy": 0.601226270198822, + "num_tokens": 3124511.0, + "step": 120 + }, + { + "epoch": 0.01328794201625302, + "ewc_loss": 7.12461769580841e-08, + "grad_norm": 6.48748254776001, + "learning_rate": 4.392386530014641e-08, + "loss": 1.4021, + "mean_token_accuracy": 0.6362745761871338, + "num_tokens": 3150451.0, + "step": 121 + }, + { + "epoch": 0.013397759718866682, + "ewc_loss": 7.264316082000732e-08, + "grad_norm": 6.1252121925354, + "learning_rate": 4.428989751098097e-08, + "loss": 1.5199, + "mean_token_accuracy": 0.6022871136665344, + "num_tokens": 3180151.0, + "step": 122 + }, + { + "epoch": 0.013507577421480343, + "ewc_loss": 7.450580596923828e-08, + "grad_norm": 6.386497974395752, + "learning_rate": 4.465592972181552e-08, + "loss": 1.5134, + "mean_token_accuracy": 0.6065632104873657, + "num_tokens": 3207702.0, + "step": 123 + }, + { + "epoch": 0.013617395124094003, + "ewc_loss": 7.636845111846924e-08, + "grad_norm": 8.089179992675781, + "learning_rate": 4.502196193265007e-08, + "loss": 1.3942, + "mean_token_accuracy": 0.6374097466468811, + "num_tokens": 3226168.0, + "step": 124 + }, + { + "epoch": 0.013727212826707665, + "ewc_loss": 7.916241884231567e-08, + "grad_norm": 6.248714923858643, + "learning_rate": 4.538799414348463e-08, + "loss": 1.4772, + "mean_token_accuracy": 0.6082082986831665, + "num_tokens": 3255714.0, + "step": 125 + }, + { + "epoch": 0.013837030529321327, + "ewc_loss": 8.335337042808533e-08, + "grad_norm": 6.495206832885742, + "learning_rate": 4.575402635431918e-08, + "loss": 1.5157, + "mean_token_accuracy": 0.6078991889953613, + "num_tokens": 3283574.0, + "step": 126 + }, + { + "epoch": 0.013946848231934988, + "ewc_loss": 8.614733815193176e-08, + "grad_norm": 6.543281555175781, + "learning_rate": 4.6120058565153734e-08, + "loss": 1.4593, + "mean_token_accuracy": 0.6191544532775879, + "num_tokens": 3309128.0, + "step": 127 + }, + { + "epoch": 0.01405666593454865, + "ewc_loss": 8.89413058757782e-08, + "grad_norm": 7.420774459838867, + "learning_rate": 4.648609077598828e-08, + "loss": 1.5125, + "mean_token_accuracy": 0.6046103239059448, + "num_tokens": 3331577.0, + "step": 128 + }, + { + "epoch": 0.014166483637162311, + "ewc_loss": 9.220093488693237e-08, + "grad_norm": 6.4982380867004395, + "learning_rate": 4.685212298682284e-08, + "loss": 1.5679, + "mean_token_accuracy": 0.5892118215560913, + "num_tokens": 3358884.0, + "step": 129 + }, + { + "epoch": 0.014276301339775971, + "ewc_loss": 9.499490261077881e-08, + "grad_norm": 8.292394638061523, + "learning_rate": 4.7218155197657394e-08, + "loss": 1.5929, + "mean_token_accuracy": 0.5899166464805603, + "num_tokens": 3378122.0, + "step": 130 + }, + { + "epoch": 0.014386119042389633, + "ewc_loss": 9.872019290924072e-08, + "grad_norm": 6.8363165855407715, + "learning_rate": 4.7584187408491943e-08, + "loss": 1.5476, + "mean_token_accuracy": 0.6019893288612366, + "num_tokens": 3404642.0, + "step": 131 + }, + { + "epoch": 0.014495936745003294, + "ewc_loss": 1.0244548320770264e-07, + "grad_norm": 6.772258281707764, + "learning_rate": 4.79502196193265e-08, + "loss": 1.4416, + "mean_token_accuracy": 0.614977240562439, + "num_tokens": 3429186.0, + "step": 132 + }, + { + "epoch": 0.014605754447616956, + "ewc_loss": 1.0523945093154907e-07, + "grad_norm": 5.9313201904296875, + "learning_rate": 4.8316251830161055e-08, + "loss": 1.4774, + "mean_token_accuracy": 0.6188403367996216, + "num_tokens": 3460810.0, + "step": 133 + }, + { + "epoch": 0.014715572150230618, + "ewc_loss": 1.0849907994270325e-07, + "grad_norm": 5.9141645431518555, + "learning_rate": 4.8682284040995604e-08, + "loss": 1.4439, + "mean_token_accuracy": 0.6179825067520142, + "num_tokens": 3491285.0, + "step": 134 + }, + { + "epoch": 0.01482538985284428, + "ewc_loss": 1.1129304766654968e-07, + "grad_norm": 6.408937454223633, + "learning_rate": 4.904831625183016e-08, + "loss": 1.4383, + "mean_token_accuracy": 0.6206715106964111, + "num_tokens": 3517893.0, + "step": 135 + }, + { + "epoch": 0.014935207555457939, + "ewc_loss": 1.1455267667770386e-07, + "grad_norm": 6.446820259094238, + "learning_rate": 4.9414348462664715e-08, + "loss": 1.4894, + "mean_token_accuracy": 0.6071805953979492, + "num_tokens": 3545614.0, + "step": 136 + }, + { + "epoch": 0.0150450252580716, + "ewc_loss": 1.1641532182693481e-07, + "grad_norm": 6.918100357055664, + "learning_rate": 4.9780380673499264e-08, + "loss": 1.5332, + "mean_token_accuracy": 0.5972411632537842, + "num_tokens": 3571156.0, + "step": 137 + }, + { + "epoch": 0.015154842960685262, + "ewc_loss": 1.1920928955078125e-07, + "grad_norm": 6.942131042480469, + "learning_rate": 5.014641288433382e-08, + "loss": 1.4293, + "mean_token_accuracy": 0.6228659749031067, + "num_tokens": 3595596.0, + "step": 138 + }, + { + "epoch": 0.015264660663298924, + "ewc_loss": 1.210719347000122e-07, + "grad_norm": 6.045258522033691, + "learning_rate": 5.0512445095168376e-08, + "loss": 1.4421, + "mean_token_accuracy": 0.6338889598846436, + "num_tokens": 3625562.0, + "step": 139 + }, + { + "epoch": 0.015374478365912585, + "ewc_loss": 1.2852251529693604e-07, + "grad_norm": 7.067945957183838, + "learning_rate": 5.0878477306002925e-08, + "loss": 1.4961, + "mean_token_accuracy": 0.6198930144309998, + "num_tokens": 3651717.0, + "step": 140 + }, + { + "epoch": 0.015484296068526247, + "ewc_loss": 1.3224780559539795e-07, + "grad_norm": 7.164541244506836, + "learning_rate": 5.124450951683748e-08, + "loss": 1.3793, + "mean_token_accuracy": 0.6367256045341492, + "num_tokens": 3674773.0, + "step": 141 + }, + { + "epoch": 0.015594113771139909, + "ewc_loss": 1.3690441846847534e-07, + "grad_norm": 6.959600448608398, + "learning_rate": 5.1610541727672036e-08, + "loss": 1.5632, + "mean_token_accuracy": 0.5885365605354309, + "num_tokens": 3701717.0, + "step": 142 + }, + { + "epoch": 0.01570393147375357, + "ewc_loss": 1.424923539161682e-07, + "grad_norm": 6.1551079750061035, + "learning_rate": 5.1976573938506585e-08, + "loss": 1.41, + "mean_token_accuracy": 0.6270005702972412, + "num_tokens": 3729901.0, + "step": 143 + }, + { + "epoch": 0.015813749176367232, + "ewc_loss": 1.471489667892456e-07, + "grad_norm": 6.211925029754639, + "learning_rate": 5.234260614934114e-08, + "loss": 1.5111, + "mean_token_accuracy": 0.5952969789505005, + "num_tokens": 3757405.0, + "step": 144 + }, + { + "epoch": 0.01592356687898089, + "ewc_loss": 1.4994293451309204e-07, + "grad_norm": 5.929476737976074, + "learning_rate": 5.2708638360175697e-08, + "loss": 1.3733, + "mean_token_accuracy": 0.636243999004364, + "num_tokens": 3787289.0, + "step": 145 + }, + { + "epoch": 0.01603338458159455, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 6.427931308746338, + "learning_rate": 5.3074670571010246e-08, + "loss": 1.4096, + "mean_token_accuracy": 0.6320145130157471, + "num_tokens": 3811919.0, + "step": 146 + }, + { + "epoch": 0.016143202284208215, + "ewc_loss": 1.6205012798309326e-07, + "grad_norm": 6.326933860778809, + "learning_rate": 5.34407027818448e-08, + "loss": 1.3517, + "mean_token_accuracy": 0.6390037536621094, + "num_tokens": 3835967.0, + "step": 147 + }, + { + "epoch": 0.016253019986821875, + "ewc_loss": 1.685693860054016e-07, + "grad_norm": 6.402769565582275, + "learning_rate": 5.380673499267936e-08, + "loss": 1.397, + "mean_token_accuracy": 0.6268244981765747, + "num_tokens": 3859268.0, + "step": 148 + }, + { + "epoch": 0.016362837689435538, + "ewc_loss": 1.7508864402770996e-07, + "grad_norm": 5.142184734344482, + "learning_rate": 5.4172767203513906e-08, + "loss": 1.4468, + "mean_token_accuracy": 0.6134747266769409, + "num_tokens": 3892417.0, + "step": 149 + }, + { + "epoch": 0.016472655392049198, + "ewc_loss": 1.7881393432617188e-07, + "grad_norm": 7.0064263343811035, + "learning_rate": 5.453879941434846e-08, + "loss": 1.3738, + "mean_token_accuracy": 0.6367802023887634, + "num_tokens": 3913323.0, + "step": 150 + }, + { + "epoch": 0.01658247309466286, + "ewc_loss": 1.8347054719924927e-07, + "grad_norm": 5.052864074707031, + "learning_rate": 5.490483162518302e-08, + "loss": 1.2671, + "mean_token_accuracy": 0.6550483107566833, + "num_tokens": 3942390.0, + "step": 151 + }, + { + "epoch": 0.01669229079727652, + "ewc_loss": 1.8719583749771118e-07, + "grad_norm": 5.8936285972595215, + "learning_rate": 5.527086383601757e-08, + "loss": 1.4521, + "mean_token_accuracy": 0.6140363216400146, + "num_tokens": 3968060.0, + "step": 152 + }, + { + "epoch": 0.01680210849989018, + "ewc_loss": 1.9185245037078857e-07, + "grad_norm": 5.120353698730469, + "learning_rate": 5.563689604685212e-08, + "loss": 1.3287, + "mean_token_accuracy": 0.6448544263839722, + "num_tokens": 3996149.0, + "step": 153 + }, + { + "epoch": 0.016911926202503844, + "ewc_loss": 1.993030309677124e-07, + "grad_norm": 5.614857196807861, + "learning_rate": 5.600292825768668e-08, + "loss": 1.4631, + "mean_token_accuracy": 0.6082680225372314, + "num_tokens": 4022708.0, + "step": 154 + }, + { + "epoch": 0.017021743905117504, + "ewc_loss": 2.0582228899002075e-07, + "grad_norm": 5.953912734985352, + "learning_rate": 5.636896046852123e-08, + "loss": 1.4981, + "mean_token_accuracy": 0.6098594665527344, + "num_tokens": 4048984.0, + "step": 155 + }, + { + "epoch": 0.017131561607731167, + "ewc_loss": 2.1513551473617554e-07, + "grad_norm": 5.366997718811035, + "learning_rate": 5.673499267935578e-08, + "loss": 1.3849, + "mean_token_accuracy": 0.6302242279052734, + "num_tokens": 4075654.0, + "step": 156 + }, + { + "epoch": 0.017241379310344827, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 5.323415756225586, + "learning_rate": 5.710102489019034e-08, + "loss": 1.5405, + "mean_token_accuracy": 0.5872704982757568, + "num_tokens": 4106959.0, + "step": 157 + }, + { + "epoch": 0.017351197012958487, + "ewc_loss": 2.2444874048233032e-07, + "grad_norm": 5.283357620239258, + "learning_rate": 5.746705710102489e-08, + "loss": 1.4297, + "mean_token_accuracy": 0.6245681047439575, + "num_tokens": 4135143.0, + "step": 158 + }, + { + "epoch": 0.01746101471557215, + "ewc_loss": 2.3189932107925415e-07, + "grad_norm": 5.078456401824951, + "learning_rate": 5.7833089311859443e-08, + "loss": 1.3718, + "mean_token_accuracy": 0.6371362805366516, + "num_tokens": 4162641.0, + "step": 159 + }, + { + "epoch": 0.01757083241818581, + "ewc_loss": 2.3469328880310059e-07, + "grad_norm": 5.1318769454956055, + "learning_rate": 5.8199121522694e-08, + "loss": 1.3301, + "mean_token_accuracy": 0.6352059245109558, + "num_tokens": 4189265.0, + "step": 160 + }, + { + "epoch": 0.017680650120799474, + "ewc_loss": 2.384185791015625e-07, + "grad_norm": 6.15118932723999, + "learning_rate": 5.856515373352855e-08, + "loss": 1.4968, + "mean_token_accuracy": 0.6057134866714478, + "num_tokens": 4209765.0, + "step": 161 + }, + { + "epoch": 0.017790467823413134, + "ewc_loss": 2.384185791015625e-07, + "grad_norm": 5.658596038818359, + "learning_rate": 5.8931185944363104e-08, + "loss": 1.3915, + "mean_token_accuracy": 0.6265054941177368, + "num_tokens": 4232444.0, + "step": 162 + }, + { + "epoch": 0.017900285526026797, + "ewc_loss": 2.4028122425079346e-07, + "grad_norm": 4.715243816375732, + "learning_rate": 5.929721815519766e-08, + "loss": 1.3952, + "mean_token_accuracy": 0.6197474002838135, + "num_tokens": 4263314.0, + "step": 163 + }, + { + "epoch": 0.018010103228640457, + "ewc_loss": 2.4400651454925537e-07, + "grad_norm": 4.838335037231445, + "learning_rate": 5.966325036603222e-08, + "loss": 1.4033, + "mean_token_accuracy": 0.6263218522071838, + "num_tokens": 4292268.0, + "step": 164 + }, + { + "epoch": 0.018119920931254117, + "ewc_loss": 2.4959444999694824e-07, + "grad_norm": 5.3221659660339355, + "learning_rate": 6.002928257686676e-08, + "loss": 1.4704, + "mean_token_accuracy": 0.6158422231674194, + "num_tokens": 4317038.0, + "step": 165 + }, + { + "epoch": 0.01822973863386778, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 5.090292453765869, + "learning_rate": 6.039531478770131e-08, + "loss": 1.4737, + "mean_token_accuracy": 0.6050178408622742, + "num_tokens": 4344439.0, + "step": 166 + }, + { + "epoch": 0.01833955633648144, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 5.0622172355651855, + "learning_rate": 6.076134699853588e-08, + "loss": 1.3819, + "mean_token_accuracy": 0.6282846927642822, + "num_tokens": 4371828.0, + "step": 167 + }, + { + "epoch": 0.018449374039095103, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 5.6312336921691895, + "learning_rate": 6.112737920937042e-08, + "loss": 1.3016, + "mean_token_accuracy": 0.6414830684661865, + "num_tokens": 4393611.0, + "step": 168 + }, + { + "epoch": 0.018559191741708763, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 5.127455711364746, + "learning_rate": 6.149341142020497e-08, + "loss": 1.3816, + "mean_token_accuracy": 0.6263933777809143, + "num_tokens": 4419458.0, + "step": 169 + }, + { + "epoch": 0.018669009444322426, + "ewc_loss": 2.6635825634002686e-07, + "grad_norm": 5.007538795471191, + "learning_rate": 6.185944363103954e-08, + "loss": 1.3351, + "mean_token_accuracy": 0.6411285996437073, + "num_tokens": 4444110.0, + "step": 170 + }, + { + "epoch": 0.018778827146936086, + "ewc_loss": 2.7008354663848877e-07, + "grad_norm": 5.744185924530029, + "learning_rate": 6.222547584187409e-08, + "loss": 1.3716, + "mean_token_accuracy": 0.6262579560279846, + "num_tokens": 4465140.0, + "step": 171 + }, + { + "epoch": 0.018888644849549746, + "ewc_loss": 2.7194619178771973e-07, + "grad_norm": 5.206790447235107, + "learning_rate": 6.259150805270863e-08, + "loss": 1.3928, + "mean_token_accuracy": 0.6216269731521606, + "num_tokens": 4489605.0, + "step": 172 + }, + { + "epoch": 0.01899846255216341, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 5.0180768966674805, + "learning_rate": 6.29575402635432e-08, + "loss": 1.3405, + "mean_token_accuracy": 0.6355627179145813, + "num_tokens": 4515095.0, + "step": 173 + }, + { + "epoch": 0.01910828025477707, + "ewc_loss": 2.7939677238464355e-07, + "grad_norm": 4.835413932800293, + "learning_rate": 6.332357247437773e-08, + "loss": 1.3749, + "mean_token_accuracy": 0.6240330338478088, + "num_tokens": 4541160.0, + "step": 174 + }, + { + "epoch": 0.019218097957390733, + "ewc_loss": 2.849847078323364e-07, + "grad_norm": 5.758450984954834, + "learning_rate": 6.36896046852123e-08, + "loss": 1.3629, + "mean_token_accuracy": 0.6259931325912476, + "num_tokens": 4563441.0, + "step": 175 + }, + { + "epoch": 0.019327915660004392, + "ewc_loss": 2.8870999813079834e-07, + "grad_norm": 5.085651397705078, + "learning_rate": 6.405563689604684e-08, + "loss": 1.3103, + "mean_token_accuracy": 0.6388376355171204, + "num_tokens": 4586881.0, + "step": 176 + }, + { + "epoch": 0.019437733362618052, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 4.614548683166504, + "learning_rate": 6.44216691068814e-08, + "loss": 1.3937, + "mean_token_accuracy": 0.621902346611023, + "num_tokens": 4616811.0, + "step": 177 + }, + { + "epoch": 0.019547551065231716, + "ewc_loss": 2.942979335784912e-07, + "grad_norm": 5.049692630767822, + "learning_rate": 6.478770131771596e-08, + "loss": 1.2358, + "mean_token_accuracy": 0.6584005951881409, + "num_tokens": 4639573.0, + "step": 178 + }, + { + "epoch": 0.019657368767845376, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 4.735479831695557, + "learning_rate": 6.515373352855052e-08, + "loss": 1.4227, + "mean_token_accuracy": 0.6181473731994629, + "num_tokens": 4666392.0, + "step": 179 + }, + { + "epoch": 0.01976718647045904, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 4.965061187744141, + "learning_rate": 6.551976573938505e-08, + "loss": 1.3214, + "mean_token_accuracy": 0.6360379457473755, + "num_tokens": 4690288.0, + "step": 180 + }, + { + "epoch": 0.0198770041730727, + "ewc_loss": 3.110617399215698e-07, + "grad_norm": 4.814040184020996, + "learning_rate": 6.588579795021962e-08, + "loss": 1.361, + "mean_token_accuracy": 0.6241999864578247, + "num_tokens": 4716583.0, + "step": 181 + }, + { + "epoch": 0.019986821875686362, + "ewc_loss": 3.203749656677246e-07, + "grad_norm": 4.908924102783203, + "learning_rate": 6.625183016105416e-08, + "loss": 1.2924, + "mean_token_accuracy": 0.6483634114265442, + "num_tokens": 4741315.0, + "step": 182 + }, + { + "epoch": 0.020096639578300022, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 4.49294376373291, + "learning_rate": 6.661786237188873e-08, + "loss": 1.4544, + "mean_token_accuracy": 0.6121571063995361, + "num_tokens": 4771862.0, + "step": 183 + }, + { + "epoch": 0.020206457280913682, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 4.723042011260986, + "learning_rate": 6.698389458272328e-08, + "loss": 1.4035, + "mean_token_accuracy": 0.6225535273551941, + "num_tokens": 4798609.0, + "step": 184 + }, + { + "epoch": 0.020316274983527345, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 4.833038330078125, + "learning_rate": 6.734992679355783e-08, + "loss": 1.3452, + "mean_token_accuracy": 0.6300339698791504, + "num_tokens": 4821516.0, + "step": 185 + }, + { + "epoch": 0.020426092686141005, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 4.942420959472656, + "learning_rate": 6.771595900439237e-08, + "loss": 1.3257, + "mean_token_accuracy": 0.6410174369812012, + "num_tokens": 4845487.0, + "step": 186 + }, + { + "epoch": 0.02053591038875467, + "ewc_loss": 3.4831464290618896e-07, + "grad_norm": 4.980788707733154, + "learning_rate": 6.808199121522694e-08, + "loss": 1.3921, + "mean_token_accuracy": 0.628697395324707, + "num_tokens": 4868981.0, + "step": 187 + }, + { + "epoch": 0.020645728091368328, + "ewc_loss": 3.501772880554199e-07, + "grad_norm": 4.350423336029053, + "learning_rate": 6.844802342606149e-08, + "loss": 1.3655, + "mean_token_accuracy": 0.6237949132919312, + "num_tokens": 4895095.0, + "step": 188 + }, + { + "epoch": 0.02075554579398199, + "ewc_loss": 3.5390257835388184e-07, + "grad_norm": 4.780983924865723, + "learning_rate": 6.881405563689605e-08, + "loss": 1.4304, + "mean_token_accuracy": 0.6133736371994019, + "num_tokens": 4920215.0, + "step": 189 + }, + { + "epoch": 0.02086536349659565, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 4.313560962677002, + "learning_rate": 6.91800878477306e-08, + "loss": 1.3943, + "mean_token_accuracy": 0.6190620064735413, + "num_tokens": 4950113.0, + "step": 190 + }, + { + "epoch": 0.02097518119920931, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 5.038139820098877, + "learning_rate": 6.954612005856515e-08, + "loss": 1.3935, + "mean_token_accuracy": 0.6189162731170654, + "num_tokens": 4970901.0, + "step": 191 + }, + { + "epoch": 0.021084998901822975, + "ewc_loss": 3.9301812648773193e-07, + "grad_norm": 4.396514415740967, + "learning_rate": 6.99121522693997e-08, + "loss": 1.332, + "mean_token_accuracy": 0.637256920337677, + "num_tokens": 4996117.0, + "step": 192 + }, + { + "epoch": 0.021194816604436634, + "ewc_loss": 3.986060619354248e-07, + "grad_norm": 4.5503997802734375, + "learning_rate": 7.027818448023426e-08, + "loss": 1.3675, + "mean_token_accuracy": 0.6263304948806763, + "num_tokens": 5020994.0, + "step": 193 + }, + { + "epoch": 0.021304634307050298, + "ewc_loss": 4.079192876815796e-07, + "grad_norm": 4.447389602661133, + "learning_rate": 7.064421669106881e-08, + "loss": 1.4247, + "mean_token_accuracy": 0.6184637546539307, + "num_tokens": 5053438.0, + "step": 194 + }, + { + "epoch": 0.021414452009663958, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 5.014675617218018, + "learning_rate": 7.101024890190337e-08, + "loss": 1.4085, + "mean_token_accuracy": 0.620928168296814, + "num_tokens": 5077854.0, + "step": 195 + }, + { + "epoch": 0.021524269712277617, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 4.5940423011779785, + "learning_rate": 7.137628111273792e-08, + "loss": 1.4044, + "mean_token_accuracy": 0.6136023998260498, + "num_tokens": 5103447.0, + "step": 196 + }, + { + "epoch": 0.02163408741489128, + "ewc_loss": 4.33996319770813e-07, + "grad_norm": 4.4161505699157715, + "learning_rate": 7.174231332357247e-08, + "loss": 1.4105, + "mean_token_accuracy": 0.6137476563453674, + "num_tokens": 5129474.0, + "step": 197 + }, + { + "epoch": 0.02174390511750494, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 4.921772480010986, + "learning_rate": 7.210834553440702e-08, + "loss": 1.3768, + "mean_token_accuracy": 0.6227342486381531, + "num_tokens": 5154331.0, + "step": 198 + }, + { + "epoch": 0.021853722820118604, + "ewc_loss": 4.507601261138916e-07, + "grad_norm": 4.40837287902832, + "learning_rate": 7.247437774524158e-08, + "loss": 1.3715, + "mean_token_accuracy": 0.6291024684906006, + "num_tokens": 5182997.0, + "step": 199 + }, + { + "epoch": 0.021963540522732264, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 4.128277778625488, + "learning_rate": 7.284040995607613e-08, + "loss": 1.3077, + "mean_token_accuracy": 0.6622510552406311, + "num_tokens": 5209941.0, + "step": 200 + }, + { + "epoch": 0.022073358225345927, + "ewc_loss": 4.675239324569702e-07, + "grad_norm": 4.394569396972656, + "learning_rate": 7.320644216691069e-08, + "loss": 1.2477, + "mean_token_accuracy": 0.6568102836608887, + "num_tokens": 5235749.0, + "step": 201 + }, + { + "epoch": 0.022183175927959587, + "ewc_loss": 4.731118679046631e-07, + "grad_norm": 4.448271751403809, + "learning_rate": 7.357247437774524e-08, + "loss": 1.3227, + "mean_token_accuracy": 0.6426297426223755, + "num_tokens": 5259439.0, + "step": 202 + }, + { + "epoch": 0.022292993630573247, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 4.585951328277588, + "learning_rate": 7.393850658857979e-08, + "loss": 1.3346, + "mean_token_accuracy": 0.6308916807174683, + "num_tokens": 5283571.0, + "step": 203 + }, + { + "epoch": 0.02240281133318691, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 4.93155574798584, + "learning_rate": 7.430453879941434e-08, + "loss": 1.2662, + "mean_token_accuracy": 0.6511291265487671, + "num_tokens": 5306606.0, + "step": 204 + }, + { + "epoch": 0.02251262903580057, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 4.181080341339111, + "learning_rate": 7.46705710102489e-08, + "loss": 1.3645, + "mean_token_accuracy": 0.6393873691558838, + "num_tokens": 5333475.0, + "step": 205 + }, + { + "epoch": 0.022622446738414233, + "ewc_loss": 4.991888999938965e-07, + "grad_norm": 4.27109432220459, + "learning_rate": 7.503660322108345e-08, + "loss": 1.4118, + "mean_token_accuracy": 0.6144008636474609, + "num_tokens": 5360729.0, + "step": 206 + }, + { + "epoch": 0.022732264441027893, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 4.884253025054932, + "learning_rate": 7.540263543191801e-08, + "loss": 1.3012, + "mean_token_accuracy": 0.6473024487495422, + "num_tokens": 5384083.0, + "step": 207 + }, + { + "epoch": 0.022842082143641557, + "ewc_loss": 5.103647708892822e-07, + "grad_norm": 4.477092742919922, + "learning_rate": 7.576866764275256e-08, + "loss": 1.3193, + "mean_token_accuracy": 0.6352516412734985, + "num_tokens": 5407198.0, + "step": 208 + }, + { + "epoch": 0.022951899846255217, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 4.835832595825195, + "learning_rate": 7.613469985358711e-08, + "loss": 1.3238, + "mean_token_accuracy": 0.6422324180603027, + "num_tokens": 5433236.0, + "step": 209 + }, + { + "epoch": 0.023061717548868876, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 4.94079065322876, + "learning_rate": 7.650073206442166e-08, + "loss": 1.3845, + "mean_token_accuracy": 0.6279256343841553, + "num_tokens": 5457306.0, + "step": 210 + }, + { + "epoch": 0.02317153525148254, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 4.619792938232422, + "learning_rate": 7.686676427525622e-08, + "loss": 1.3097, + "mean_token_accuracy": 0.6418183445930481, + "num_tokens": 5477673.0, + "step": 211 + }, + { + "epoch": 0.0232813529540962, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 4.40778923034668, + "learning_rate": 7.723279648609077e-08, + "loss": 1.3771, + "mean_token_accuracy": 0.6280021667480469, + "num_tokens": 5506648.0, + "step": 212 + }, + { + "epoch": 0.023391170656709863, + "ewc_loss": 5.438923835754395e-07, + "grad_norm": 3.9324278831481934, + "learning_rate": 7.759882869692533e-08, + "loss": 1.4426, + "mean_token_accuracy": 0.6063117980957031, + "num_tokens": 5537209.0, + "step": 213 + }, + { + "epoch": 0.023500988359323523, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 5.092946529388428, + "learning_rate": 7.796486090775987e-08, + "loss": 1.3907, + "mean_token_accuracy": 0.630696177482605, + "num_tokens": 5559612.0, + "step": 214 + }, + { + "epoch": 0.023610806061937183, + "ewc_loss": 5.513429641723633e-07, + "grad_norm": 4.683177471160889, + "learning_rate": 7.833089311859443e-08, + "loss": 1.3008, + "mean_token_accuracy": 0.6477435827255249, + "num_tokens": 5581853.0, + "step": 215 + }, + { + "epoch": 0.023720623764550846, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 4.48418664932251, + "learning_rate": 7.869692532942898e-08, + "loss": 1.3128, + "mean_token_accuracy": 0.6421679258346558, + "num_tokens": 5605429.0, + "step": 216 + }, + { + "epoch": 0.023830441467164506, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 5.271427154541016, + "learning_rate": 7.906295754026354e-08, + "loss": 1.2866, + "mean_token_accuracy": 0.6423166990280151, + "num_tokens": 5635630.0, + "step": 217 + }, + { + "epoch": 0.02394025916977817, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 4.781907558441162, + "learning_rate": 7.942898975109809e-08, + "loss": 1.3954, + "mean_token_accuracy": 0.618062436580658, + "num_tokens": 5661702.0, + "step": 218 + }, + { + "epoch": 0.02405007687239183, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 5.574324607849121, + "learning_rate": 7.979502196193265e-08, + "loss": 1.2157, + "mean_token_accuracy": 0.6560882925987244, + "num_tokens": 5682274.0, + "step": 219 + }, + { + "epoch": 0.024159894575005492, + "ewc_loss": 5.885958671569824e-07, + "grad_norm": 3.657257080078125, + "learning_rate": 8.016105417276719e-08, + "loss": 1.313, + "mean_token_accuracy": 0.643455445766449, + "num_tokens": 5715819.0, + "step": 220 + }, + { + "epoch": 0.024269712277619152, + "ewc_loss": 5.923211574554443e-07, + "grad_norm": 4.871540546417236, + "learning_rate": 8.052708638360175e-08, + "loss": 1.3472, + "mean_token_accuracy": 0.6315381526947021, + "num_tokens": 5740534.0, + "step": 221 + }, + { + "epoch": 0.024379529980232812, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 5.6658711433410645, + "learning_rate": 8.08931185944363e-08, + "loss": 1.3921, + "mean_token_accuracy": 0.6217183470726013, + "num_tokens": 5762357.0, + "step": 222 + }, + { + "epoch": 0.024489347682846475, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 4.757065773010254, + "learning_rate": 8.125915080527086e-08, + "loss": 1.2185, + "mean_token_accuracy": 0.6528187990188599, + "num_tokens": 5787211.0, + "step": 223 + }, + { + "epoch": 0.024599165385460135, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 4.172914028167725, + "learning_rate": 8.162518301610541e-08, + "loss": 1.346, + "mean_token_accuracy": 0.6399050951004028, + "num_tokens": 5820989.0, + "step": 224 + }, + { + "epoch": 0.0247089830880738, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 4.628762245178223, + "learning_rate": 8.199121522693997e-08, + "loss": 1.2985, + "mean_token_accuracy": 0.6353253126144409, + "num_tokens": 5844903.0, + "step": 225 + }, + { + "epoch": 0.02481880079068746, + "ewc_loss": 6.183981895446777e-07, + "grad_norm": 5.086416721343994, + "learning_rate": 8.235724743777451e-08, + "loss": 1.2993, + "mean_token_accuracy": 0.644708514213562, + "num_tokens": 5867365.0, + "step": 226 + }, + { + "epoch": 0.024928618493301122, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 4.83430814743042, + "learning_rate": 8.272327964860907e-08, + "loss": 1.3755, + "mean_token_accuracy": 0.621787428855896, + "num_tokens": 5895675.0, + "step": 227 + }, + { + "epoch": 0.02503843619591478, + "ewc_loss": 6.295740604400635e-07, + "grad_norm": 4.69564962387085, + "learning_rate": 8.308931185944362e-08, + "loss": 1.3314, + "mean_token_accuracy": 0.6392855048179626, + "num_tokens": 5920138.0, + "step": 228 + }, + { + "epoch": 0.02514825389852844, + "ewc_loss": 6.295740604400635e-07, + "grad_norm": 5.278421401977539, + "learning_rate": 8.345534407027818e-08, + "loss": 1.3096, + "mean_token_accuracy": 0.6337404847145081, + "num_tokens": 5945763.0, + "step": 229 + }, + { + "epoch": 0.025258071601142105, + "ewc_loss": 6.48200511932373e-07, + "grad_norm": 5.036571979522705, + "learning_rate": 8.382137628111273e-08, + "loss": 1.2266, + "mean_token_accuracy": 0.6616706848144531, + "num_tokens": 5971811.0, + "step": 230 + }, + { + "epoch": 0.025367889303755765, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 4.486859321594238, + "learning_rate": 8.41874084919473e-08, + "loss": 1.3613, + "mean_token_accuracy": 0.632822573184967, + "num_tokens": 5995852.0, + "step": 231 + }, + { + "epoch": 0.025477707006369428, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 4.763795375823975, + "learning_rate": 8.455344070278183e-08, + "loss": 1.2913, + "mean_token_accuracy": 0.6413557529449463, + "num_tokens": 6018357.0, + "step": 232 + }, + { + "epoch": 0.025587524708983088, + "ewc_loss": 6.891787052154541e-07, + "grad_norm": 4.442183971405029, + "learning_rate": 8.491947291361639e-08, + "loss": 1.3396, + "mean_token_accuracy": 0.6350165605545044, + "num_tokens": 6046503.0, + "step": 233 + }, + { + "epoch": 0.025697342411596748, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 5.559413433074951, + "learning_rate": 8.528550512445094e-08, + "loss": 1.272, + "mean_token_accuracy": 0.645202100276947, + "num_tokens": 6067210.0, + "step": 234 + }, + { + "epoch": 0.02580716011421041, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 5.095271110534668, + "learning_rate": 8.56515373352855e-08, + "loss": 1.2492, + "mean_token_accuracy": 0.6495173573493958, + "num_tokens": 6091649.0, + "step": 235 + }, + { + "epoch": 0.02591697781682407, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 5.155811786651611, + "learning_rate": 8.601756954612005e-08, + "loss": 1.3097, + "mean_token_accuracy": 0.6426445245742798, + "num_tokens": 6119052.0, + "step": 236 + }, + { + "epoch": 0.026026795519437734, + "ewc_loss": 6.891787052154541e-07, + "grad_norm": 5.176318168640137, + "learning_rate": 8.638360175695462e-08, + "loss": 1.2519, + "mean_token_accuracy": 0.6572808623313904, + "num_tokens": 6141688.0, + "step": 237 + }, + { + "epoch": 0.026136613222051394, + "ewc_loss": 7.189810276031494e-07, + "grad_norm": 4.686403274536133, + "learning_rate": 8.674963396778915e-08, + "loss": 1.3315, + "mean_token_accuracy": 0.6296212077140808, + "num_tokens": 6168591.0, + "step": 238 + }, + { + "epoch": 0.026246430924665057, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 6.225805759429932, + "learning_rate": 8.711566617862371e-08, + "loss": 1.2039, + "mean_token_accuracy": 0.6593726873397827, + "num_tokens": 6186942.0, + "step": 239 + }, + { + "epoch": 0.026356248627278717, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 4.754128932952881, + "learning_rate": 8.748169838945826e-08, + "loss": 1.3, + "mean_token_accuracy": 0.6412358283996582, + "num_tokens": 6214324.0, + "step": 240 + }, + { + "epoch": 0.026466066329892377, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 4.673898220062256, + "learning_rate": 8.784773060029283e-08, + "loss": 1.3176, + "mean_token_accuracy": 0.6363711357116699, + "num_tokens": 6240342.0, + "step": 241 + }, + { + "epoch": 0.02657588403250604, + "ewc_loss": 7.487833499908447e-07, + "grad_norm": 5.154688835144043, + "learning_rate": 8.821376281112737e-08, + "loss": 1.3687, + "mean_token_accuracy": 0.6260125637054443, + "num_tokens": 6266070.0, + "step": 242 + }, + { + "epoch": 0.0266857017351197, + "ewc_loss": 7.487833499908447e-07, + "grad_norm": 4.7961297035217285, + "learning_rate": 8.857979502196194e-08, + "loss": 1.3202, + "mean_token_accuracy": 0.636433482170105, + "num_tokens": 6294087.0, + "step": 243 + }, + { + "epoch": 0.026795519437733364, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 4.685704708099365, + "learning_rate": 8.894582723279647e-08, + "loss": 1.2852, + "mean_token_accuracy": 0.6391541957855225, + "num_tokens": 6318429.0, + "step": 244 + }, + { + "epoch": 0.026905337140347024, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 4.599842548370361, + "learning_rate": 8.931185944363104e-08, + "loss": 1.3173, + "mean_token_accuracy": 0.6406774520874023, + "num_tokens": 6348101.0, + "step": 245 + }, + { + "epoch": 0.027015154842960687, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 5.2912797927856445, + "learning_rate": 8.967789165446558e-08, + "loss": 1.2921, + "mean_token_accuracy": 0.6476960182189941, + "num_tokens": 6373529.0, + "step": 246 + }, + { + "epoch": 0.027124972545574347, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 4.563386917114258, + "learning_rate": 9.004392386530015e-08, + "loss": 1.3998, + "mean_token_accuracy": 0.6182597279548645, + "num_tokens": 6400776.0, + "step": 247 + }, + { + "epoch": 0.027234790248188007, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 4.637787342071533, + "learning_rate": 9.04099560761347e-08, + "loss": 1.2661, + "mean_token_accuracy": 0.6410377621650696, + "num_tokens": 6425702.0, + "step": 248 + }, + { + "epoch": 0.02734460795080167, + "ewc_loss": 7.860362529754639e-07, + "grad_norm": 4.496292591094971, + "learning_rate": 9.077598828696926e-08, + "loss": 1.2713, + "mean_token_accuracy": 0.6493752598762512, + "num_tokens": 6447754.0, + "step": 249 + }, + { + "epoch": 0.02745442565341533, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 4.023219585418701, + "learning_rate": 9.11420204978038e-08, + "loss": 1.2157, + "mean_token_accuracy": 0.6668927669525146, + "num_tokens": 6477535.0, + "step": 250 + }, + { + "epoch": 0.027564243356028993, + "ewc_loss": 7.934868335723877e-07, + "grad_norm": 4.263810157775879, + "learning_rate": 9.150805270863836e-08, + "loss": 1.2568, + "mean_token_accuracy": 0.6518642902374268, + "num_tokens": 6504584.0, + "step": 251 + }, + { + "epoch": 0.027674061058642653, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 5.383790969848633, + "learning_rate": 9.18740849194729e-08, + "loss": 1.1518, + "mean_token_accuracy": 0.6794507503509521, + "num_tokens": 6529726.0, + "step": 252 + }, + { + "epoch": 0.027783878761256313, + "ewc_loss": 8.009374141693115e-07, + "grad_norm": 6.24947452545166, + "learning_rate": 9.224011713030747e-08, + "loss": 1.3496, + "mean_token_accuracy": 0.6336156725883484, + "num_tokens": 6550265.0, + "step": 253 + }, + { + "epoch": 0.027893696463869976, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 4.69013786315918, + "learning_rate": 9.260614934114202e-08, + "loss": 1.4217, + "mean_token_accuracy": 0.616210401058197, + "num_tokens": 6579733.0, + "step": 254 + }, + { + "epoch": 0.028003514166483636, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 4.304104804992676, + "learning_rate": 9.297218155197657e-08, + "loss": 1.3099, + "mean_token_accuracy": 0.6405676603317261, + "num_tokens": 6613781.0, + "step": 255 + }, + { + "epoch": 0.0281133318690973, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 6.910017967224121, + "learning_rate": 9.333821376281111e-08, + "loss": 1.2222, + "mean_token_accuracy": 0.6495168209075928, + "num_tokens": 6635982.0, + "step": 256 + }, + { + "epoch": 0.02822314957171096, + "ewc_loss": 8.23289155960083e-07, + "grad_norm": 6.2349348068237305, + "learning_rate": 9.370424597364568e-08, + "loss": 1.2877, + "mean_token_accuracy": 0.6370530724525452, + "num_tokens": 6660180.0, + "step": 257 + }, + { + "epoch": 0.028332967274324623, + "ewc_loss": 8.381903171539307e-07, + "grad_norm": 5.731904983520508, + "learning_rate": 9.407027818448023e-08, + "loss": 1.2868, + "mean_token_accuracy": 0.640412449836731, + "num_tokens": 6687798.0, + "step": 258 + }, + { + "epoch": 0.028442784976938282, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 5.214197635650635, + "learning_rate": 9.443631039531479e-08, + "loss": 1.3113, + "mean_token_accuracy": 0.6406944394111633, + "num_tokens": 6716005.0, + "step": 259 + }, + { + "epoch": 0.028552602679551942, + "ewc_loss": 8.568167686462402e-07, + "grad_norm": 4.816781997680664, + "learning_rate": 9.480234260614934e-08, + "loss": 1.2912, + "mean_token_accuracy": 0.6429517865180969, + "num_tokens": 6744026.0, + "step": 260 + }, + { + "epoch": 0.028662420382165606, + "ewc_loss": 8.67992639541626e-07, + "grad_norm": 8.042421340942383, + "learning_rate": 9.516837481698389e-08, + "loss": 1.2604, + "mean_token_accuracy": 0.6452855467796326, + "num_tokens": 6765799.0, + "step": 261 + }, + { + "epoch": 0.028772238084779266, + "ewc_loss": 8.717179298400879e-07, + "grad_norm": 5.37493371963501, + "learning_rate": 9.553440702781844e-08, + "loss": 1.3256, + "mean_token_accuracy": 0.6293798685073853, + "num_tokens": 6792636.0, + "step": 262 + }, + { + "epoch": 0.02888205578739293, + "ewc_loss": 8.754432201385498e-07, + "grad_norm": 8.5234375, + "learning_rate": 9.5900439238653e-08, + "loss": 1.1274, + "mean_token_accuracy": 0.6827950477600098, + "num_tokens": 6814122.0, + "step": 263 + }, + { + "epoch": 0.02899187349000659, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 5.119579315185547, + "learning_rate": 9.626647144948755e-08, + "loss": 1.3386, + "mean_token_accuracy": 0.6275668740272522, + "num_tokens": 6844162.0, + "step": 264 + }, + { + "epoch": 0.029101691192620252, + "ewc_loss": 8.866190910339355e-07, + "grad_norm": 7.213894844055176, + "learning_rate": 9.663250366032211e-08, + "loss": 1.1349, + "mean_token_accuracy": 0.6776092648506165, + "num_tokens": 6867856.0, + "step": 265 + }, + { + "epoch": 0.029211508895233912, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 5.969963073730469, + "learning_rate": 9.699853587115666e-08, + "loss": 1.2351, + "mean_token_accuracy": 0.6626081466674805, + "num_tokens": 6894300.0, + "step": 266 + }, + { + "epoch": 0.029321326597847572, + "ewc_loss": 8.940696716308594e-07, + "grad_norm": 6.299384117126465, + "learning_rate": 9.736456808199121e-08, + "loss": 1.2503, + "mean_token_accuracy": 0.650696873664856, + "num_tokens": 6919880.0, + "step": 267 + }, + { + "epoch": 0.029431144300461235, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 4.985145092010498, + "learning_rate": 9.773060029282576e-08, + "loss": 1.2938, + "mean_token_accuracy": 0.644767701625824, + "num_tokens": 6948610.0, + "step": 268 + }, + { + "epoch": 0.029540962003074895, + "ewc_loss": 9.313225746154785e-07, + "grad_norm": 5.491035461425781, + "learning_rate": 9.809663250366032e-08, + "loss": 1.2726, + "mean_token_accuracy": 0.6433548331260681, + "num_tokens": 6974058.0, + "step": 269 + }, + { + "epoch": 0.02965077970568856, + "ewc_loss": 9.350478649139404e-07, + "grad_norm": 7.108407020568848, + "learning_rate": 9.846266471449487e-08, + "loss": 1.3255, + "mean_token_accuracy": 0.6336990594863892, + "num_tokens": 7001527.0, + "step": 270 + }, + { + "epoch": 0.029760597408302218, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 5.566695213317871, + "learning_rate": 9.882869692532943e-08, + "loss": 1.301, + "mean_token_accuracy": 0.6330639719963074, + "num_tokens": 7025908.0, + "step": 271 + }, + { + "epoch": 0.029870415110915878, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 5.925024032592773, + "learning_rate": 9.919472913616398e-08, + "loss": 1.1982, + "mean_token_accuracy": 0.6575433015823364, + "num_tokens": 7048417.0, + "step": 272 + }, + { + "epoch": 0.02998023281352954, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 8.029288291931152, + "learning_rate": 9.956076134699853e-08, + "loss": 1.235, + "mean_token_accuracy": 0.6522362232208252, + "num_tokens": 7072239.0, + "step": 273 + }, + { + "epoch": 0.0300900505161432, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 5.359362602233887, + "learning_rate": 9.992679355783308e-08, + "loss": 1.2977, + "mean_token_accuracy": 0.639901876449585, + "num_tokens": 7097960.0, + "step": 274 + }, + { + "epoch": 0.030199868218756865, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 6.179714202880859, + "learning_rate": 1.0029282576866764e-07, + "loss": 1.2775, + "mean_token_accuracy": 0.6483479142189026, + "num_tokens": 7129347.0, + "step": 275 + }, + { + "epoch": 0.030309685921370524, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 5.435110569000244, + "learning_rate": 1.0065885797950219e-07, + "loss": 1.2689, + "mean_token_accuracy": 0.643577516078949, + "num_tokens": 7155811.0, + "step": 276 + }, + { + "epoch": 0.030419503623984188, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 6.796477794647217, + "learning_rate": 1.0102489019033675e-07, + "loss": 1.1619, + "mean_token_accuracy": 0.6718885898590088, + "num_tokens": 7175864.0, + "step": 277 + }, + { + "epoch": 0.030529321326597848, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 6.472039699554443, + "learning_rate": 1.013909224011713e-07, + "loss": 1.1907, + "mean_token_accuracy": 0.659977912902832, + "num_tokens": 7197905.0, + "step": 278 + }, + { + "epoch": 0.030639139029211507, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 6.291360855102539, + "learning_rate": 1.0175695461200585e-07, + "loss": 1.2761, + "mean_token_accuracy": 0.6462410688400269, + "num_tokens": 7223710.0, + "step": 279 + }, + { + "epoch": 0.03074895673182517, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 5.545547008514404, + "learning_rate": 1.021229868228404e-07, + "loss": 1.2294, + "mean_token_accuracy": 0.6538143157958984, + "num_tokens": 7245870.0, + "step": 280 + }, + { + "epoch": 0.03085877443443883, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 5.934746742248535, + "learning_rate": 1.0248901903367496e-07, + "loss": 1.2684, + "mean_token_accuracy": 0.6459676623344421, + "num_tokens": 7271138.0, + "step": 281 + }, + { + "epoch": 0.030968592137052494, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 5.767671585083008, + "learning_rate": 1.0285505124450951e-07, + "loss": 1.2853, + "mean_token_accuracy": 0.6483883857727051, + "num_tokens": 7297052.0, + "step": 282 + }, + { + "epoch": 0.031078409839666154, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 5.469801425933838, + "learning_rate": 1.0322108345534407e-07, + "loss": 1.3525, + "mean_token_accuracy": 0.6286933422088623, + "num_tokens": 7326733.0, + "step": 283 + }, + { + "epoch": 0.031188227542279817, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 5.431668758392334, + "learning_rate": 1.0358711566617862e-07, + "loss": 1.2995, + "mean_token_accuracy": 0.6358798742294312, + "num_tokens": 7358626.0, + "step": 284 + }, + { + "epoch": 0.03129804524489348, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 5.117305755615234, + "learning_rate": 1.0395314787701317e-07, + "loss": 1.2992, + "mean_token_accuracy": 0.637411892414093, + "num_tokens": 7385058.0, + "step": 285 + }, + { + "epoch": 0.03140786294750714, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 4.753612518310547, + "learning_rate": 1.0431918008784772e-07, + "loss": 1.255, + "mean_token_accuracy": 0.6519181728363037, + "num_tokens": 7414433.0, + "step": 286 + }, + { + "epoch": 0.0315176806501208, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 6.647675514221191, + "learning_rate": 1.0468521229868228e-07, + "loss": 1.3707, + "mean_token_accuracy": 0.6209607720375061, + "num_tokens": 7446196.0, + "step": 287 + }, + { + "epoch": 0.031627498352734464, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 6.939875602722168, + "learning_rate": 1.0505124450951683e-07, + "loss": 1.2299, + "mean_token_accuracy": 0.6481863856315613, + "num_tokens": 7469003.0, + "step": 288 + }, + { + "epoch": 0.03173731605534812, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 6.932000160217285, + "learning_rate": 1.0541727672035139e-07, + "loss": 1.2934, + "mean_token_accuracy": 0.6415103673934937, + "num_tokens": 7494794.0, + "step": 289 + }, + { + "epoch": 0.03184713375796178, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 6.079611778259277, + "learning_rate": 1.0578330893118593e-07, + "loss": 1.2913, + "mean_token_accuracy": 0.6381270289421082, + "num_tokens": 7531151.0, + "step": 290 + }, + { + "epoch": 0.03195695146057544, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 7.64888334274292, + "learning_rate": 1.0614934114202049e-07, + "loss": 1.1987, + "mean_token_accuracy": 0.6589974164962769, + "num_tokens": 7556362.0, + "step": 291 + }, + { + "epoch": 0.0320667691631891, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 7.048026084899902, + "learning_rate": 1.0651537335285504e-07, + "loss": 1.2869, + "mean_token_accuracy": 0.6391170620918274, + "num_tokens": 7580209.0, + "step": 292 + }, + { + "epoch": 0.03217658686580277, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 5.0174641609191895, + "learning_rate": 1.068814055636896e-07, + "loss": 1.3429, + "mean_token_accuracy": 0.6290076971054077, + "num_tokens": 7607823.0, + "step": 293 + }, + { + "epoch": 0.03228640456841643, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 6.285788536071777, + "learning_rate": 1.0724743777452415e-07, + "loss": 1.3544, + "mean_token_accuracy": 0.6290600299835205, + "num_tokens": 7630083.0, + "step": 294 + }, + { + "epoch": 0.03239622227103009, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 6.251888275146484, + "learning_rate": 1.0761346998535871e-07, + "loss": 1.3431, + "mean_token_accuracy": 0.6320508122444153, + "num_tokens": 7657339.0, + "step": 295 + }, + { + "epoch": 0.03250603997364375, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 7.200750350952148, + "learning_rate": 1.0797950219619325e-07, + "loss": 1.1319, + "mean_token_accuracy": 0.6688212752342224, + "num_tokens": 7678789.0, + "step": 296 + }, + { + "epoch": 0.03261585767625741, + "ewc_loss": 1.0505318641662598e-06, + "grad_norm": 5.35218620300293, + "learning_rate": 1.0834553440702781e-07, + "loss": 1.2132, + "mean_token_accuracy": 0.6623676419258118, + "num_tokens": 7705531.0, + "step": 297 + }, + { + "epoch": 0.032725675378871076, + "ewc_loss": 1.0505318641662598e-06, + "grad_norm": 6.240689277648926, + "learning_rate": 1.0871156661786236e-07, + "loss": 1.2369, + "mean_token_accuracy": 0.6507631540298462, + "num_tokens": 7726907.0, + "step": 298 + }, + { + "epoch": 0.032835493081484736, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 5.265483379364014, + "learning_rate": 1.0907759882869692e-07, + "loss": 1.4032, + "mean_token_accuracy": 0.6289670467376709, + "num_tokens": 7754917.0, + "step": 299 + }, + { + "epoch": 0.032945310784098396, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 7.101071834564209, + "learning_rate": 1.0944363103953147e-07, + "loss": 1.2573, + "mean_token_accuracy": 0.6430339217185974, + "num_tokens": 7780922.0, + "step": 300 + }, + { + "epoch": 0.033055128486712056, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 7.154463291168213, + "learning_rate": 1.0980966325036604e-07, + "loss": 1.2773, + "mean_token_accuracy": 0.6463287472724915, + "num_tokens": 7807004.0, + "step": 301 + }, + { + "epoch": 0.03316494618932572, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 7.503502368927002, + "learning_rate": 1.1017569546120057e-07, + "loss": 1.1577, + "mean_token_accuracy": 0.6651064157485962, + "num_tokens": 7830480.0, + "step": 302 + }, + { + "epoch": 0.03327476389193938, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 6.255960464477539, + "learning_rate": 1.1054172767203513e-07, + "loss": 1.2618, + "mean_token_accuracy": 0.6431233882904053, + "num_tokens": 7856537.0, + "step": 303 + }, + { + "epoch": 0.03338458159455304, + "ewc_loss": 1.0952353477478027e-06, + "grad_norm": 7.912069320678711, + "learning_rate": 1.1090775988286968e-07, + "loss": 1.1908, + "mean_token_accuracy": 0.6631734371185303, + "num_tokens": 7875428.0, + "step": 304 + }, + { + "epoch": 0.0334943992971667, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 6.567294597625732, + "learning_rate": 1.1127379209370424e-07, + "loss": 1.2212, + "mean_token_accuracy": 0.6525173187255859, + "num_tokens": 7899765.0, + "step": 305 + }, + { + "epoch": 0.03360421699978036, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 6.539906978607178, + "learning_rate": 1.116398243045388e-07, + "loss": 1.2536, + "mean_token_accuracy": 0.647377610206604, + "num_tokens": 7922002.0, + "step": 306 + }, + { + "epoch": 0.03371403470239403, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 6.526037216186523, + "learning_rate": 1.1200585651537336e-07, + "loss": 1.2614, + "mean_token_accuracy": 0.6526790261268616, + "num_tokens": 7945165.0, + "step": 307 + }, + { + "epoch": 0.03382385240500769, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 7.121865272521973, + "learning_rate": 1.1237188872620789e-07, + "loss": 1.1835, + "mean_token_accuracy": 0.6719474792480469, + "num_tokens": 7967060.0, + "step": 308 + }, + { + "epoch": 0.03393367010762135, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 4.780085563659668, + "learning_rate": 1.1273792093704245e-07, + "loss": 1.2028, + "mean_token_accuracy": 0.6556016802787781, + "num_tokens": 7988205.0, + "step": 309 + }, + { + "epoch": 0.03404348781023501, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 5.608370780944824, + "learning_rate": 1.13103953147877e-07, + "loss": 1.1457, + "mean_token_accuracy": 0.6649461388587952, + "num_tokens": 8018516.0, + "step": 310 + }, + { + "epoch": 0.03415330551284867, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 4.694137096405029, + "learning_rate": 1.1346998535871157e-07, + "loss": 1.2421, + "mean_token_accuracy": 0.6548426151275635, + "num_tokens": 8049713.0, + "step": 311 + }, + { + "epoch": 0.034263123215462335, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 5.089285373687744, + "learning_rate": 1.1383601756954611e-07, + "loss": 1.2411, + "mean_token_accuracy": 0.6536561250686646, + "num_tokens": 8077271.0, + "step": 312 + }, + { + "epoch": 0.034372940918075995, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 6.406395435333252, + "learning_rate": 1.1420204978038068e-07, + "loss": 1.2498, + "mean_token_accuracy": 0.6557976007461548, + "num_tokens": 8100334.0, + "step": 313 + }, + { + "epoch": 0.034482758620689655, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 5.169366359710693, + "learning_rate": 1.1456808199121521e-07, + "loss": 1.2688, + "mean_token_accuracy": 0.6614073514938354, + "num_tokens": 8127291.0, + "step": 314 + }, + { + "epoch": 0.034592576323303315, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 4.108391284942627, + "learning_rate": 1.1493411420204978e-07, + "loss": 1.3012, + "mean_token_accuracy": 0.6378589868545532, + "num_tokens": 8158273.0, + "step": 315 + }, + { + "epoch": 0.034702394025916974, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 6.834139347076416, + "learning_rate": 1.1530014641288432e-07, + "loss": 1.1688, + "mean_token_accuracy": 0.6637930274009705, + "num_tokens": 8180056.0, + "step": 316 + }, + { + "epoch": 0.03481221172853064, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 6.559394836425781, + "learning_rate": 1.1566617862371889e-07, + "loss": 1.2431, + "mean_token_accuracy": 0.6539621949195862, + "num_tokens": 8203267.0, + "step": 317 + }, + { + "epoch": 0.0349220294311443, + "ewc_loss": 1.214444637298584e-06, + "grad_norm": 6.405194282531738, + "learning_rate": 1.1603221083455344e-07, + "loss": 1.2344, + "mean_token_accuracy": 0.6571062803268433, + "num_tokens": 8225551.0, + "step": 318 + }, + { + "epoch": 0.03503184713375796, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 5.800381183624268, + "learning_rate": 1.16398243045388e-07, + "loss": 1.2847, + "mean_token_accuracy": 0.6336801648139954, + "num_tokens": 8249591.0, + "step": 319 + }, + { + "epoch": 0.03514166483637162, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 5.880638599395752, + "learning_rate": 1.1676427525622253e-07, + "loss": 1.2528, + "mean_token_accuracy": 0.6431673765182495, + "num_tokens": 8274565.0, + "step": 320 + }, + { + "epoch": 0.03525148253898529, + "ewc_loss": 1.2442469596862793e-06, + "grad_norm": 7.32682991027832, + "learning_rate": 1.171303074670571e-07, + "loss": 1.2725, + "mean_token_accuracy": 0.6422048211097717, + "num_tokens": 8301284.0, + "step": 321 + }, + { + "epoch": 0.03536130024159895, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 5.999466419219971, + "learning_rate": 1.1749633967789165e-07, + "loss": 1.2444, + "mean_token_accuracy": 0.6443246603012085, + "num_tokens": 8322113.0, + "step": 322 + }, + { + "epoch": 0.03547111794421261, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 6.7216715812683105, + "learning_rate": 1.1786237188872621e-07, + "loss": 1.2796, + "mean_token_accuracy": 0.6372165083885193, + "num_tokens": 8349315.0, + "step": 323 + }, + { + "epoch": 0.03558093564682627, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 5.786715507507324, + "learning_rate": 1.1822840409956076e-07, + "loss": 1.2385, + "mean_token_accuracy": 0.6486908793449402, + "num_tokens": 8369953.0, + "step": 324 + }, + { + "epoch": 0.03569075334943993, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 4.792079448699951, + "learning_rate": 1.1859443631039532e-07, + "loss": 1.27, + "mean_token_accuracy": 0.6449906826019287, + "num_tokens": 8394310.0, + "step": 325 + }, + { + "epoch": 0.035800571052053594, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 6.021547794342041, + "learning_rate": 1.1896046852122986e-07, + "loss": 1.3212, + "mean_token_accuracy": 0.6314225196838379, + "num_tokens": 8419318.0, + "step": 326 + }, + { + "epoch": 0.035910388754667254, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 6.1286725997924805, + "learning_rate": 1.1932650073206443e-07, + "loss": 1.1006, + "mean_token_accuracy": 0.6802542805671692, + "num_tokens": 8443229.0, + "step": 327 + }, + { + "epoch": 0.036020206457280914, + "ewc_loss": 1.296401023864746e-06, + "grad_norm": 5.7851738929748535, + "learning_rate": 1.1969253294289895e-07, + "loss": 1.2526, + "mean_token_accuracy": 0.6514350175857544, + "num_tokens": 8467929.0, + "step": 328 + }, + { + "epoch": 0.03613002415989457, + "ewc_loss": 1.3113021850585938e-06, + "grad_norm": 5.181347846984863, + "learning_rate": 1.2005856515373353e-07, + "loss": 1.2151, + "mean_token_accuracy": 0.6570374369621277, + "num_tokens": 8491470.0, + "step": 329 + }, + { + "epoch": 0.03623984186250823, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 5.409581184387207, + "learning_rate": 1.2042459736456808e-07, + "loss": 1.1834, + "mean_token_accuracy": 0.6633471250534058, + "num_tokens": 8519715.0, + "step": 330 + }, + { + "epoch": 0.0363496595651219, + "ewc_loss": 1.3336539268493652e-06, + "grad_norm": 7.511041164398193, + "learning_rate": 1.2079062957540263e-07, + "loss": 1.2428, + "mean_token_accuracy": 0.6485503315925598, + "num_tokens": 8542376.0, + "step": 331 + }, + { + "epoch": 0.03645947726773556, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 5.172972679138184, + "learning_rate": 1.2115666178623718e-07, + "loss": 1.2532, + "mean_token_accuracy": 0.6517549157142639, + "num_tokens": 8564513.0, + "step": 332 + }, + { + "epoch": 0.03656929497034922, + "ewc_loss": 1.3336539268493652e-06, + "grad_norm": 6.18115234375, + "learning_rate": 1.2152269399707175e-07, + "loss": 1.1743, + "mean_token_accuracy": 0.6680352687835693, + "num_tokens": 8587957.0, + "step": 333 + }, + { + "epoch": 0.03667911267296288, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 6.260185718536377, + "learning_rate": 1.2188872620790627e-07, + "loss": 1.194, + "mean_token_accuracy": 0.6672407984733582, + "num_tokens": 8612155.0, + "step": 334 + }, + { + "epoch": 0.03678893037557654, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 5.7313408851623535, + "learning_rate": 1.2225475841874085e-07, + "loss": 1.2154, + "mean_token_accuracy": 0.6619741916656494, + "num_tokens": 8643677.0, + "step": 335 + }, + { + "epoch": 0.036898748078190206, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 5.681445121765137, + "learning_rate": 1.226207906295754e-07, + "loss": 1.1713, + "mean_token_accuracy": 0.6642433404922485, + "num_tokens": 8672599.0, + "step": 336 + }, + { + "epoch": 0.037008565780803866, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 5.222816467285156, + "learning_rate": 1.2298682284040995e-07, + "loss": 1.2635, + "mean_token_accuracy": 0.6556886434555054, + "num_tokens": 8696109.0, + "step": 337 + }, + { + "epoch": 0.037118383483417526, + "ewc_loss": 1.385807991027832e-06, + "grad_norm": 5.157867431640625, + "learning_rate": 1.233528550512445e-07, + "loss": 1.169, + "mean_token_accuracy": 0.6711331605911255, + "num_tokens": 8725617.0, + "step": 338 + }, + { + "epoch": 0.037228201186031186, + "ewc_loss": 1.385807991027832e-06, + "grad_norm": 5.758976936340332, + "learning_rate": 1.2371888726207907e-07, + "loss": 1.1797, + "mean_token_accuracy": 0.6660201549530029, + "num_tokens": 8753416.0, + "step": 339 + }, + { + "epoch": 0.03733801888864485, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 5.2311553955078125, + "learning_rate": 1.240849194729136e-07, + "loss": 1.2435, + "mean_token_accuracy": 0.6442667245864868, + "num_tokens": 8781610.0, + "step": 340 + }, + { + "epoch": 0.03744783659125851, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 4.696457386016846, + "learning_rate": 1.2445095168374817e-07, + "loss": 1.1766, + "mean_token_accuracy": 0.6575323343276978, + "num_tokens": 8805386.0, + "step": 341 + }, + { + "epoch": 0.03755765429387217, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 6.274916648864746, + "learning_rate": 1.2481698389458272e-07, + "loss": 1.3159, + "mean_token_accuracy": 0.6364549398422241, + "num_tokens": 8830009.0, + "step": 342 + }, + { + "epoch": 0.03766747199648583, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 6.75684928894043, + "learning_rate": 1.2518301610541727e-07, + "loss": 1.1853, + "mean_token_accuracy": 0.6593161225318909, + "num_tokens": 8848499.0, + "step": 343 + }, + { + "epoch": 0.03777728969909949, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 5.857545852661133, + "learning_rate": 1.2554904831625182e-07, + "loss": 1.16, + "mean_token_accuracy": 0.6642367839813232, + "num_tokens": 8869974.0, + "step": 344 + }, + { + "epoch": 0.03788710740171316, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 6.58172082901001, + "learning_rate": 1.259150805270864e-07, + "loss": 1.2841, + "mean_token_accuracy": 0.6391716003417969, + "num_tokens": 8897952.0, + "step": 345 + }, + { + "epoch": 0.03799692510432682, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 4.94627571105957, + "learning_rate": 1.2628111273792094e-07, + "loss": 1.208, + "mean_token_accuracy": 0.6609097719192505, + "num_tokens": 8931394.0, + "step": 346 + }, + { + "epoch": 0.03810674280694048, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 4.67088508605957, + "learning_rate": 1.2664714494875547e-07, + "loss": 1.2585, + "mean_token_accuracy": 0.6492675542831421, + "num_tokens": 8962484.0, + "step": 347 + }, + { + "epoch": 0.03821656050955414, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 5.695640563964844, + "learning_rate": 1.2701317715959004e-07, + "loss": 1.2234, + "mean_token_accuracy": 0.6501524448394775, + "num_tokens": 8988810.0, + "step": 348 + }, + { + "epoch": 0.0383263782121678, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 7.077723979949951, + "learning_rate": 1.273792093704246e-07, + "loss": 1.1689, + "mean_token_accuracy": 0.6685583591461182, + "num_tokens": 9011086.0, + "step": 349 + }, + { + "epoch": 0.038436195914781465, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 5.8831787109375, + "learning_rate": 1.2774524158125917e-07, + "loss": 1.2451, + "mean_token_accuracy": 0.6503923535346985, + "num_tokens": 9038964.0, + "step": 350 + }, + { + "epoch": 0.038546013617395125, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 5.700845241546631, + "learning_rate": 1.281112737920937e-07, + "loss": 1.3143, + "mean_token_accuracy": 0.6297289133071899, + "num_tokens": 9065347.0, + "step": 351 + }, + { + "epoch": 0.038655831320008785, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 7.125729560852051, + "learning_rate": 1.2847730600292824e-07, + "loss": 1.1451, + "mean_token_accuracy": 0.6755622625350952, + "num_tokens": 9089414.0, + "step": 352 + }, + { + "epoch": 0.038765649022622445, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 4.646429061889648, + "learning_rate": 1.288433382137628e-07, + "loss": 1.3095, + "mean_token_accuracy": 0.6372082233428955, + "num_tokens": 9120551.0, + "step": 353 + }, + { + "epoch": 0.038875466725236105, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 5.661139965057373, + "learning_rate": 1.2920937042459736e-07, + "loss": 1.2108, + "mean_token_accuracy": 0.6589045524597168, + "num_tokens": 9143936.0, + "step": 354 + }, + { + "epoch": 0.03898528442784977, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 5.454245567321777, + "learning_rate": 1.295754026354319e-07, + "loss": 1.2943, + "mean_token_accuracy": 0.6338184475898743, + "num_tokens": 9171765.0, + "step": 355 + }, + { + "epoch": 0.03909510213046343, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 5.538103103637695, + "learning_rate": 1.2994143484626646e-07, + "loss": 1.2496, + "mean_token_accuracy": 0.6455378532409668, + "num_tokens": 9198275.0, + "step": 356 + }, + { + "epoch": 0.03920491983307709, + "ewc_loss": 1.4677643775939941e-06, + "grad_norm": 6.884566783905029, + "learning_rate": 1.3030746705710104e-07, + "loss": 1.2002, + "mean_token_accuracy": 0.6511995792388916, + "num_tokens": 9221218.0, + "step": 357 + }, + { + "epoch": 0.03931473753569075, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 4.709895610809326, + "learning_rate": 1.3067349926793558e-07, + "loss": 1.2005, + "mean_token_accuracy": 0.6512102484703064, + "num_tokens": 9249309.0, + "step": 358 + }, + { + "epoch": 0.03942455523830442, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 6.382948398590088, + "learning_rate": 1.310395314787701e-07, + "loss": 1.2305, + "mean_token_accuracy": 0.6491957902908325, + "num_tokens": 9270701.0, + "step": 359 + }, + { + "epoch": 0.03953437294091808, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 5.083617687225342, + "learning_rate": 1.3140556368960468e-07, + "loss": 1.2635, + "mean_token_accuracy": 0.6476063132286072, + "num_tokens": 9303156.0, + "step": 360 + }, + { + "epoch": 0.03964419064353174, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 6.72424840927124, + "learning_rate": 1.3177159590043923e-07, + "loss": 1.1595, + "mean_token_accuracy": 0.6732094883918762, + "num_tokens": 9325163.0, + "step": 361 + }, + { + "epoch": 0.0397540083461454, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 6.433728218078613, + "learning_rate": 1.321376281112738e-07, + "loss": 1.1317, + "mean_token_accuracy": 0.6767134666442871, + "num_tokens": 9349474.0, + "step": 362 + }, + { + "epoch": 0.03986382604875906, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 5.025449752807617, + "learning_rate": 1.3250366032210833e-07, + "loss": 1.1967, + "mean_token_accuracy": 0.6584434509277344, + "num_tokens": 9376227.0, + "step": 363 + }, + { + "epoch": 0.039973643751372724, + "ewc_loss": 1.5050172805786133e-06, + "grad_norm": 4.931293487548828, + "learning_rate": 1.3286969253294288e-07, + "loss": 1.1465, + "mean_token_accuracy": 0.6694327592849731, + "num_tokens": 9398802.0, + "step": 364 + }, + { + "epoch": 0.040083461453986384, + "ewc_loss": 1.5050172805786133e-06, + "grad_norm": 4.495023727416992, + "learning_rate": 1.3323572474377745e-07, + "loss": 1.2288, + "mean_token_accuracy": 0.6521861553192139, + "num_tokens": 9425588.0, + "step": 365 + }, + { + "epoch": 0.040193279156600044, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 3.9464118480682373, + "learning_rate": 1.33601756954612e-07, + "loss": 1.2801, + "mean_token_accuracy": 0.6334925889968872, + "num_tokens": 9457907.0, + "step": 366 + }, + { + "epoch": 0.040303096859213704, + "ewc_loss": 1.5273690223693848e-06, + "grad_norm": 5.694772720336914, + "learning_rate": 1.3396778916544655e-07, + "loss": 1.1297, + "mean_token_accuracy": 0.6829952597618103, + "num_tokens": 9482988.0, + "step": 367 + }, + { + "epoch": 0.040412914561827364, + "ewc_loss": 1.5348196029663086e-06, + "grad_norm": 6.670355319976807, + "learning_rate": 1.343338213762811e-07, + "loss": 1.2469, + "mean_token_accuracy": 0.6466431021690369, + "num_tokens": 9503553.0, + "step": 368 + }, + { + "epoch": 0.04052273226444103, + "ewc_loss": 1.5348196029663086e-06, + "grad_norm": 4.608431816101074, + "learning_rate": 1.3469985358711565e-07, + "loss": 1.2456, + "mean_token_accuracy": 0.6495547294616699, + "num_tokens": 9528274.0, + "step": 369 + }, + { + "epoch": 0.04063254996705469, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 4.043929576873779, + "learning_rate": 1.3506588579795023e-07, + "loss": 1.2072, + "mean_token_accuracy": 0.6532905697822571, + "num_tokens": 9563887.0, + "step": 370 + }, + { + "epoch": 0.04074236766966835, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 7.15656852722168, + "learning_rate": 1.3543191800878475e-07, + "loss": 1.2582, + "mean_token_accuracy": 0.647301435470581, + "num_tokens": 9586556.0, + "step": 371 + }, + { + "epoch": 0.04085218537228201, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 5.163967609405518, + "learning_rate": 1.3579795021961932e-07, + "loss": 1.2083, + "mean_token_accuracy": 0.65693199634552, + "num_tokens": 9616303.0, + "step": 372 + }, + { + "epoch": 0.04096200307489567, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 5.476652145385742, + "learning_rate": 1.3616398243045387e-07, + "loss": 1.2337, + "mean_token_accuracy": 0.6514513492584229, + "num_tokens": 9645693.0, + "step": 373 + }, + { + "epoch": 0.04107182077750934, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 6.408543586730957, + "learning_rate": 1.3653001464128845e-07, + "loss": 1.2272, + "mean_token_accuracy": 0.6561509370803833, + "num_tokens": 9672371.0, + "step": 374 + }, + { + "epoch": 0.041181638480122996, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 4.60734224319458, + "learning_rate": 1.3689604685212297e-07, + "loss": 1.1883, + "mean_token_accuracy": 0.6727520823478699, + "num_tokens": 9705264.0, + "step": 375 + }, + { + "epoch": 0.041291456182736656, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 5.057756423950195, + "learning_rate": 1.3726207906295752e-07, + "loss": 1.1309, + "mean_token_accuracy": 0.6755083799362183, + "num_tokens": 9729726.0, + "step": 376 + }, + { + "epoch": 0.041401273885350316, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 5.418838977813721, + "learning_rate": 1.376281112737921e-07, + "loss": 1.2545, + "mean_token_accuracy": 0.6536494493484497, + "num_tokens": 9756079.0, + "step": 377 + }, + { + "epoch": 0.04151109158796398, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 5.456227779388428, + "learning_rate": 1.3799414348462665e-07, + "loss": 1.1122, + "mean_token_accuracy": 0.6804099082946777, + "num_tokens": 9778317.0, + "step": 378 + }, + { + "epoch": 0.04162090929057764, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 5.444605827331543, + "learning_rate": 1.383601756954612e-07, + "loss": 1.244, + "mean_token_accuracy": 0.6498316526412964, + "num_tokens": 9808114.0, + "step": 379 + }, + { + "epoch": 0.0417307269931913, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 4.530700206756592, + "learning_rate": 1.3872620790629574e-07, + "loss": 1.2229, + "mean_token_accuracy": 0.6433988809585571, + "num_tokens": 9833909.0, + "step": 380 + }, + { + "epoch": 0.04184054469580496, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 5.327883243560791, + "learning_rate": 1.390922401171303e-07, + "loss": 1.1866, + "mean_token_accuracy": 0.6639984846115112, + "num_tokens": 9857091.0, + "step": 381 + }, + { + "epoch": 0.04195036239841862, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 5.420687675476074, + "learning_rate": 1.3945827232796487e-07, + "loss": 1.1526, + "mean_token_accuracy": 0.6670289635658264, + "num_tokens": 9880395.0, + "step": 382 + }, + { + "epoch": 0.04206018010103229, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 5.588293552398682, + "learning_rate": 1.398243045387994e-07, + "loss": 1.2747, + "mean_token_accuracy": 0.6413223743438721, + "num_tokens": 9904483.0, + "step": 383 + }, + { + "epoch": 0.04216999780364595, + "ewc_loss": 1.6242265701293945e-06, + "grad_norm": 4.939534664154053, + "learning_rate": 1.4019033674963397e-07, + "loss": 1.1556, + "mean_token_accuracy": 0.6664032340049744, + "num_tokens": 9927865.0, + "step": 384 + }, + { + "epoch": 0.04227981550625961, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 5.443539142608643, + "learning_rate": 1.4055636896046852e-07, + "loss": 1.2374, + "mean_token_accuracy": 0.6559125781059265, + "num_tokens": 9956712.0, + "step": 385 + }, + { + "epoch": 0.04238963320887327, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 4.811751365661621, + "learning_rate": 1.409224011713031e-07, + "loss": 1.0861, + "mean_token_accuracy": 0.6902910470962524, + "num_tokens": 9979222.0, + "step": 386 + }, + { + "epoch": 0.04249945091148693, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 6.387205600738525, + "learning_rate": 1.4128843338213761e-07, + "loss": 1.2202, + "mean_token_accuracy": 0.6480622887611389, + "num_tokens": 10007509.0, + "step": 387 + }, + { + "epoch": 0.042609268614100596, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 4.9873480796813965, + "learning_rate": 1.4165446559297216e-07, + "loss": 1.2559, + "mean_token_accuracy": 0.649761438369751, + "num_tokens": 10037020.0, + "step": 388 + }, + { + "epoch": 0.042719086316714255, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 5.386353015899658, + "learning_rate": 1.4202049780380674e-07, + "loss": 1.2781, + "mean_token_accuracy": 0.6387033462524414, + "num_tokens": 10062076.0, + "step": 389 + }, + { + "epoch": 0.042828904019327915, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 4.241064071655273, + "learning_rate": 1.423865300146413e-07, + "loss": 1.1923, + "mean_token_accuracy": 0.6594594717025757, + "num_tokens": 10093034.0, + "step": 390 + }, + { + "epoch": 0.042938721721941575, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 5.647027969360352, + "learning_rate": 1.4275256222547584e-07, + "loss": 1.2234, + "mean_token_accuracy": 0.6586716771125793, + "num_tokens": 10117118.0, + "step": 391 + }, + { + "epoch": 0.043048539424555235, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 5.06908655166626, + "learning_rate": 1.4311859443631039e-07, + "loss": 1.2657, + "mean_token_accuracy": 0.6443292498588562, + "num_tokens": 10151018.0, + "step": 392 + }, + { + "epoch": 0.0431583571271689, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 9.049090385437012, + "learning_rate": 1.4348462664714493e-07, + "loss": 1.1085, + "mean_token_accuracy": 0.6774696707725525, + "num_tokens": 10168012.0, + "step": 393 + }, + { + "epoch": 0.04326817482978256, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 5.650341033935547, + "learning_rate": 1.438506588579795e-07, + "loss": 1.2403, + "mean_token_accuracy": 0.6462737917900085, + "num_tokens": 10193267.0, + "step": 394 + }, + { + "epoch": 0.04337799253239622, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 4.474823951721191, + "learning_rate": 1.4421669106881403e-07, + "loss": 1.1465, + "mean_token_accuracy": 0.6627634763717651, + "num_tokens": 10219769.0, + "step": 395 + }, + { + "epoch": 0.04348781023500988, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 5.668895244598389, + "learning_rate": 1.445827232796486e-07, + "loss": 1.2524, + "mean_token_accuracy": 0.6446805000305176, + "num_tokens": 10244899.0, + "step": 396 + }, + { + "epoch": 0.04359762793762355, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 7.270735263824463, + "learning_rate": 1.4494875549048316e-07, + "loss": 1.192, + "mean_token_accuracy": 0.6596347093582153, + "num_tokens": 10263877.0, + "step": 397 + }, + { + "epoch": 0.04370744564023721, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 5.148193836212158, + "learning_rate": 1.4531478770131773e-07, + "loss": 1.2513, + "mean_token_accuracy": 0.6459014415740967, + "num_tokens": 10287450.0, + "step": 398 + }, + { + "epoch": 0.04381726334285087, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 4.856070041656494, + "learning_rate": 1.4568081991215226e-07, + "loss": 1.2062, + "mean_token_accuracy": 0.6656529307365417, + "num_tokens": 10314939.0, + "step": 399 + }, + { + "epoch": 0.04392708104546453, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 4.8150434494018555, + "learning_rate": 1.460468521229868e-07, + "loss": 1.096, + "mean_token_accuracy": 0.6861969828605652, + "num_tokens": 10340075.0, + "step": 400 + }, + { + "epoch": 0.04403689874807819, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 4.957571506500244, + "learning_rate": 1.4641288433382138e-07, + "loss": 1.249, + "mean_token_accuracy": 0.65215003490448, + "num_tokens": 10364116.0, + "step": 401 + }, + { + "epoch": 0.044146716450691854, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 5.2244391441345215, + "learning_rate": 1.4677891654465593e-07, + "loss": 1.2619, + "mean_token_accuracy": 0.6491000652313232, + "num_tokens": 10391839.0, + "step": 402 + }, + { + "epoch": 0.044256534153305514, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 4.236721992492676, + "learning_rate": 1.4714494875549048e-07, + "loss": 1.1965, + "mean_token_accuracy": 0.6737003326416016, + "num_tokens": 10419954.0, + "step": 403 + }, + { + "epoch": 0.044366351855919174, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 6.095024585723877, + "learning_rate": 1.4751098096632503e-07, + "loss": 1.2576, + "mean_token_accuracy": 0.6434000134468079, + "num_tokens": 10444143.0, + "step": 404 + }, + { + "epoch": 0.044476169558532834, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 4.971898078918457, + "learning_rate": 1.4787701317715958e-07, + "loss": 1.19, + "mean_token_accuracy": 0.6603885889053345, + "num_tokens": 10473697.0, + "step": 405 + }, + { + "epoch": 0.044585987261146494, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 5.867884635925293, + "learning_rate": 1.4824304538799415e-07, + "loss": 1.1847, + "mean_token_accuracy": 0.6643599271774292, + "num_tokens": 10502612.0, + "step": 406 + }, + { + "epoch": 0.04469580496376016, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 4.99357271194458, + "learning_rate": 1.4860907759882867e-07, + "loss": 1.228, + "mean_token_accuracy": 0.6500723361968994, + "num_tokens": 10529266.0, + "step": 407 + }, + { + "epoch": 0.04480562266637382, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 5.305943965911865, + "learning_rate": 1.4897510980966325e-07, + "loss": 1.2076, + "mean_token_accuracy": 0.6561135649681091, + "num_tokens": 10550616.0, + "step": 408 + }, + { + "epoch": 0.04491544036898748, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 4.714290618896484, + "learning_rate": 1.493411420204978e-07, + "loss": 1.1716, + "mean_token_accuracy": 0.6656228303909302, + "num_tokens": 10582682.0, + "step": 409 + }, + { + "epoch": 0.04502525807160114, + "ewc_loss": 1.7285346984863281e-06, + "grad_norm": 5.177734375, + "learning_rate": 1.4970717423133235e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.6802870035171509, + "num_tokens": 10606229.0, + "step": 410 + }, + { + "epoch": 0.0451350757742148, + "ewc_loss": 1.7285346984863281e-06, + "grad_norm": 4.130097389221191, + "learning_rate": 1.500732064421669e-07, + "loss": 1.2489, + "mean_token_accuracy": 0.6445169448852539, + "num_tokens": 10639889.0, + "step": 411 + }, + { + "epoch": 0.04524489347682847, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 4.472767353057861, + "learning_rate": 1.5043923865300145e-07, + "loss": 1.255, + "mean_token_accuracy": 0.6403805613517761, + "num_tokens": 10664352.0, + "step": 412 + }, + { + "epoch": 0.04535471117944213, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 4.908764839172363, + "learning_rate": 1.5080527086383602e-07, + "loss": 1.1648, + "mean_token_accuracy": 0.6703236103057861, + "num_tokens": 10692255.0, + "step": 413 + }, + { + "epoch": 0.04546452888205579, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 6.310358047485352, + "learning_rate": 1.5117130307467057e-07, + "loss": 1.1884, + "mean_token_accuracy": 0.6666712164878845, + "num_tokens": 10713360.0, + "step": 414 + }, + { + "epoch": 0.045574346584669446, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 5.045009613037109, + "learning_rate": 1.5153733528550512e-07, + "loss": 1.1758, + "mean_token_accuracy": 0.6663499474525452, + "num_tokens": 10743290.0, + "step": 415 + }, + { + "epoch": 0.04568416428728311, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 5.636015892028809, + "learning_rate": 1.5190336749633967e-07, + "loss": 1.1917, + "mean_token_accuracy": 0.6531072854995728, + "num_tokens": 10771567.0, + "step": 416 + }, + { + "epoch": 0.04579398198989677, + "ewc_loss": 1.7583370208740234e-06, + "grad_norm": 4.448980808258057, + "learning_rate": 1.5226939970717422e-07, + "loss": 1.1544, + "mean_token_accuracy": 0.6805377006530762, + "num_tokens": 10796435.0, + "step": 417 + }, + { + "epoch": 0.04590379969251043, + "ewc_loss": 1.7657876014709473e-06, + "grad_norm": 4.692486763000488, + "learning_rate": 1.526354319180088e-07, + "loss": 1.2179, + "mean_token_accuracy": 0.6631459593772888, + "num_tokens": 10826350.0, + "step": 418 + }, + { + "epoch": 0.04601361739512409, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 5.179934978485107, + "learning_rate": 1.5300146412884332e-07, + "loss": 1.1555, + "mean_token_accuracy": 0.6683366298675537, + "num_tokens": 10851407.0, + "step": 419 + }, + { + "epoch": 0.04612343509773775, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 5.0400919914245605, + "learning_rate": 1.533674963396779e-07, + "loss": 1.3076, + "mean_token_accuracy": 0.645281195640564, + "num_tokens": 10885508.0, + "step": 420 + }, + { + "epoch": 0.04623325280035142, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 4.881284713745117, + "learning_rate": 1.5373352855051244e-07, + "loss": 1.1892, + "mean_token_accuracy": 0.6551609635353088, + "num_tokens": 10911158.0, + "step": 421 + }, + { + "epoch": 0.04634307050296508, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 5.828506946563721, + "learning_rate": 1.54099560761347e-07, + "loss": 1.1243, + "mean_token_accuracy": 0.6759604811668396, + "num_tokens": 10932147.0, + "step": 422 + }, + { + "epoch": 0.04645288820557874, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 5.237981796264648, + "learning_rate": 1.5446559297218154e-07, + "loss": 1.1774, + "mean_token_accuracy": 0.6640245914459229, + "num_tokens": 10952566.0, + "step": 423 + }, + { + "epoch": 0.0465627059081924, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 6.54088830947876, + "learning_rate": 1.548316251830161e-07, + "loss": 1.1468, + "mean_token_accuracy": 0.6698017120361328, + "num_tokens": 10971181.0, + "step": 424 + }, + { + "epoch": 0.04667252361080606, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 5.224094867706299, + "learning_rate": 1.5519765739385066e-07, + "loss": 1.109, + "mean_token_accuracy": 0.6792911291122437, + "num_tokens": 10991562.0, + "step": 425 + }, + { + "epoch": 0.046782341313419726, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 4.995647430419922, + "learning_rate": 1.5556368960468521e-07, + "loss": 1.1831, + "mean_token_accuracy": 0.6600173711776733, + "num_tokens": 11016836.0, + "step": 426 + }, + { + "epoch": 0.046892159016033386, + "ewc_loss": 1.7955899238586426e-06, + "grad_norm": 4.155873775482178, + "learning_rate": 1.5592972181551974e-07, + "loss": 1.2262, + "mean_token_accuracy": 0.6515845060348511, + "num_tokens": 11040431.0, + "step": 427 + }, + { + "epoch": 0.047001976718647046, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 4.408470630645752, + "learning_rate": 1.562957540263543e-07, + "loss": 1.1734, + "mean_token_accuracy": 0.6656243205070496, + "num_tokens": 11065612.0, + "step": 428 + }, + { + "epoch": 0.047111794421260705, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 5.928896903991699, + "learning_rate": 1.5666178623718886e-07, + "loss": 1.1224, + "mean_token_accuracy": 0.6804638504981995, + "num_tokens": 11090027.0, + "step": 429 + }, + { + "epoch": 0.047221612123874365, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 3.960087299346924, + "learning_rate": 1.5702781844802344e-07, + "loss": 1.2872, + "mean_token_accuracy": 0.6346864104270935, + "num_tokens": 11121097.0, + "step": 430 + }, + { + "epoch": 0.04733142982648803, + "ewc_loss": 1.8253922462463379e-06, + "grad_norm": 5.855001449584961, + "learning_rate": 1.5739385065885796e-07, + "loss": 1.2029, + "mean_token_accuracy": 0.6531655788421631, + "num_tokens": 11144282.0, + "step": 431 + }, + { + "epoch": 0.04744124752910169, + "ewc_loss": 1.8253922462463379e-06, + "grad_norm": 4.616157531738281, + "learning_rate": 1.5775988286969253e-07, + "loss": 1.205, + "mean_token_accuracy": 0.6611601710319519, + "num_tokens": 11172508.0, + "step": 432 + }, + { + "epoch": 0.04755106523171535, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 4.950656890869141, + "learning_rate": 1.5812591508052708e-07, + "loss": 1.1557, + "mean_token_accuracy": 0.6717765927314758, + "num_tokens": 11194936.0, + "step": 433 + }, + { + "epoch": 0.04766088293432901, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 4.271273136138916, + "learning_rate": 1.5849194729136163e-07, + "loss": 1.0583, + "mean_token_accuracy": 0.6908203363418579, + "num_tokens": 11220873.0, + "step": 434 + }, + { + "epoch": 0.04777070063694268, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 5.172623157501221, + "learning_rate": 1.5885797950219618e-07, + "loss": 1.2001, + "mean_token_accuracy": 0.6556992530822754, + "num_tokens": 11245691.0, + "step": 435 + }, + { + "epoch": 0.04788051833955634, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 4.720003128051758, + "learning_rate": 1.5922401171303073e-07, + "loss": 1.2476, + "mean_token_accuracy": 0.6640490293502808, + "num_tokens": 11272386.0, + "step": 436 + }, + { + "epoch": 0.04799033604217, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 4.239912986755371, + "learning_rate": 1.595900439238653e-07, + "loss": 1.2529, + "mean_token_accuracy": 0.6491910219192505, + "num_tokens": 11301211.0, + "step": 437 + }, + { + "epoch": 0.04810015374478366, + "ewc_loss": 1.8849968910217285e-06, + "grad_norm": 3.7796292304992676, + "learning_rate": 1.5995607613469986e-07, + "loss": 1.2441, + "mean_token_accuracy": 0.6523476839065552, + "num_tokens": 11331253.0, + "step": 438 + }, + { + "epoch": 0.04820997144739732, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 3.9105098247528076, + "learning_rate": 1.6032210834553438e-07, + "loss": 1.1671, + "mean_token_accuracy": 0.6748892068862915, + "num_tokens": 11360900.0, + "step": 439 + }, + { + "epoch": 0.048319789150010985, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.126638889312744, + "learning_rate": 1.6068814055636895e-07, + "loss": 1.2803, + "mean_token_accuracy": 0.6362018585205078, + "num_tokens": 11391599.0, + "step": 440 + }, + { + "epoch": 0.048429606852624645, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 4.0376482009887695, + "learning_rate": 1.610541727672035e-07, + "loss": 1.2057, + "mean_token_accuracy": 0.658727765083313, + "num_tokens": 11424019.0, + "step": 441 + }, + { + "epoch": 0.048539424555238304, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 4.654540538787842, + "learning_rate": 1.6142020497803808e-07, + "loss": 1.1771, + "mean_token_accuracy": 0.6569749116897583, + "num_tokens": 11446707.0, + "step": 442 + }, + { + "epoch": 0.048649242257851964, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 5.275589942932129, + "learning_rate": 1.617862371888726e-07, + "loss": 1.1339, + "mean_token_accuracy": 0.6704713106155396, + "num_tokens": 11467310.0, + "step": 443 + }, + { + "epoch": 0.048759059960465624, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 3.9600307941436768, + "learning_rate": 1.6215226939970718e-07, + "loss": 1.1871, + "mean_token_accuracy": 0.6600963473320007, + "num_tokens": 11493973.0, + "step": 444 + }, + { + "epoch": 0.04886887766307929, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 5.119204044342041, + "learning_rate": 1.6251830161054173e-07, + "loss": 1.1261, + "mean_token_accuracy": 0.6844925284385681, + "num_tokens": 11514611.0, + "step": 445 + }, + { + "epoch": 0.04897869536569295, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 5.115761756896973, + "learning_rate": 1.6288433382137627e-07, + "loss": 1.271, + "mean_token_accuracy": 0.6458368301391602, + "num_tokens": 11538158.0, + "step": 446 + }, + { + "epoch": 0.04908851306830661, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.244462966918945, + "learning_rate": 1.6325036603221082e-07, + "loss": 1.1864, + "mean_token_accuracy": 0.6630254983901978, + "num_tokens": 11570295.0, + "step": 447 + }, + { + "epoch": 0.04919833077092027, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.974661350250244, + "learning_rate": 1.6361639824304537e-07, + "loss": 1.1566, + "mean_token_accuracy": 0.6714481115341187, + "num_tokens": 11597780.0, + "step": 448 + }, + { + "epoch": 0.04930814847353393, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.211415767669678, + "learning_rate": 1.6398243045387995e-07, + "loss": 1.2121, + "mean_token_accuracy": 0.6616039872169495, + "num_tokens": 11623199.0, + "step": 449 + }, + { + "epoch": 0.0494179661761476, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.06248140335083, + "learning_rate": 1.643484626647145e-07, + "loss": 1.248, + "mean_token_accuracy": 0.6411899328231812, + "num_tokens": 11651657.0, + "step": 450 + }, + { + "epoch": 0.04952778387876126, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 4.317858695983887, + "learning_rate": 1.6471449487554902e-07, + "loss": 1.1863, + "mean_token_accuracy": 0.6587136387825012, + "num_tokens": 11680381.0, + "step": 451 + }, + { + "epoch": 0.04963760158137492, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.594311714172363, + "learning_rate": 1.650805270863836e-07, + "loss": 1.1449, + "mean_token_accuracy": 0.6705244183540344, + "num_tokens": 11705013.0, + "step": 452 + }, + { + "epoch": 0.04974741928398858, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.216645240783691, + "learning_rate": 1.6544655929721814e-07, + "loss": 1.0859, + "mean_token_accuracy": 0.6861400604248047, + "num_tokens": 11731140.0, + "step": 453 + }, + { + "epoch": 0.049857236986602244, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.100368499755859, + "learning_rate": 1.6581259150805272e-07, + "loss": 1.2941, + "mean_token_accuracy": 0.6354910731315613, + "num_tokens": 11761261.0, + "step": 454 + }, + { + "epoch": 0.0499670546892159, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.233551502227783, + "learning_rate": 1.6617862371888724e-07, + "loss": 1.1656, + "mean_token_accuracy": 0.6739563941955566, + "num_tokens": 11787241.0, + "step": 455 + }, + { + "epoch": 0.05007687239182956, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.357273578643799, + "learning_rate": 1.6654465592972182e-07, + "loss": 1.1991, + "mean_token_accuracy": 0.6599701046943665, + "num_tokens": 11808564.0, + "step": 456 + }, + { + "epoch": 0.05018669009444322, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 3.855468273162842, + "learning_rate": 1.6691068814055637e-07, + "loss": 1.1703, + "mean_token_accuracy": 0.6672635078430176, + "num_tokens": 11835170.0, + "step": 457 + }, + { + "epoch": 0.05029650779705688, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 3.7655115127563477, + "learning_rate": 1.6727672035139092e-07, + "loss": 1.1716, + "mean_token_accuracy": 0.668229341506958, + "num_tokens": 11862838.0, + "step": 458 + }, + { + "epoch": 0.05040632549967055, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 3.971010684967041, + "learning_rate": 1.6764275256222547e-07, + "loss": 1.3027, + "mean_token_accuracy": 0.6406217813491821, + "num_tokens": 11892865.0, + "step": 459 + }, + { + "epoch": 0.05051614320228421, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.956164360046387, + "learning_rate": 1.6800878477306001e-07, + "loss": 1.119, + "mean_token_accuracy": 0.6802177429199219, + "num_tokens": 11913385.0, + "step": 460 + }, + { + "epoch": 0.05062596090489787, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.270848274230957, + "learning_rate": 1.683748169838946e-07, + "loss": 1.2153, + "mean_token_accuracy": 0.6541323661804199, + "num_tokens": 11942433.0, + "step": 461 + }, + { + "epoch": 0.05073577860751153, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 4.6793670654296875, + "learning_rate": 1.6874084919472914e-07, + "loss": 1.1878, + "mean_token_accuracy": 0.6581928133964539, + "num_tokens": 11968602.0, + "step": 462 + }, + { + "epoch": 0.05084559631012519, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 4.559617042541504, + "learning_rate": 1.6910688140556366e-07, + "loss": 1.1756, + "mean_token_accuracy": 0.6669316291809082, + "num_tokens": 11989652.0, + "step": 463 + }, + { + "epoch": 0.050955414012738856, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 4.353244781494141, + "learning_rate": 1.6947291361639824e-07, + "loss": 1.2065, + "mean_token_accuracy": 0.6499226093292236, + "num_tokens": 12014428.0, + "step": 464 + }, + { + "epoch": 0.051065231715352516, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 4.144178867340088, + "learning_rate": 1.6983894582723279e-07, + "loss": 1.1062, + "mean_token_accuracy": 0.6789345741271973, + "num_tokens": 12041211.0, + "step": 465 + }, + { + "epoch": 0.051175049417966176, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 3.5100395679473877, + "learning_rate": 1.7020497803806736e-07, + "loss": 1.2354, + "mean_token_accuracy": 0.6608623266220093, + "num_tokens": 12071813.0, + "step": 466 + }, + { + "epoch": 0.051284867120579836, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 4.238401412963867, + "learning_rate": 1.7057101024890188e-07, + "loss": 1.2369, + "mean_token_accuracy": 0.6502400636672974, + "num_tokens": 12099495.0, + "step": 467 + }, + { + "epoch": 0.051394684823193496, + "ewc_loss": 1.9818544387817383e-06, + "grad_norm": 4.196129322052002, + "learning_rate": 1.7093704245973643e-07, + "loss": 1.235, + "mean_token_accuracy": 0.6437828540802002, + "num_tokens": 12125207.0, + "step": 468 + }, + { + "epoch": 0.05150450252580716, + "ewc_loss": 1.9818544387817383e-06, + "grad_norm": 5.210012435913086, + "learning_rate": 1.71303074670571e-07, + "loss": 1.1912, + "mean_token_accuracy": 0.652012288570404, + "num_tokens": 12146398.0, + "step": 469 + }, + { + "epoch": 0.05161432022842082, + "ewc_loss": 1.9818544387817383e-06, + "grad_norm": 3.713099241256714, + "learning_rate": 1.7166910688140556e-07, + "loss": 1.0903, + "mean_token_accuracy": 0.6797963380813599, + "num_tokens": 12170215.0, + "step": 470 + }, + { + "epoch": 0.05172413793103448, + "ewc_loss": 1.9818544387817383e-06, + "grad_norm": 4.975036144256592, + "learning_rate": 1.720351390922401e-07, + "loss": 1.2426, + "mean_token_accuracy": 0.6441497802734375, + "num_tokens": 12192610.0, + "step": 471 + }, + { + "epoch": 0.05183395563364814, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 5.298534870147705, + "learning_rate": 1.7240117130307466e-07, + "loss": 1.1874, + "mean_token_accuracy": 0.6655162572860718, + "num_tokens": 12220138.0, + "step": 472 + }, + { + "epoch": 0.05194377333626181, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 3.850010871887207, + "learning_rate": 1.7276720351390923e-07, + "loss": 1.1943, + "mean_token_accuracy": 0.6609439253807068, + "num_tokens": 12251326.0, + "step": 473 + }, + { + "epoch": 0.05205359103887547, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 3.747896671295166, + "learning_rate": 1.7313323572474378e-07, + "loss": 1.2528, + "mean_token_accuracy": 0.6479406356811523, + "num_tokens": 12280837.0, + "step": 474 + }, + { + "epoch": 0.05216340874148913, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 4.513009071350098, + "learning_rate": 1.734992679355783e-07, + "loss": 1.0743, + "mean_token_accuracy": 0.6899221539497375, + "num_tokens": 12304158.0, + "step": 475 + }, + { + "epoch": 0.05227322644410279, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 3.955232620239258, + "learning_rate": 1.7386530014641288e-07, + "loss": 1.2499, + "mean_token_accuracy": 0.6435796618461609, + "num_tokens": 12331193.0, + "step": 476 + }, + { + "epoch": 0.05238304414671645, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 3.9656882286071777, + "learning_rate": 1.7423133235724743e-07, + "loss": 1.2308, + "mean_token_accuracy": 0.6534770727157593, + "num_tokens": 12357251.0, + "step": 477 + }, + { + "epoch": 0.052492861849330115, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 3.7584173679351807, + "learning_rate": 1.74597364568082e-07, + "loss": 1.2452, + "mean_token_accuracy": 0.6471052169799805, + "num_tokens": 12384863.0, + "step": 478 + }, + { + "epoch": 0.052602679551943775, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 3.5429399013519287, + "learning_rate": 1.7496339677891653e-07, + "loss": 1.1049, + "mean_token_accuracy": 0.6800698041915894, + "num_tokens": 12415065.0, + "step": 479 + }, + { + "epoch": 0.052712497254557435, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 4.526996612548828, + "learning_rate": 1.7532942898975108e-07, + "loss": 1.1188, + "mean_token_accuracy": 0.6751502156257629, + "num_tokens": 12439060.0, + "step": 480 + }, + { + "epoch": 0.052822314957171095, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 3.9698147773742676, + "learning_rate": 1.7569546120058565e-07, + "loss": 1.1266, + "mean_token_accuracy": 0.6703639626502991, + "num_tokens": 12462856.0, + "step": 481 + }, + { + "epoch": 0.052932132659784754, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 4.043781280517578, + "learning_rate": 1.760614934114202e-07, + "loss": 1.0932, + "mean_token_accuracy": 0.6806239485740662, + "num_tokens": 12487168.0, + "step": 482 + }, + { + "epoch": 0.05304195036239842, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 4.282992839813232, + "learning_rate": 1.7642752562225475e-07, + "loss": 1.2512, + "mean_token_accuracy": 0.643398642539978, + "num_tokens": 12516282.0, + "step": 483 + }, + { + "epoch": 0.05315176806501208, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 4.375101566314697, + "learning_rate": 1.767935578330893e-07, + "loss": 1.1118, + "mean_token_accuracy": 0.6782684922218323, + "num_tokens": 12541665.0, + "step": 484 + }, + { + "epoch": 0.05326158576762574, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 3.8619017601013184, + "learning_rate": 1.7715959004392387e-07, + "loss": 1.2413, + "mean_token_accuracy": 0.6544163227081299, + "num_tokens": 12573589.0, + "step": 485 + }, + { + "epoch": 0.0533714034702394, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 3.6421804428100586, + "learning_rate": 1.7752562225475842e-07, + "loss": 1.1352, + "mean_token_accuracy": 0.6727409362792969, + "num_tokens": 12601106.0, + "step": 486 + }, + { + "epoch": 0.05348122117285306, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 3.515551805496216, + "learning_rate": 1.7789165446559295e-07, + "loss": 1.1587, + "mean_token_accuracy": 0.6700510382652283, + "num_tokens": 12631770.0, + "step": 487 + }, + { + "epoch": 0.05359103887546673, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 4.104073524475098, + "learning_rate": 1.7825768667642752e-07, + "loss": 1.1427, + "mean_token_accuracy": 0.6622171401977539, + "num_tokens": 12655234.0, + "step": 488 + }, + { + "epoch": 0.05370085657808039, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.696122646331787, + "learning_rate": 1.7862371888726207e-07, + "loss": 1.192, + "mean_token_accuracy": 0.6597285866737366, + "num_tokens": 12682179.0, + "step": 489 + }, + { + "epoch": 0.05381067428069405, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.9385170936584473, + "learning_rate": 1.7898975109809665e-07, + "loss": 1.1529, + "mean_token_accuracy": 0.6675717830657959, + "num_tokens": 12711843.0, + "step": 490 + }, + { + "epoch": 0.05392049198330771, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 4.079559803009033, + "learning_rate": 1.7935578330893117e-07, + "loss": 1.1952, + "mean_token_accuracy": 0.656440258026123, + "num_tokens": 12741567.0, + "step": 491 + }, + { + "epoch": 0.054030309685921374, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.925924301147461, + "learning_rate": 1.7972181551976572e-07, + "loss": 1.1947, + "mean_token_accuracy": 0.6619614958763123, + "num_tokens": 12767598.0, + "step": 492 + }, + { + "epoch": 0.054140127388535034, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 4.234228610992432, + "learning_rate": 1.800878477306003e-07, + "loss": 1.1486, + "mean_token_accuracy": 0.6738371849060059, + "num_tokens": 12790026.0, + "step": 493 + }, + { + "epoch": 0.054249945091148694, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.5109026432037354, + "learning_rate": 1.8045387994143484e-07, + "loss": 1.1542, + "mean_token_accuracy": 0.6731902956962585, + "num_tokens": 12818197.0, + "step": 494 + }, + { + "epoch": 0.05435976279376235, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 4.371522903442383, + "learning_rate": 1.808199121522694e-07, + "loss": 1.1636, + "mean_token_accuracy": 0.6644998788833618, + "num_tokens": 12840299.0, + "step": 495 + }, + { + "epoch": 0.05446958049637601, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 4.461458206176758, + "learning_rate": 1.8118594436310394e-07, + "loss": 1.2403, + "mean_token_accuracy": 0.6457723379135132, + "num_tokens": 12865428.0, + "step": 496 + }, + { + "epoch": 0.05457939819898968, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 3.1207895278930664, + "learning_rate": 1.8155197657393852e-07, + "loss": 1.1544, + "mean_token_accuracy": 0.6748308539390564, + "num_tokens": 12892978.0, + "step": 497 + }, + { + "epoch": 0.05468921590160334, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 3.5414559841156006, + "learning_rate": 1.8191800878477307e-07, + "loss": 1.2002, + "mean_token_accuracy": 0.6674566268920898, + "num_tokens": 12921515.0, + "step": 498 + }, + { + "epoch": 0.054799033604217, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 4.10932731628418, + "learning_rate": 1.822840409956076e-07, + "loss": 1.1832, + "mean_token_accuracy": 0.6641497015953064, + "num_tokens": 12947231.0, + "step": 499 + }, + { + "epoch": 0.05490885130683066, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.9454195499420166, + "learning_rate": 1.8265007320644216e-07, + "loss": 1.193, + "mean_token_accuracy": 0.657302737236023, + "num_tokens": 12975796.0, + "step": 500 + }, + { + "epoch": 0.05501866900944432, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 3.437746286392212, + "learning_rate": 1.830161054172767e-07, + "loss": 1.1298, + "mean_token_accuracy": 0.6777533888816833, + "num_tokens": 13005740.0, + "step": 501 + }, + { + "epoch": 0.055128486712057986, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 3.9849915504455566, + "learning_rate": 1.833821376281113e-07, + "loss": 1.1639, + "mean_token_accuracy": 0.6595414876937866, + "num_tokens": 13029608.0, + "step": 502 + }, + { + "epoch": 0.055238304414671646, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 4.309621334075928, + "learning_rate": 1.837481698389458e-07, + "loss": 1.0725, + "mean_token_accuracy": 0.6899320483207703, + "num_tokens": 13050822.0, + "step": 503 + }, + { + "epoch": 0.055348122117285306, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 4.333123683929443, + "learning_rate": 1.8411420204978036e-07, + "loss": 0.9969, + "mean_token_accuracy": 0.7045212984085083, + "num_tokens": 13076150.0, + "step": 504 + }, + { + "epoch": 0.055457939819898966, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 3.5887258052825928, + "learning_rate": 1.8448023426061494e-07, + "loss": 1.2145, + "mean_token_accuracy": 0.6519338488578796, + "num_tokens": 13103581.0, + "step": 505 + }, + { + "epoch": 0.055567757522512626, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 3.524449586868286, + "learning_rate": 1.8484626647144948e-07, + "loss": 1.1466, + "mean_token_accuracy": 0.6627981066703796, + "num_tokens": 13131904.0, + "step": 506 + }, + { + "epoch": 0.05567757522512629, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 3.701289415359497, + "learning_rate": 1.8521229868228403e-07, + "loss": 1.0532, + "mean_token_accuracy": 0.6907854080200195, + "num_tokens": 13155934.0, + "step": 507 + }, + { + "epoch": 0.05578739292773995, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 3.748992443084717, + "learning_rate": 1.8557833089311858e-07, + "loss": 1.11, + "mean_token_accuracy": 0.6768948435783386, + "num_tokens": 13179650.0, + "step": 508 + }, + { + "epoch": 0.05589721063035361, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 3.808011054992676, + "learning_rate": 1.8594436310395313e-07, + "loss": 1.201, + "mean_token_accuracy": 0.6541142463684082, + "num_tokens": 13205174.0, + "step": 509 + }, + { + "epoch": 0.05600702833296727, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 3.6028194427490234, + "learning_rate": 1.863103953147877e-07, + "loss": 1.1428, + "mean_token_accuracy": 0.6657320261001587, + "num_tokens": 13231544.0, + "step": 510 + }, + { + "epoch": 0.05611684603558094, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 4.66948127746582, + "learning_rate": 1.8667642752562223e-07, + "loss": 1.1725, + "mean_token_accuracy": 0.6608936786651611, + "num_tokens": 13252845.0, + "step": 511 + }, + { + "epoch": 0.0562266637381946, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 3.4522554874420166, + "learning_rate": 1.870424597364568e-07, + "loss": 1.2661, + "mean_token_accuracy": 0.6383472084999084, + "num_tokens": 13285009.0, + "step": 512 + }, + { + "epoch": 0.05633648144080826, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 3.481654644012451, + "learning_rate": 1.8740849194729135e-07, + "loss": 1.0853, + "mean_token_accuracy": 0.6881771087646484, + "num_tokens": 13312518.0, + "step": 513 + }, + { + "epoch": 0.05644629914342192, + "ewc_loss": 2.205371856689453e-06, + "grad_norm": 3.484863519668579, + "learning_rate": 1.8777452415812593e-07, + "loss": 1.0849, + "mean_token_accuracy": 0.6880935430526733, + "num_tokens": 13335578.0, + "step": 514 + }, + { + "epoch": 0.05655611684603558, + "ewc_loss": 2.205371856689453e-06, + "grad_norm": 3.0803449153900146, + "learning_rate": 1.8814055636896045e-07, + "loss": 1.1907, + "mean_token_accuracy": 0.6661964058876038, + "num_tokens": 13368449.0, + "step": 515 + }, + { + "epoch": 0.056665934548649245, + "ewc_loss": 2.205371856689453e-06, + "grad_norm": 3.748828887939453, + "learning_rate": 1.88506588579795e-07, + "loss": 1.2199, + "mean_token_accuracy": 0.6541237831115723, + "num_tokens": 13397250.0, + "step": 516 + }, + { + "epoch": 0.056775752251262905, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 4.09463357925415, + "learning_rate": 1.8887262079062958e-07, + "loss": 1.1608, + "mean_token_accuracy": 0.658206582069397, + "num_tokens": 13422857.0, + "step": 517 + }, + { + "epoch": 0.056885569953876565, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.4133055210113525, + "learning_rate": 1.8923865300146413e-07, + "loss": 1.086, + "mean_token_accuracy": 0.6772743463516235, + "num_tokens": 13447180.0, + "step": 518 + }, + { + "epoch": 0.056995387656490225, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 4.021984100341797, + "learning_rate": 1.8960468521229868e-07, + "loss": 1.2382, + "mean_token_accuracy": 0.6407957077026367, + "num_tokens": 13469860.0, + "step": 519 + }, + { + "epoch": 0.057105205359103885, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.5965516567230225, + "learning_rate": 1.8997071742313322e-07, + "loss": 1.1858, + "mean_token_accuracy": 0.6684839725494385, + "num_tokens": 13495606.0, + "step": 520 + }, + { + "epoch": 0.05721502306171755, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.3231518268585205, + "learning_rate": 1.9033674963396777e-07, + "loss": 1.1671, + "mean_token_accuracy": 0.6651613116264343, + "num_tokens": 13527790.0, + "step": 521 + }, + { + "epoch": 0.05732484076433121, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.51886248588562, + "learning_rate": 1.9070278184480235e-07, + "loss": 1.1489, + "mean_token_accuracy": 0.679713249206543, + "num_tokens": 13551986.0, + "step": 522 + }, + { + "epoch": 0.05743465846694487, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.427412271499634, + "learning_rate": 1.9106881405563687e-07, + "loss": 1.195, + "mean_token_accuracy": 0.6573151350021362, + "num_tokens": 13577954.0, + "step": 523 + }, + { + "epoch": 0.05754447616955853, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.4791438579559326, + "learning_rate": 1.9143484626647145e-07, + "loss": 1.1947, + "mean_token_accuracy": 0.6534801125526428, + "num_tokens": 13604870.0, + "step": 524 + }, + { + "epoch": 0.05765429387217219, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.7915754318237305, + "learning_rate": 1.91800878477306e-07, + "loss": 1.2108, + "mean_token_accuracy": 0.6597251892089844, + "num_tokens": 13637200.0, + "step": 525 + }, + { + "epoch": 0.05776411157478586, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 5.183077335357666, + "learning_rate": 1.9216691068814057e-07, + "loss": 1.2042, + "mean_token_accuracy": 0.6511518955230713, + "num_tokens": 13656718.0, + "step": 526 + }, + { + "epoch": 0.05787392927739952, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 4.343263149261475, + "learning_rate": 1.925329428989751e-07, + "loss": 1.1185, + "mean_token_accuracy": 0.6748745441436768, + "num_tokens": 13680498.0, + "step": 527 + }, + { + "epoch": 0.05798374698001318, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.9439635276794434, + "learning_rate": 1.9289897510980964e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.6860435009002686, + "num_tokens": 13701339.0, + "step": 528 + }, + { + "epoch": 0.05809356468262684, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.7802977561950684, + "learning_rate": 1.9326500732064422e-07, + "loss": 1.1728, + "mean_token_accuracy": 0.6644089221954346, + "num_tokens": 13724927.0, + "step": 529 + }, + { + "epoch": 0.058203382385240504, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 4.864259719848633, + "learning_rate": 1.9363103953147877e-07, + "loss": 1.079, + "mean_token_accuracy": 0.6867203116416931, + "num_tokens": 13745791.0, + "step": 530 + }, + { + "epoch": 0.058313200087854164, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 3.7283055782318115, + "learning_rate": 1.9399707174231332e-07, + "loss": 1.0977, + "mean_token_accuracy": 0.6784430742263794, + "num_tokens": 13770891.0, + "step": 531 + }, + { + "epoch": 0.058423017790467824, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 4.130971431732178, + "learning_rate": 1.9436310395314787e-07, + "loss": 1.1225, + "mean_token_accuracy": 0.6674481630325317, + "num_tokens": 13790625.0, + "step": 532 + }, + { + "epoch": 0.058532835493081484, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 3.7423813343048096, + "learning_rate": 1.9472913616398242e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6800126433372498, + "num_tokens": 13813346.0, + "step": 533 + }, + { + "epoch": 0.058642653195695144, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 3.5685267448425293, + "learning_rate": 1.95095168374817e-07, + "loss": 1.2014, + "mean_token_accuracy": 0.6558937430381775, + "num_tokens": 13837250.0, + "step": 534 + }, + { + "epoch": 0.05875247089830881, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 3.708714485168457, + "learning_rate": 1.9546120058565151e-07, + "loss": 1.1262, + "mean_token_accuracy": 0.6780138611793518, + "num_tokens": 13865730.0, + "step": 535 + }, + { + "epoch": 0.05886228860092247, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 3.878647565841675, + "learning_rate": 1.958272327964861e-07, + "loss": 1.1128, + "mean_token_accuracy": 0.6759821176528931, + "num_tokens": 13893215.0, + "step": 536 + }, + { + "epoch": 0.05897210630353613, + "ewc_loss": 2.3096799850463867e-06, + "grad_norm": 3.2797040939331055, + "learning_rate": 1.9619326500732064e-07, + "loss": 1.163, + "mean_token_accuracy": 0.6649113893508911, + "num_tokens": 13919232.0, + "step": 537 + }, + { + "epoch": 0.05908192400614979, + "ewc_loss": 2.3096799850463867e-06, + "grad_norm": 3.213812828063965, + "learning_rate": 1.9655929721815521e-07, + "loss": 1.1697, + "mean_token_accuracy": 0.6649912595748901, + "num_tokens": 13943661.0, + "step": 538 + }, + { + "epoch": 0.05919174170876345, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 4.065835952758789, + "learning_rate": 1.9692532942898974e-07, + "loss": 1.1737, + "mean_token_accuracy": 0.6571035385131836, + "num_tokens": 13967209.0, + "step": 539 + }, + { + "epoch": 0.05930155941137712, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 3.9897964000701904, + "learning_rate": 1.9729136163982429e-07, + "loss": 1.0375, + "mean_token_accuracy": 0.6970860958099365, + "num_tokens": 13987948.0, + "step": 540 + }, + { + "epoch": 0.059411377113990776, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 3.1486992835998535, + "learning_rate": 1.9765739385065886e-07, + "loss": 1.2021, + "mean_token_accuracy": 0.6598591804504395, + "num_tokens": 14016862.0, + "step": 541 + }, + { + "epoch": 0.059521194816604436, + "ewc_loss": 2.339482307434082e-06, + "grad_norm": 3.384098768234253, + "learning_rate": 1.980234260614934e-07, + "loss": 1.1858, + "mean_token_accuracy": 0.6640410423278809, + "num_tokens": 14041242.0, + "step": 542 + }, + { + "epoch": 0.059631012519218096, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 3.6954147815704346, + "learning_rate": 1.9838945827232796e-07, + "loss": 1.0883, + "mean_token_accuracy": 0.6851974129676819, + "num_tokens": 14064732.0, + "step": 543 + }, + { + "epoch": 0.059740830221831756, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 3.650726795196533, + "learning_rate": 1.987554904831625e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6869637370109558, + "num_tokens": 14090186.0, + "step": 544 + }, + { + "epoch": 0.05985064792444542, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 3.1342921257019043, + "learning_rate": 1.9912152269399706e-07, + "loss": 1.2151, + "mean_token_accuracy": 0.6550718545913696, + "num_tokens": 14121747.0, + "step": 545 + }, + { + "epoch": 0.05996046562705908, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 3.0052402019500732, + "learning_rate": 1.9948755490483163e-07, + "loss": 1.2258, + "mean_token_accuracy": 0.650588870048523, + "num_tokens": 14152053.0, + "step": 546 + }, + { + "epoch": 0.06007028332967274, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 4.458683013916016, + "learning_rate": 1.9985358711566616e-07, + "loss": 1.1247, + "mean_token_accuracy": 0.6800134181976318, + "num_tokens": 14170890.0, + "step": 547 + }, + { + "epoch": 0.0601801010322864, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.9502532482147217, + "learning_rate": 2.0021961932650073e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.6767905354499817, + "num_tokens": 14200001.0, + "step": 548 + }, + { + "epoch": 0.06028991873490007, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 3.425642251968384, + "learning_rate": 2.0058565153733528e-07, + "loss": 1.1532, + "mean_token_accuracy": 0.67006915807724, + "num_tokens": 14225062.0, + "step": 549 + }, + { + "epoch": 0.06039973643751373, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 3.365894079208374, + "learning_rate": 2.0095168374816983e-07, + "loss": 1.1248, + "mean_token_accuracy": 0.6719911098480225, + "num_tokens": 14248169.0, + "step": 550 + }, + { + "epoch": 0.06050955414012739, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 3.4710755348205566, + "learning_rate": 2.0131771595900438e-07, + "loss": 1.146, + "mean_token_accuracy": 0.6759955286979675, + "num_tokens": 14273309.0, + "step": 551 + }, + { + "epoch": 0.06061937184274105, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 3.5632541179656982, + "learning_rate": 2.0168374816983893e-07, + "loss": 1.1057, + "mean_token_accuracy": 0.6809133291244507, + "num_tokens": 14300494.0, + "step": 552 + }, + { + "epoch": 0.06072918954535471, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 3.247516632080078, + "learning_rate": 2.020497803806735e-07, + "loss": 1.1406, + "mean_token_accuracy": 0.6656294465065002, + "num_tokens": 14327576.0, + "step": 553 + }, + { + "epoch": 0.060839007247968376, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.7417945861816406, + "learning_rate": 2.0241581259150805e-07, + "loss": 1.1639, + "mean_token_accuracy": 0.6608722805976868, + "num_tokens": 14359413.0, + "step": 554 + }, + { + "epoch": 0.060948824950582035, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 3.113405227661133, + "learning_rate": 2.027818448023426e-07, + "loss": 1.2827, + "mean_token_accuracy": 0.6497783064842224, + "num_tokens": 14389772.0, + "step": 555 + }, + { + "epoch": 0.061058642653195695, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 3.493290901184082, + "learning_rate": 2.0314787701317715e-07, + "loss": 1.1978, + "mean_token_accuracy": 0.6626379489898682, + "num_tokens": 14416775.0, + "step": 556 + }, + { + "epoch": 0.061168460355809355, + "ewc_loss": 2.4437904357910156e-06, + "grad_norm": 3.5299904346466064, + "learning_rate": 2.035139092240117e-07, + "loss": 1.2374, + "mean_token_accuracy": 0.6436784863471985, + "num_tokens": 14445547.0, + "step": 557 + }, + { + "epoch": 0.061278278058423015, + "ewc_loss": 2.4437904357910156e-06, + "grad_norm": 3.5852761268615723, + "learning_rate": 2.0387994143484627e-07, + "loss": 1.2094, + "mean_token_accuracy": 0.6510529518127441, + "num_tokens": 14469951.0, + "step": 558 + }, + { + "epoch": 0.06138809576103668, + "ewc_loss": 2.4437904357910156e-06, + "grad_norm": 3.368460178375244, + "learning_rate": 2.042459736456808e-07, + "loss": 1.1696, + "mean_token_accuracy": 0.6665197610855103, + "num_tokens": 14492322.0, + "step": 559 + }, + { + "epoch": 0.06149791346365034, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 3.0389859676361084, + "learning_rate": 2.0461200585651537e-07, + "loss": 1.1604, + "mean_token_accuracy": 0.6670112013816833, + "num_tokens": 14517881.0, + "step": 560 + }, + { + "epoch": 0.061607731166264, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 3.8658359050750732, + "learning_rate": 2.0497803806734992e-07, + "loss": 1.1898, + "mean_token_accuracy": 0.6641483306884766, + "num_tokens": 14543018.0, + "step": 561 + }, + { + "epoch": 0.06171754886887766, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 3.735053300857544, + "learning_rate": 2.0534407027818447e-07, + "loss": 1.1702, + "mean_token_accuracy": 0.6598902940750122, + "num_tokens": 14570418.0, + "step": 562 + }, + { + "epoch": 0.06182736657149132, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 3.748450756072998, + "learning_rate": 2.0571010248901902e-07, + "loss": 1.0265, + "mean_token_accuracy": 0.6956539154052734, + "num_tokens": 14589713.0, + "step": 563 + }, + { + "epoch": 0.06193718427410499, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.0408647060394287, + "learning_rate": 2.0607613469985357e-07, + "loss": 1.1701, + "mean_token_accuracy": 0.6623352766036987, + "num_tokens": 14617418.0, + "step": 564 + }, + { + "epoch": 0.06204700197671865, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 3.3365821838378906, + "learning_rate": 2.0644216691068814e-07, + "loss": 1.142, + "mean_token_accuracy": 0.675154983997345, + "num_tokens": 14645074.0, + "step": 565 + }, + { + "epoch": 0.06215681967933231, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.1264357566833496, + "learning_rate": 2.068081991215227e-07, + "loss": 1.2154, + "mean_token_accuracy": 0.6551796793937683, + "num_tokens": 14671552.0, + "step": 566 + }, + { + "epoch": 0.06226663738194597, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.4631645679473877, + "learning_rate": 2.0717423133235724e-07, + "loss": 1.1372, + "mean_token_accuracy": 0.6712393760681152, + "num_tokens": 14694223.0, + "step": 567 + }, + { + "epoch": 0.062376455084559634, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.62479305267334, + "learning_rate": 2.075402635431918e-07, + "loss": 1.0836, + "mean_token_accuracy": 0.6899666786193848, + "num_tokens": 14719184.0, + "step": 568 + }, + { + "epoch": 0.062486272787173294, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.1330795288085938, + "learning_rate": 2.0790629575402634e-07, + "loss": 1.1424, + "mean_token_accuracy": 0.6717578172683716, + "num_tokens": 14742235.0, + "step": 569 + }, + { + "epoch": 0.06259609048978695, + "ewc_loss": 2.4884939193725586e-06, + "grad_norm": 3.2669906616210938, + "learning_rate": 2.0827232796486092e-07, + "loss": 1.2203, + "mean_token_accuracy": 0.6525669097900391, + "num_tokens": 14769348.0, + "step": 570 + }, + { + "epoch": 0.06270590819240061, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 3.2663466930389404, + "learning_rate": 2.0863836017569544e-07, + "loss": 1.0509, + "mean_token_accuracy": 0.6920151710510254, + "num_tokens": 14794555.0, + "step": 571 + }, + { + "epoch": 0.06281572589501427, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 3.302503824234009, + "learning_rate": 2.0900439238653001e-07, + "loss": 1.1405, + "mean_token_accuracy": 0.6700888872146606, + "num_tokens": 14817364.0, + "step": 572 + }, + { + "epoch": 0.06292554359762793, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 3.605518102645874, + "learning_rate": 2.0937042459736456e-07, + "loss": 1.1734, + "mean_token_accuracy": 0.6685770750045776, + "num_tokens": 14837472.0, + "step": 573 + }, + { + "epoch": 0.0630353613002416, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.8576700687408447, + "learning_rate": 2.097364568081991e-07, + "loss": 1.1162, + "mean_token_accuracy": 0.6787682175636292, + "num_tokens": 14865112.0, + "step": 574 + }, + { + "epoch": 0.06314517900285525, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 3.191875696182251, + "learning_rate": 2.1010248901903366e-07, + "loss": 1.1525, + "mean_token_accuracy": 0.6633908748626709, + "num_tokens": 14894808.0, + "step": 575 + }, + { + "epoch": 0.06325499670546893, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 3.687798261642456, + "learning_rate": 2.104685212298682e-07, + "loss": 1.0442, + "mean_token_accuracy": 0.6959407329559326, + "num_tokens": 14913195.0, + "step": 576 + }, + { + "epoch": 0.06336481440808259, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.9347269535064697, + "learning_rate": 2.1083455344070279e-07, + "loss": 1.1894, + "mean_token_accuracy": 0.6601865887641907, + "num_tokens": 14943929.0, + "step": 577 + }, + { + "epoch": 0.06347463211069625, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 3.1385576725006104, + "learning_rate": 2.1120058565153734e-07, + "loss": 1.1709, + "mean_token_accuracy": 0.6648850440979004, + "num_tokens": 14969696.0, + "step": 578 + }, + { + "epoch": 0.0635844498133099, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 3.5835182666778564, + "learning_rate": 2.1156661786237186e-07, + "loss": 1.0127, + "mean_token_accuracy": 0.6967551708221436, + "num_tokens": 14989320.0, + "step": 579 + }, + { + "epoch": 0.06369426751592357, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.987428903579712, + "learning_rate": 2.1193265007320643e-07, + "loss": 1.0777, + "mean_token_accuracy": 0.6836534738540649, + "num_tokens": 15014145.0, + "step": 580 + }, + { + "epoch": 0.06380408521853723, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.7617948055267334, + "learning_rate": 2.1229868228404098e-07, + "loss": 1.157, + "mean_token_accuracy": 0.6617491245269775, + "num_tokens": 15040827.0, + "step": 581 + }, + { + "epoch": 0.06391390292115089, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.702993154525757, + "learning_rate": 2.1266471449487556e-07, + "loss": 1.2053, + "mean_token_accuracy": 0.6585469841957092, + "num_tokens": 15069707.0, + "step": 582 + }, + { + "epoch": 0.06402372062376455, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 3.5941474437713623, + "learning_rate": 2.1303074670571008e-07, + "loss": 1.1427, + "mean_token_accuracy": 0.6679142713546753, + "num_tokens": 15091567.0, + "step": 583 + }, + { + "epoch": 0.0641335383263782, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 3.3321430683135986, + "learning_rate": 2.1339677891654466e-07, + "loss": 1.2179, + "mean_token_accuracy": 0.6526588797569275, + "num_tokens": 15117176.0, + "step": 584 + }, + { + "epoch": 0.06424335602899188, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.5944175720214844, + "learning_rate": 2.137628111273792e-07, + "loss": 1.1607, + "mean_token_accuracy": 0.6678200364112854, + "num_tokens": 15151656.0, + "step": 585 + }, + { + "epoch": 0.06435317373160554, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 3.110474109649658, + "learning_rate": 2.1412884333821375e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6942856907844543, + "num_tokens": 15176031.0, + "step": 586 + }, + { + "epoch": 0.0644629914342192, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 3.299387216567993, + "learning_rate": 2.144948755490483e-07, + "loss": 1.1199, + "mean_token_accuracy": 0.6694384813308716, + "num_tokens": 15197744.0, + "step": 587 + }, + { + "epoch": 0.06457280913683286, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.7160675525665283, + "learning_rate": 2.1486090775988285e-07, + "loss": 1.0988, + "mean_token_accuracy": 0.680168092250824, + "num_tokens": 15219377.0, + "step": 588 + }, + { + "epoch": 0.06468262683944652, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.0631520748138428, + "learning_rate": 2.1522693997071743e-07, + "loss": 1.1128, + "mean_token_accuracy": 0.6756501793861389, + "num_tokens": 15244196.0, + "step": 589 + }, + { + "epoch": 0.06479244454206018, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.960157871246338, + "learning_rate": 2.1559297218155198e-07, + "loss": 1.165, + "mean_token_accuracy": 0.6746006011962891, + "num_tokens": 15273093.0, + "step": 590 + }, + { + "epoch": 0.06490226224467384, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.0634102821350098, + "learning_rate": 2.159590043923865e-07, + "loss": 1.1219, + "mean_token_accuracy": 0.6769947409629822, + "num_tokens": 15299076.0, + "step": 591 + }, + { + "epoch": 0.0650120799472875, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.202039957046509, + "learning_rate": 2.1632503660322108e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.7001637816429138, + "num_tokens": 15323632.0, + "step": 592 + }, + { + "epoch": 0.06512189764990116, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.72428035736084, + "learning_rate": 2.1669106881405563e-07, + "loss": 1.1375, + "mean_token_accuracy": 0.6719533205032349, + "num_tokens": 15356700.0, + "step": 593 + }, + { + "epoch": 0.06523171535251482, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.7534587383270264, + "learning_rate": 2.170571010248902e-07, + "loss": 1.0452, + "mean_token_accuracy": 0.6972594261169434, + "num_tokens": 15382486.0, + "step": 594 + }, + { + "epoch": 0.06534153305512849, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.529531955718994, + "learning_rate": 2.1742313323572472e-07, + "loss": 1.0711, + "mean_token_accuracy": 0.6923235654830933, + "num_tokens": 15403037.0, + "step": 595 + }, + { + "epoch": 0.06545135075774215, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.3325841426849365, + "learning_rate": 2.177891654465593e-07, + "loss": 1.0671, + "mean_token_accuracy": 0.6901249885559082, + "num_tokens": 15427884.0, + "step": 596 + }, + { + "epoch": 0.06556116846035581, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.6617019176483154, + "learning_rate": 2.1815519765739385e-07, + "loss": 1.1403, + "mean_token_accuracy": 0.667095422744751, + "num_tokens": 15461669.0, + "step": 597 + }, + { + "epoch": 0.06567098616296947, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 3.393249750137329, + "learning_rate": 2.185212298682284e-07, + "loss": 1.0819, + "mean_token_accuracy": 0.6795247793197632, + "num_tokens": 15484624.0, + "step": 598 + }, + { + "epoch": 0.06578080386558313, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.9930543899536133, + "learning_rate": 2.1888726207906295e-07, + "loss": 1.1776, + "mean_token_accuracy": 0.670164167881012, + "num_tokens": 15508390.0, + "step": 599 + }, + { + "epoch": 0.06589062156819679, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.9728775024414062, + "learning_rate": 2.192532942898975e-07, + "loss": 1.1043, + "mean_token_accuracy": 0.6792255640029907, + "num_tokens": 15535744.0, + "step": 600 + }, + { + "epoch": 0.06600043927081045, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.8046350479125977, + "learning_rate": 2.1961932650073207e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.679145336151123, + "num_tokens": 15565036.0, + "step": 601 + }, + { + "epoch": 0.06611025697342411, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.426377058029175, + "learning_rate": 2.1998535871156662e-07, + "loss": 1.2022, + "mean_token_accuracy": 0.6580361723899841, + "num_tokens": 15597806.0, + "step": 602 + }, + { + "epoch": 0.06622007467603777, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 3.1255743503570557, + "learning_rate": 2.2035139092240114e-07, + "loss": 1.1158, + "mean_token_accuracy": 0.6728607416152954, + "num_tokens": 15623941.0, + "step": 603 + }, + { + "epoch": 0.06632989237865144, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 3.3045454025268555, + "learning_rate": 2.2071742313323572e-07, + "loss": 1.1549, + "mean_token_accuracy": 0.6617885231971741, + "num_tokens": 15649843.0, + "step": 604 + }, + { + "epoch": 0.0664397100812651, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.807548761367798, + "learning_rate": 2.2108345534407027e-07, + "loss": 1.049, + "mean_token_accuracy": 0.6895259618759155, + "num_tokens": 15677720.0, + "step": 605 + }, + { + "epoch": 0.06654952778387876, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 3.03367018699646, + "learning_rate": 2.2144948755490484e-07, + "loss": 1.059, + "mean_token_accuracy": 0.6884419322013855, + "num_tokens": 15704251.0, + "step": 606 + }, + { + "epoch": 0.06665934548649242, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 3.238043785095215, + "learning_rate": 2.2181551976573937e-07, + "loss": 1.1383, + "mean_token_accuracy": 0.6676411628723145, + "num_tokens": 15729100.0, + "step": 607 + }, + { + "epoch": 0.06676916318910608, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 3.18061900138855, + "learning_rate": 2.2218155197657394e-07, + "loss": 1.1648, + "mean_token_accuracy": 0.6561328172683716, + "num_tokens": 15753408.0, + "step": 608 + }, + { + "epoch": 0.06687898089171974, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 3.100020170211792, + "learning_rate": 2.225475841874085e-07, + "loss": 1.0786, + "mean_token_accuracy": 0.6916307806968689, + "num_tokens": 15777400.0, + "step": 609 + }, + { + "epoch": 0.0669887985943334, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.89501690864563, + "learning_rate": 2.2291361639824304e-07, + "loss": 1.1926, + "mean_token_accuracy": 0.6633914709091187, + "num_tokens": 15808367.0, + "step": 610 + }, + { + "epoch": 0.06709861629694706, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 3.136356830596924, + "learning_rate": 2.232796486090776e-07, + "loss": 1.0215, + "mean_token_accuracy": 0.7068818211555481, + "num_tokens": 15831105.0, + "step": 611 + }, + { + "epoch": 0.06720843399956072, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 3.0370373725891113, + "learning_rate": 2.2364568081991214e-07, + "loss": 1.1307, + "mean_token_accuracy": 0.673744797706604, + "num_tokens": 15854816.0, + "step": 612 + }, + { + "epoch": 0.06731825170217438, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 3.410865306854248, + "learning_rate": 2.240117130307467e-07, + "loss": 1.1181, + "mean_token_accuracy": 0.6738482117652893, + "num_tokens": 15873146.0, + "step": 613 + }, + { + "epoch": 0.06742806940478806, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.746913194656372, + "learning_rate": 2.2437774524158126e-07, + "loss": 1.1667, + "mean_token_accuracy": 0.669766902923584, + "num_tokens": 15898651.0, + "step": 614 + }, + { + "epoch": 0.06753788710740172, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 3.0717520713806152, + "learning_rate": 2.2474377745241578e-07, + "loss": 1.1649, + "mean_token_accuracy": 0.6566926836967468, + "num_tokens": 15921202.0, + "step": 615 + }, + { + "epoch": 0.06764770481001538, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 3.1748790740966797, + "learning_rate": 2.2510980966325036e-07, + "loss": 1.015, + "mean_token_accuracy": 0.6961795091629028, + "num_tokens": 15940488.0, + "step": 616 + }, + { + "epoch": 0.06775752251262904, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 3.0332884788513184, + "learning_rate": 2.254758418740849e-07, + "loss": 1.1003, + "mean_token_accuracy": 0.6843134760856628, + "num_tokens": 15967477.0, + "step": 617 + }, + { + "epoch": 0.0678673402152427, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.7794930934906006, + "learning_rate": 2.2584187408491948e-07, + "loss": 1.1156, + "mean_token_accuracy": 0.674743115901947, + "num_tokens": 15995250.0, + "step": 618 + }, + { + "epoch": 0.06797715791785636, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.859501838684082, + "learning_rate": 2.26207906295754e-07, + "loss": 1.2263, + "mean_token_accuracy": 0.6547420024871826, + "num_tokens": 16020883.0, + "step": 619 + }, + { + "epoch": 0.06808697562047002, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.7335431575775146, + "learning_rate": 2.2657393850658856e-07, + "loss": 1.2407, + "mean_token_accuracy": 0.649570882320404, + "num_tokens": 16047600.0, + "step": 620 + }, + { + "epoch": 0.06819679332308368, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 3.1583147048950195, + "learning_rate": 2.2693997071742313e-07, + "loss": 1.1318, + "mean_token_accuracy": 0.6715865135192871, + "num_tokens": 16070115.0, + "step": 621 + }, + { + "epoch": 0.06830661102569734, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.820842742919922, + "learning_rate": 2.2730600292825768e-07, + "loss": 1.1615, + "mean_token_accuracy": 0.6689437627792358, + "num_tokens": 16096102.0, + "step": 622 + }, + { + "epoch": 0.06841642872831101, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.8823795318603516, + "learning_rate": 2.2767203513909223e-07, + "loss": 1.1967, + "mean_token_accuracy": 0.6622527241706848, + "num_tokens": 16122671.0, + "step": 623 + }, + { + "epoch": 0.06852624643092467, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.8827033042907715, + "learning_rate": 2.2803806734992678e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6870512962341309, + "num_tokens": 16149145.0, + "step": 624 + }, + { + "epoch": 0.06863606413353833, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.552290439605713, + "learning_rate": 2.2840409956076135e-07, + "loss": 1.1313, + "mean_token_accuracy": 0.6818734407424927, + "num_tokens": 16178336.0, + "step": 625 + }, + { + "epoch": 0.06874588183615199, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 3.178760051727295, + "learning_rate": 2.287701317715959e-07, + "loss": 1.1339, + "mean_token_accuracy": 0.6744726300239563, + "num_tokens": 16200808.0, + "step": 626 + }, + { + "epoch": 0.06885569953876565, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.6842079162597656, + "learning_rate": 2.2913616398243043e-07, + "loss": 1.2544, + "mean_token_accuracy": 0.6424664855003357, + "num_tokens": 16228780.0, + "step": 627 + }, + { + "epoch": 0.06896551724137931, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.966001510620117, + "learning_rate": 2.29502196193265e-07, + "loss": 1.1888, + "mean_token_accuracy": 0.6523909568786621, + "num_tokens": 16256783.0, + "step": 628 + }, + { + "epoch": 0.06907533494399297, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.798992395401001, + "learning_rate": 2.2986822840409955e-07, + "loss": 1.1635, + "mean_token_accuracy": 0.658286988735199, + "num_tokens": 16281018.0, + "step": 629 + }, + { + "epoch": 0.06918515264660663, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 3.007223129272461, + "learning_rate": 2.3023426061493413e-07, + "loss": 1.0585, + "mean_token_accuracy": 0.688757061958313, + "num_tokens": 16306575.0, + "step": 630 + }, + { + "epoch": 0.06929497034922029, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.585773468017578, + "learning_rate": 2.3060029282576865e-07, + "loss": 1.1955, + "mean_token_accuracy": 0.6565602421760559, + "num_tokens": 16337454.0, + "step": 631 + }, + { + "epoch": 0.06940478805183395, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 3.111231803894043, + "learning_rate": 2.309663250366032e-07, + "loss": 1.1328, + "mean_token_accuracy": 0.6654672622680664, + "num_tokens": 16360648.0, + "step": 632 + }, + { + "epoch": 0.06951460575444762, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.7138352394104004, + "learning_rate": 2.3133235724743777e-07, + "loss": 1.1315, + "mean_token_accuracy": 0.6700668931007385, + "num_tokens": 16389325.0, + "step": 633 + }, + { + "epoch": 0.06962442345706128, + "ewc_loss": 2.771615982055664e-06, + "grad_norm": 3.2558813095092773, + "learning_rate": 2.3169838945827232e-07, + "loss": 1.1631, + "mean_token_accuracy": 0.6670319437980652, + "num_tokens": 16412500.0, + "step": 634 + }, + { + "epoch": 0.06973424115967494, + "ewc_loss": 2.771615982055664e-06, + "grad_norm": 2.3284881114959717, + "learning_rate": 2.3206442166910687e-07, + "loss": 1.2101, + "mean_token_accuracy": 0.6468881368637085, + "num_tokens": 16449528.0, + "step": 635 + }, + { + "epoch": 0.0698440588622886, + "ewc_loss": 2.771615982055664e-06, + "grad_norm": 2.6391966342926025, + "learning_rate": 2.3243045387994142e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.7051090002059937, + "num_tokens": 16475464.0, + "step": 636 + }, + { + "epoch": 0.06995387656490226, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.6903634071350098, + "learning_rate": 2.32796486090776e-07, + "loss": 1.1206, + "mean_token_accuracy": 0.6729589700698853, + "num_tokens": 16503191.0, + "step": 637 + }, + { + "epoch": 0.07006369426751592, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.716987371444702, + "learning_rate": 2.3316251830161055e-07, + "loss": 1.1448, + "mean_token_accuracy": 0.6647812128067017, + "num_tokens": 16532650.0, + "step": 638 + }, + { + "epoch": 0.07017351197012958, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.743025064468384, + "learning_rate": 2.3352855051244507e-07, + "loss": 1.2168, + "mean_token_accuracy": 0.658327043056488, + "num_tokens": 16560255.0, + "step": 639 + }, + { + "epoch": 0.07028332967274324, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.458547353744507, + "learning_rate": 2.3389458272327964e-07, + "loss": 1.1736, + "mean_token_accuracy": 0.6586856842041016, + "num_tokens": 16591859.0, + "step": 640 + }, + { + "epoch": 0.0703931473753569, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.5968194007873535, + "learning_rate": 2.342606149341142e-07, + "loss": 1.1654, + "mean_token_accuracy": 0.6663509607315063, + "num_tokens": 16620782.0, + "step": 641 + }, + { + "epoch": 0.07050296507797058, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.876702070236206, + "learning_rate": 2.3462664714494877e-07, + "loss": 1.1729, + "mean_token_accuracy": 0.654839277267456, + "num_tokens": 16648051.0, + "step": 642 + }, + { + "epoch": 0.07061278278058424, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.6030731201171875, + "learning_rate": 2.349926793557833e-07, + "loss": 1.2091, + "mean_token_accuracy": 0.6580452919006348, + "num_tokens": 16677294.0, + "step": 643 + }, + { + "epoch": 0.0707226004831979, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.6541194915771484, + "learning_rate": 2.3535871156661784e-07, + "loss": 1.2419, + "mean_token_accuracy": 0.6461889743804932, + "num_tokens": 16707122.0, + "step": 644 + }, + { + "epoch": 0.07083241818581155, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.557046890258789, + "learning_rate": 2.3572474377745242e-07, + "loss": 1.1455, + "mean_token_accuracy": 0.6653429865837097, + "num_tokens": 16735039.0, + "step": 645 + }, + { + "epoch": 0.07094223588842521, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 3.160977363586426, + "learning_rate": 2.3609077598828696e-07, + "loss": 1.0868, + "mean_token_accuracy": 0.6867120862007141, + "num_tokens": 16759433.0, + "step": 646 + }, + { + "epoch": 0.07105205359103887, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.6770999431610107, + "learning_rate": 2.3645680819912151e-07, + "loss": 1.0953, + "mean_token_accuracy": 0.6768073439598083, + "num_tokens": 16785559.0, + "step": 647 + }, + { + "epoch": 0.07116187129365253, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 3.185124635696411, + "learning_rate": 2.3682284040995606e-07, + "loss": 1.0793, + "mean_token_accuracy": 0.683904767036438, + "num_tokens": 16805432.0, + "step": 648 + }, + { + "epoch": 0.0712716889962662, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.9612648487091064, + "learning_rate": 2.3718887262079064e-07, + "loss": 1.1015, + "mean_token_accuracy": 0.6798085570335388, + "num_tokens": 16828225.0, + "step": 649 + }, + { + "epoch": 0.07138150669887985, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.9354946613311768, + "learning_rate": 2.375549048316252e-07, + "loss": 1.1396, + "mean_token_accuracy": 0.6762633323669434, + "num_tokens": 16853283.0, + "step": 650 + }, + { + "epoch": 0.07149132440149351, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.779543399810791, + "learning_rate": 2.379209370424597e-07, + "loss": 1.1573, + "mean_token_accuracy": 0.6659355759620667, + "num_tokens": 16880661.0, + "step": 651 + }, + { + "epoch": 0.07160114210410719, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.8464932441711426, + "learning_rate": 2.3828696925329429e-07, + "loss": 1.1955, + "mean_token_accuracy": 0.6555575132369995, + "num_tokens": 16905765.0, + "step": 652 + }, + { + "epoch": 0.07171095980672085, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 3.389655828475952, + "learning_rate": 2.3865300146412886e-07, + "loss": 1.1289, + "mean_token_accuracy": 0.6726859211921692, + "num_tokens": 16926151.0, + "step": 653 + }, + { + "epoch": 0.07182077750933451, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.660480260848999, + "learning_rate": 2.390190336749634e-07, + "loss": 1.1508, + "mean_token_accuracy": 0.6664538383483887, + "num_tokens": 16955098.0, + "step": 654 + }, + { + "epoch": 0.07193059521194817, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 3.362518310546875, + "learning_rate": 2.393850658857979e-07, + "loss": 1.0798, + "mean_token_accuracy": 0.6846036911010742, + "num_tokens": 16975049.0, + "step": 655 + }, + { + "epoch": 0.07204041291456183, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 3.557884454727173, + "learning_rate": 2.397510980966325e-07, + "loss": 1.0044, + "mean_token_accuracy": 0.7017557621002197, + "num_tokens": 16993119.0, + "step": 656 + }, + { + "epoch": 0.07215023061717549, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.955540180206299, + "learning_rate": 2.4011713030746706e-07, + "loss": 1.0825, + "mean_token_accuracy": 0.6823548078536987, + "num_tokens": 17016760.0, + "step": 657 + }, + { + "epoch": 0.07226004831978915, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.306865930557251, + "learning_rate": 2.404831625183016e-07, + "loss": 1.1444, + "mean_token_accuracy": 0.664080798625946, + "num_tokens": 17046841.0, + "step": 658 + }, + { + "epoch": 0.0723698660224028, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.6601719856262207, + "learning_rate": 2.4084919472913616e-07, + "loss": 1.2297, + "mean_token_accuracy": 0.6476345062255859, + "num_tokens": 17073782.0, + "step": 659 + }, + { + "epoch": 0.07247968372501647, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.514773368835449, + "learning_rate": 2.412152269399707e-07, + "loss": 1.1662, + "mean_token_accuracy": 0.6676384806632996, + "num_tokens": 17105514.0, + "step": 660 + }, + { + "epoch": 0.07258950142763014, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.8902511596679688, + "learning_rate": 2.4158125915080525e-07, + "loss": 1.1095, + "mean_token_accuracy": 0.6756229400634766, + "num_tokens": 17128183.0, + "step": 661 + }, + { + "epoch": 0.0726993191302438, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 3.0955350399017334, + "learning_rate": 2.419472913616398e-07, + "loss": 1.1258, + "mean_token_accuracy": 0.6689935922622681, + "num_tokens": 17149575.0, + "step": 662 + }, + { + "epoch": 0.07280913683285746, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.79459810256958, + "learning_rate": 2.4231332357247435e-07, + "loss": 1.0797, + "mean_token_accuracy": 0.6868448853492737, + "num_tokens": 17175446.0, + "step": 663 + }, + { + "epoch": 0.07291895453547112, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.5052926540374756, + "learning_rate": 2.426793557833089e-07, + "loss": 1.0798, + "mean_token_accuracy": 0.6830324530601501, + "num_tokens": 17206437.0, + "step": 664 + }, + { + "epoch": 0.07302877223808478, + "ewc_loss": 2.8759241104125977e-06, + "grad_norm": 2.608150005340576, + "learning_rate": 2.430453879941435e-07, + "loss": 1.1821, + "mean_token_accuracy": 0.6517108082771301, + "num_tokens": 17232640.0, + "step": 665 + }, + { + "epoch": 0.07313858994069844, + "ewc_loss": 2.8908252716064453e-06, + "grad_norm": 2.6350927352905273, + "learning_rate": 2.4341142020497805e-07, + "loss": 1.1701, + "mean_token_accuracy": 0.6621098518371582, + "num_tokens": 17260421.0, + "step": 666 + }, + { + "epoch": 0.0732484076433121, + "ewc_loss": 2.8908252716064453e-06, + "grad_norm": 2.8632686138153076, + "learning_rate": 2.4377745241581255e-07, + "loss": 0.9737, + "mean_token_accuracy": 0.7122992873191833, + "num_tokens": 17283253.0, + "step": 667 + }, + { + "epoch": 0.07335822534592576, + "ewc_loss": 2.905726432800293e-06, + "grad_norm": 3.131007671356201, + "learning_rate": 2.4414348462664715e-07, + "loss": 1.1335, + "mean_token_accuracy": 0.664124608039856, + "num_tokens": 17303932.0, + "step": 668 + }, + { + "epoch": 0.07346804304853942, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 3.0405075550079346, + "learning_rate": 2.445095168374817e-07, + "loss": 1.0161, + "mean_token_accuracy": 0.6982955932617188, + "num_tokens": 17324634.0, + "step": 669 + }, + { + "epoch": 0.07357786075115308, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.7544589042663574, + "learning_rate": 2.4487554904831625e-07, + "loss": 0.9915, + "mean_token_accuracy": 0.7061298489570618, + "num_tokens": 17347927.0, + "step": 670 + }, + { + "epoch": 0.07368767845376675, + "ewc_loss": 2.9355287551879883e-06, + "grad_norm": 2.672532558441162, + "learning_rate": 2.452415812591508e-07, + "loss": 1.0491, + "mean_token_accuracy": 0.69495689868927, + "num_tokens": 17374596.0, + "step": 671 + }, + { + "epoch": 0.07379749615638041, + "ewc_loss": 2.9355287551879883e-06, + "grad_norm": 3.023151159286499, + "learning_rate": 2.4560761346998535e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.6856123208999634, + "num_tokens": 17397546.0, + "step": 672 + }, + { + "epoch": 0.07390731385899407, + "ewc_loss": 2.9355287551879883e-06, + "grad_norm": 3.049583673477173, + "learning_rate": 2.459736456808199e-07, + "loss": 1.2245, + "mean_token_accuracy": 0.6553959846496582, + "num_tokens": 17420746.0, + "step": 673 + }, + { + "epoch": 0.07401713156160773, + "ewc_loss": 2.9355287551879883e-06, + "grad_norm": 3.0657527446746826, + "learning_rate": 2.4633967789165444e-07, + "loss": 1.0769, + "mean_token_accuracy": 0.6918307542800903, + "num_tokens": 17441694.0, + "step": 674 + }, + { + "epoch": 0.07412694926422139, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.615191698074341, + "learning_rate": 2.46705710102489e-07, + "loss": 1.154, + "mean_token_accuracy": 0.6699793934822083, + "num_tokens": 17468898.0, + "step": 675 + }, + { + "epoch": 0.07423676696683505, + "ewc_loss": 2.9355287551879883e-06, + "grad_norm": 3.129668712615967, + "learning_rate": 2.4707174231332354e-07, + "loss": 1.0358, + "mean_token_accuracy": 0.6936391592025757, + "num_tokens": 17488143.0, + "step": 676 + }, + { + "epoch": 0.07434658466944871, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.645782947540283, + "learning_rate": 2.4743777452415815e-07, + "loss": 1.0557, + "mean_token_accuracy": 0.68935227394104, + "num_tokens": 17514042.0, + "step": 677 + }, + { + "epoch": 0.07445640237206237, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.8193838596343994, + "learning_rate": 2.478038067349927e-07, + "loss": 1.0438, + "mean_token_accuracy": 0.689500629901886, + "num_tokens": 17538766.0, + "step": 678 + }, + { + "epoch": 0.07456622007467603, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.498974323272705, + "learning_rate": 2.481698389458272e-07, + "loss": 1.1589, + "mean_token_accuracy": 0.6794425845146179, + "num_tokens": 17567824.0, + "step": 679 + }, + { + "epoch": 0.0746760377772897, + "ewc_loss": 2.9802322387695312e-06, + "grad_norm": 2.432724714279175, + "learning_rate": 2.485358711566618e-07, + "loss": 1.1819, + "mean_token_accuracy": 0.6602034568786621, + "num_tokens": 17599355.0, + "step": 680 + }, + { + "epoch": 0.07478585547990337, + "ewc_loss": 2.9802322387695312e-06, + "grad_norm": 2.594818353652954, + "learning_rate": 2.4890190336749634e-07, + "loss": 1.1537, + "mean_token_accuracy": 0.6631859540939331, + "num_tokens": 17625251.0, + "step": 681 + }, + { + "epoch": 0.07489567318251703, + "ewc_loss": 2.9802322387695312e-06, + "grad_norm": 2.590724229812622, + "learning_rate": 2.492679355783309e-07, + "loss": 1.1519, + "mean_token_accuracy": 0.66457200050354, + "num_tokens": 17654107.0, + "step": 682 + }, + { + "epoch": 0.07500549088513069, + "ewc_loss": 2.9802322387695312e-06, + "grad_norm": 3.013054370880127, + "learning_rate": 2.4963396778916544e-07, + "loss": 1.0533, + "mean_token_accuracy": 0.6956056952476501, + "num_tokens": 17677403.0, + "step": 683 + }, + { + "epoch": 0.07511530858774434, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.521329879760742, + "learning_rate": 2.5e-07, + "loss": 1.1713, + "mean_token_accuracy": 0.6632685661315918, + "num_tokens": 17706274.0, + "step": 684 + }, + { + "epoch": 0.075225126290358, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.633620262145996, + "learning_rate": 2.5036603221083454e-07, + "loss": 1.1436, + "mean_token_accuracy": 0.6587508320808411, + "num_tokens": 17732923.0, + "step": 685 + }, + { + "epoch": 0.07533494399297166, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.6797986030578613, + "learning_rate": 2.507320644216691e-07, + "loss": 1.0879, + "mean_token_accuracy": 0.6864843368530273, + "num_tokens": 17758057.0, + "step": 686 + }, + { + "epoch": 0.07544476169558532, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.4624745845794678, + "learning_rate": 2.5109809663250364e-07, + "loss": 1.0876, + "mean_token_accuracy": 0.6790511608123779, + "num_tokens": 17787037.0, + "step": 687 + }, + { + "epoch": 0.07555457939819898, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 3.1460678577423096, + "learning_rate": 2.514641288433382e-07, + "loss": 1.1434, + "mean_token_accuracy": 0.6722925901412964, + "num_tokens": 17807531.0, + "step": 688 + }, + { + "epoch": 0.07566439710081264, + "ewc_loss": 3.0547380447387695e-06, + "grad_norm": 3.505460262298584, + "learning_rate": 2.518301610541728e-07, + "loss": 1.0351, + "mean_token_accuracy": 0.6909915804862976, + "num_tokens": 17825910.0, + "step": 689 + }, + { + "epoch": 0.07577421480342632, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 3.0042784214019775, + "learning_rate": 2.521961932650073e-07, + "loss": 1.193, + "mean_token_accuracy": 0.6509366631507874, + "num_tokens": 17849674.0, + "step": 690 + }, + { + "epoch": 0.07588403250603998, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.3855245113372803, + "learning_rate": 2.525622254758419e-07, + "loss": 1.2067, + "mean_token_accuracy": 0.6594902873039246, + "num_tokens": 17884281.0, + "step": 691 + }, + { + "epoch": 0.07599385020865364, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.7065911293029785, + "learning_rate": 2.5292825768667643e-07, + "loss": 1.1219, + "mean_token_accuracy": 0.6752793788909912, + "num_tokens": 17907327.0, + "step": 692 + }, + { + "epoch": 0.0761036679112673, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.6782915592193604, + "learning_rate": 2.5329428989751093e-07, + "loss": 1.1176, + "mean_token_accuracy": 0.6763037443161011, + "num_tokens": 17931212.0, + "step": 693 + }, + { + "epoch": 0.07621348561388096, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.7928249835968018, + "learning_rate": 2.5366032210834553e-07, + "loss": 1.1818, + "mean_token_accuracy": 0.6597703099250793, + "num_tokens": 17957452.0, + "step": 694 + }, + { + "epoch": 0.07632330331649462, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 3.6029040813446045, + "learning_rate": 2.540263543191801e-07, + "loss": 1.0588, + "mean_token_accuracy": 0.6819174289703369, + "num_tokens": 17974591.0, + "step": 695 + }, + { + "epoch": 0.07643312101910828, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.7233903408050537, + "learning_rate": 2.543923865300146e-07, + "loss": 1.1042, + "mean_token_accuracy": 0.6848292350769043, + "num_tokens": 18002528.0, + "step": 696 + }, + { + "epoch": 0.07654293872172194, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.6970813274383545, + "learning_rate": 2.547584187408492e-07, + "loss": 1.1209, + "mean_token_accuracy": 0.6682605743408203, + "num_tokens": 18027367.0, + "step": 697 + }, + { + "epoch": 0.0766527564243356, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.6857564449310303, + "learning_rate": 2.5512445095168373e-07, + "loss": 1.1527, + "mean_token_accuracy": 0.6632688045501709, + "num_tokens": 18052873.0, + "step": 698 + }, + { + "epoch": 0.07676257412694927, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.4645321369171143, + "learning_rate": 2.5549048316251833e-07, + "loss": 1.1394, + "mean_token_accuracy": 0.6658499836921692, + "num_tokens": 18086772.0, + "step": 699 + }, + { + "epoch": 0.07687239182956293, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.6930432319641113, + "learning_rate": 2.5585651537335283e-07, + "loss": 1.1758, + "mean_token_accuracy": 0.6596505641937256, + "num_tokens": 18115267.0, + "step": 700 + }, + { + "epoch": 0.07698220953217659, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.65765118598938, + "learning_rate": 2.562225475841874e-07, + "loss": 1.1313, + "mean_token_accuracy": 0.6707584857940674, + "num_tokens": 18144689.0, + "step": 701 + }, + { + "epoch": 0.07709202723479025, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 3.00638747215271, + "learning_rate": 2.56588579795022e-07, + "loss": 1.1425, + "mean_token_accuracy": 0.6739086508750916, + "num_tokens": 18168911.0, + "step": 702 + }, + { + "epoch": 0.07720184493740391, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.84539794921875, + "learning_rate": 2.569546120058565e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.6756370067596436, + "num_tokens": 18193758.0, + "step": 703 + }, + { + "epoch": 0.07731166264001757, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.914504289627075, + "learning_rate": 2.57320644216691e-07, + "loss": 1.1024, + "mean_token_accuracy": 0.6734566688537598, + "num_tokens": 18216265.0, + "step": 704 + }, + { + "epoch": 0.07742148034263123, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.605637311935425, + "learning_rate": 2.576866764275256e-07, + "loss": 1.0711, + "mean_token_accuracy": 0.6832761764526367, + "num_tokens": 18240261.0, + "step": 705 + }, + { + "epoch": 0.07753129804524489, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.569373369216919, + "learning_rate": 2.580527086383602e-07, + "loss": 1.208, + "mean_token_accuracy": 0.6507420539855957, + "num_tokens": 18267819.0, + "step": 706 + }, + { + "epoch": 0.07764111574785855, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.5933337211608887, + "learning_rate": 2.584187408491947e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.6784372925758362, + "num_tokens": 18295498.0, + "step": 707 + }, + { + "epoch": 0.07775093345047221, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.4960269927978516, + "learning_rate": 2.5878477306002927e-07, + "loss": 1.148, + "mean_token_accuracy": 0.6619108319282532, + "num_tokens": 18324344.0, + "step": 708 + }, + { + "epoch": 0.07786075115308588, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 3.1856460571289062, + "learning_rate": 2.591508052708638e-07, + "loss": 1.1146, + "mean_token_accuracy": 0.6738110780715942, + "num_tokens": 18345249.0, + "step": 709 + }, + { + "epoch": 0.07797056885569954, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.791465997695923, + "learning_rate": 2.5951683748169837e-07, + "loss": 1.2176, + "mean_token_accuracy": 0.6527667045593262, + "num_tokens": 18369976.0, + "step": 710 + }, + { + "epoch": 0.0780803865583132, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.899369716644287, + "learning_rate": 2.598828696925329e-07, + "loss": 1.0495, + "mean_token_accuracy": 0.6879669427871704, + "num_tokens": 18389828.0, + "step": 711 + }, + { + "epoch": 0.07819020426092686, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4769227504730225, + "learning_rate": 2.6024890190336747e-07, + "loss": 1.1247, + "mean_token_accuracy": 0.6799423694610596, + "num_tokens": 18418228.0, + "step": 712 + }, + { + "epoch": 0.07830002196354052, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.7010204792022705, + "learning_rate": 2.6061493411420207e-07, + "loss": 1.1239, + "mean_token_accuracy": 0.6638659238815308, + "num_tokens": 18442229.0, + "step": 713 + }, + { + "epoch": 0.07840983966615418, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.495328426361084, + "learning_rate": 2.6098096632503657e-07, + "loss": 1.1548, + "mean_token_accuracy": 0.665289580821991, + "num_tokens": 18469419.0, + "step": 714 + }, + { + "epoch": 0.07851965736876784, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.591280460357666, + "learning_rate": 2.6134699853587117e-07, + "loss": 0.9564, + "mean_token_accuracy": 0.7134084105491638, + "num_tokens": 18493762.0, + "step": 715 + }, + { + "epoch": 0.0786294750713815, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 3.0498788356781006, + "learning_rate": 2.617130307467057e-07, + "loss": 1.1226, + "mean_token_accuracy": 0.6718751192092896, + "num_tokens": 18516052.0, + "step": 716 + }, + { + "epoch": 0.07873929277399516, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.473729372024536, + "learning_rate": 2.620790629575402e-07, + "loss": 1.1116, + "mean_token_accuracy": 0.6808727383613586, + "num_tokens": 18542974.0, + "step": 717 + }, + { + "epoch": 0.07884911047660884, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 3.139542818069458, + "learning_rate": 2.624450951683748e-07, + "loss": 1.049, + "mean_token_accuracy": 0.6866231560707092, + "num_tokens": 18561988.0, + "step": 718 + }, + { + "epoch": 0.0789589281792225, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.6797242164611816, + "learning_rate": 2.6281112737920937e-07, + "loss": 1.168, + "mean_token_accuracy": 0.6626036167144775, + "num_tokens": 18588546.0, + "step": 719 + }, + { + "epoch": 0.07906874588183616, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.728243350982666, + "learning_rate": 2.6317715959004386e-07, + "loss": 1.1955, + "mean_token_accuracy": 0.6526910066604614, + "num_tokens": 18615106.0, + "step": 720 + }, + { + "epoch": 0.07917856358444982, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.3904528617858887, + "learning_rate": 2.6354319180087846e-07, + "loss": 1.111, + "mean_token_accuracy": 0.6762189865112305, + "num_tokens": 18642433.0, + "step": 721 + }, + { + "epoch": 0.07928838128706348, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.440176248550415, + "learning_rate": 2.63909224011713e-07, + "loss": 1.1055, + "mean_token_accuracy": 0.6740891337394714, + "num_tokens": 18670567.0, + "step": 722 + }, + { + "epoch": 0.07939819898967714, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.9348301887512207, + "learning_rate": 2.642752562225476e-07, + "loss": 1.116, + "mean_token_accuracy": 0.6707602739334106, + "num_tokens": 18691814.0, + "step": 723 + }, + { + "epoch": 0.0795080166922908, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.5818846225738525, + "learning_rate": 2.646412884333821e-07, + "loss": 1.1512, + "mean_token_accuracy": 0.6664471626281738, + "num_tokens": 18717182.0, + "step": 724 + }, + { + "epoch": 0.07961783439490445, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.383172035217285, + "learning_rate": 2.6500732064421666e-07, + "loss": 1.131, + "mean_token_accuracy": 0.6648705005645752, + "num_tokens": 18746831.0, + "step": 725 + }, + { + "epoch": 0.07972765209751811, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.3961715698242188, + "learning_rate": 2.6537335285505126e-07, + "loss": 1.0725, + "mean_token_accuracy": 0.6828041076660156, + "num_tokens": 18777233.0, + "step": 726 + }, + { + "epoch": 0.07983746980013177, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.653681516647339, + "learning_rate": 2.6573938506588576e-07, + "loss": 1.1235, + "mean_token_accuracy": 0.6767450571060181, + "num_tokens": 18802479.0, + "step": 727 + }, + { + "epoch": 0.07994728750274545, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.3816885948181152, + "learning_rate": 2.661054172767203e-07, + "loss": 1.1462, + "mean_token_accuracy": 0.6628798246383667, + "num_tokens": 18831630.0, + "step": 728 + }, + { + "epoch": 0.08005710520535911, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.898271083831787, + "learning_rate": 2.664714494875549e-07, + "loss": 0.9651, + "mean_token_accuracy": 0.715522289276123, + "num_tokens": 18853499.0, + "step": 729 + }, + { + "epoch": 0.08016692290797277, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.778386354446411, + "learning_rate": 2.6683748169838946e-07, + "loss": 1.1809, + "mean_token_accuracy": 0.6647053956985474, + "num_tokens": 18875616.0, + "step": 730 + }, + { + "epoch": 0.08027674061058643, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.857696533203125, + "learning_rate": 2.67203513909224e-07, + "loss": 1.1066, + "mean_token_accuracy": 0.6772404313087463, + "num_tokens": 18895943.0, + "step": 731 + }, + { + "epoch": 0.08038655831320009, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.1359176635742188, + "learning_rate": 2.6756954612005856e-07, + "loss": 1.1632, + "mean_token_accuracy": 0.6630911827087402, + "num_tokens": 18930183.0, + "step": 732 + }, + { + "epoch": 0.08049637601581375, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.767929792404175, + "learning_rate": 2.679355783308931e-07, + "loss": 1.1613, + "mean_token_accuracy": 0.6643685698509216, + "num_tokens": 18953193.0, + "step": 733 + }, + { + "epoch": 0.08060619371842741, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.424227476119995, + "learning_rate": 2.6830161054172765e-07, + "loss": 1.1929, + "mean_token_accuracy": 0.6585039496421814, + "num_tokens": 18981251.0, + "step": 734 + }, + { + "epoch": 0.08071601142104107, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.4403676986694336, + "learning_rate": 2.686676427525622e-07, + "loss": 1.1427, + "mean_token_accuracy": 0.6767055988311768, + "num_tokens": 19009491.0, + "step": 735 + }, + { + "epoch": 0.08082582912365473, + "ewc_loss": 3.2633543014526367e-06, + "grad_norm": 2.449937582015991, + "learning_rate": 2.6903367496339675e-07, + "loss": 1.0732, + "mean_token_accuracy": 0.6842563152313232, + "num_tokens": 19037182.0, + "step": 736 + }, + { + "epoch": 0.0809356468262684, + "ewc_loss": 3.2633543014526367e-06, + "grad_norm": 2.8130483627319336, + "learning_rate": 2.693997071742313e-07, + "loss": 1.0855, + "mean_token_accuracy": 0.6808425784111023, + "num_tokens": 19059901.0, + "step": 737 + }, + { + "epoch": 0.08104546452888206, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.642768621444702, + "learning_rate": 2.6976573938506585e-07, + "loss": 1.1688, + "mean_token_accuracy": 0.6638927459716797, + "num_tokens": 19085117.0, + "step": 738 + }, + { + "epoch": 0.08115528223149572, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.476020336151123, + "learning_rate": 2.7013177159590045e-07, + "loss": 1.0756, + "mean_token_accuracy": 0.6820361614227295, + "num_tokens": 19111093.0, + "step": 739 + }, + { + "epoch": 0.08126509993410938, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.7094736099243164, + "learning_rate": 2.70497803806735e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6758209466934204, + "num_tokens": 19135550.0, + "step": 740 + }, + { + "epoch": 0.08137491763672304, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.476199150085449, + "learning_rate": 2.708638360175695e-07, + "loss": 1.0618, + "mean_token_accuracy": 0.6854633092880249, + "num_tokens": 19162633.0, + "step": 741 + }, + { + "epoch": 0.0814847353393367, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.584932565689087, + "learning_rate": 2.712298682284041e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6970076560974121, + "num_tokens": 19186813.0, + "step": 742 + }, + { + "epoch": 0.08159455304195036, + "ewc_loss": 3.3229589462280273e-06, + "grad_norm": 2.3393356800079346, + "learning_rate": 2.7159590043923865e-07, + "loss": 1.1028, + "mean_token_accuracy": 0.6750641465187073, + "num_tokens": 19217664.0, + "step": 743 + }, + { + "epoch": 0.08170437074456402, + "ewc_loss": 3.3527612686157227e-06, + "grad_norm": 2.937598466873169, + "learning_rate": 2.7196193265007315e-07, + "loss": 1.0823, + "mean_token_accuracy": 0.6764939427375793, + "num_tokens": 19239127.0, + "step": 744 + }, + { + "epoch": 0.08181418844717768, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.667562246322632, + "learning_rate": 2.7232796486090775e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.678024411201477, + "num_tokens": 19263028.0, + "step": 745 + }, + { + "epoch": 0.08192400614979134, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.353382110595703, + "learning_rate": 2.726939970717423e-07, + "loss": 1.1847, + "mean_token_accuracy": 0.6521667838096619, + "num_tokens": 19292808.0, + "step": 746 + }, + { + "epoch": 0.08203382385240501, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.6867778301239014, + "learning_rate": 2.730600292825769e-07, + "loss": 1.116, + "mean_token_accuracy": 0.6740553975105286, + "num_tokens": 19314443.0, + "step": 747 + }, + { + "epoch": 0.08214364155501867, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.680598497390747, + "learning_rate": 2.734260614934114e-07, + "loss": 1.0645, + "mean_token_accuracy": 0.6912717819213867, + "num_tokens": 19339945.0, + "step": 748 + }, + { + "epoch": 0.08225345925763233, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.428165912628174, + "learning_rate": 2.7379209370424594e-07, + "loss": 1.1836, + "mean_token_accuracy": 0.6590647101402283, + "num_tokens": 19370487.0, + "step": 749 + }, + { + "epoch": 0.08236327696024599, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.4966790676116943, + "learning_rate": 2.7415812591508055e-07, + "loss": 1.0848, + "mean_token_accuracy": 0.6794870495796204, + "num_tokens": 19395644.0, + "step": 750 + }, + { + "epoch": 0.08247309466285965, + "ewc_loss": 3.382563591003418e-06, + "grad_norm": 2.238874673843384, + "learning_rate": 2.7452415812591504e-07, + "loss": 1.1649, + "mean_token_accuracy": 0.6595980525016785, + "num_tokens": 19426218.0, + "step": 751 + }, + { + "epoch": 0.08258291236547331, + "ewc_loss": 3.3527612686157227e-06, + "grad_norm": 2.5487895011901855, + "learning_rate": 2.748901903367496e-07, + "loss": 1.1629, + "mean_token_accuracy": 0.6664713621139526, + "num_tokens": 19451846.0, + "step": 752 + }, + { + "epoch": 0.08269273006808697, + "ewc_loss": 3.382563591003418e-06, + "grad_norm": 2.5914266109466553, + "learning_rate": 2.752562225475842e-07, + "loss": 1.0858, + "mean_token_accuracy": 0.6861602067947388, + "num_tokens": 19476296.0, + "step": 753 + }, + { + "epoch": 0.08280254777070063, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.406494140625, + "learning_rate": 2.756222547584187e-07, + "loss": 1.1254, + "mean_token_accuracy": 0.6787965893745422, + "num_tokens": 19504200.0, + "step": 754 + }, + { + "epoch": 0.08291236547331429, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.701171398162842, + "learning_rate": 2.759882869692533e-07, + "loss": 1.2031, + "mean_token_accuracy": 0.6565680503845215, + "num_tokens": 19527134.0, + "step": 755 + }, + { + "epoch": 0.08302218317592797, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.427933692932129, + "learning_rate": 2.7635431918008784e-07, + "loss": 1.0866, + "mean_token_accuracy": 0.6839548945426941, + "num_tokens": 19556429.0, + "step": 756 + }, + { + "epoch": 0.08313200087854163, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.190009593963623, + "learning_rate": 2.767203513909224e-07, + "loss": 1.1598, + "mean_token_accuracy": 0.6620863676071167, + "num_tokens": 19587759.0, + "step": 757 + }, + { + "epoch": 0.08324181858115529, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 3.116607427597046, + "learning_rate": 2.7708638360175694e-07, + "loss": 1.1088, + "mean_token_accuracy": 0.6822325587272644, + "num_tokens": 19607412.0, + "step": 758 + }, + { + "epoch": 0.08335163628376895, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.6321864128112793, + "learning_rate": 2.774524158125915e-07, + "loss": 1.0047, + "mean_token_accuracy": 0.7105526924133301, + "num_tokens": 19630184.0, + "step": 759 + }, + { + "epoch": 0.0834614539863826, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.6747660636901855, + "learning_rate": 2.7781844802342604e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6818566918373108, + "num_tokens": 19655787.0, + "step": 760 + }, + { + "epoch": 0.08357127168899627, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.4278781414031982, + "learning_rate": 2.781844802342606e-07, + "loss": 1.0524, + "mean_token_accuracy": 0.6932591199874878, + "num_tokens": 19683813.0, + "step": 761 + }, + { + "epoch": 0.08368108939160993, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.786100387573242, + "learning_rate": 2.7855051244509513e-07, + "loss": 1.081, + "mean_token_accuracy": 0.6817461848258972, + "num_tokens": 19706512.0, + "step": 762 + }, + { + "epoch": 0.08379090709422359, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.3553805351257324, + "learning_rate": 2.7891654465592974e-07, + "loss": 1.1499, + "mean_token_accuracy": 0.6704550981521606, + "num_tokens": 19737097.0, + "step": 763 + }, + { + "epoch": 0.08390072479683724, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.609205484390259, + "learning_rate": 2.792825768667643e-07, + "loss": 1.0052, + "mean_token_accuracy": 0.7027167677879333, + "num_tokens": 19762815.0, + "step": 764 + }, + { + "epoch": 0.0840105424994509, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.7851130962371826, + "learning_rate": 2.796486090775988e-07, + "loss": 1.1178, + "mean_token_accuracy": 0.6755751967430115, + "num_tokens": 19785444.0, + "step": 765 + }, + { + "epoch": 0.08412036020206458, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.719066619873047, + "learning_rate": 2.800146412884334e-07, + "loss": 1.1256, + "mean_token_accuracy": 0.6717267632484436, + "num_tokens": 19809678.0, + "step": 766 + }, + { + "epoch": 0.08423017790467824, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.522826671600342, + "learning_rate": 2.8038067349926793e-07, + "loss": 1.0995, + "mean_token_accuracy": 0.6751440763473511, + "num_tokens": 19834658.0, + "step": 767 + }, + { + "epoch": 0.0843399956072919, + "ewc_loss": 3.4570693969726562e-06, + "grad_norm": 2.470951795578003, + "learning_rate": 2.8074670571010243e-07, + "loss": 1.1852, + "mean_token_accuracy": 0.6611358523368835, + "num_tokens": 19860076.0, + "step": 768 + }, + { + "epoch": 0.08444981330990556, + "ewc_loss": 3.4570693969726562e-06, + "grad_norm": 2.68727970123291, + "learning_rate": 2.8111273792093703e-07, + "loss": 1.1401, + "mean_token_accuracy": 0.6680662035942078, + "num_tokens": 19885069.0, + "step": 769 + }, + { + "epoch": 0.08455963101251922, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.7535665035247803, + "learning_rate": 2.814787701317716e-07, + "loss": 1.2001, + "mean_token_accuracy": 0.6601008772850037, + "num_tokens": 19909921.0, + "step": 770 + }, + { + "epoch": 0.08466944871513288, + "ewc_loss": 3.4570693969726562e-06, + "grad_norm": 2.564481019973755, + "learning_rate": 2.818448023426062e-07, + "loss": 1.0552, + "mean_token_accuracy": 0.6977459192276001, + "num_tokens": 19935788.0, + "step": 771 + }, + { + "epoch": 0.08477926641774654, + "ewc_loss": 3.4570693969726562e-06, + "grad_norm": 2.338052988052368, + "learning_rate": 2.822108345534407e-07, + "loss": 1.1329, + "mean_token_accuracy": 0.6675270199775696, + "num_tokens": 19965350.0, + "step": 772 + }, + { + "epoch": 0.0848890841203602, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.710341215133667, + "learning_rate": 2.8257686676427523e-07, + "loss": 1.006, + "mean_token_accuracy": 0.704796314239502, + "num_tokens": 19985991.0, + "step": 773 + }, + { + "epoch": 0.08499890182297386, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.4576575756073, + "learning_rate": 2.8294289897510983e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6836866140365601, + "num_tokens": 20013789.0, + "step": 774 + }, + { + "epoch": 0.08510871952558753, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.7983248233795166, + "learning_rate": 2.833089311859443e-07, + "loss": 1.0248, + "mean_token_accuracy": 0.6922900080680847, + "num_tokens": 20034001.0, + "step": 775 + }, + { + "epoch": 0.08521853722820119, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.6021225452423096, + "learning_rate": 2.836749633967789e-07, + "loss": 1.145, + "mean_token_accuracy": 0.6643486618995667, + "num_tokens": 20060222.0, + "step": 776 + }, + { + "epoch": 0.08532835493081485, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.3146543502807617, + "learning_rate": 2.840409956076135e-07, + "loss": 1.0793, + "mean_token_accuracy": 0.6840351819992065, + "num_tokens": 20093804.0, + "step": 777 + }, + { + "epoch": 0.08543817263342851, + "ewc_loss": 3.5315752029418945e-06, + "grad_norm": 2.259274482727051, + "learning_rate": 2.8440702781844797e-07, + "loss": 1.2164, + "mean_token_accuracy": 0.643078625202179, + "num_tokens": 20126639.0, + "step": 778 + }, + { + "epoch": 0.08554799033604217, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.5648505687713623, + "learning_rate": 2.847730600292826e-07, + "loss": 1.0146, + "mean_token_accuracy": 0.7052335739135742, + "num_tokens": 20154205.0, + "step": 779 + }, + { + "epoch": 0.08565780803865583, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2766361236572266, + "learning_rate": 2.851390922401171e-07, + "loss": 1.2067, + "mean_token_accuracy": 0.6638845801353455, + "num_tokens": 20185261.0, + "step": 780 + }, + { + "epoch": 0.08576762574126949, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.8586246967315674, + "learning_rate": 2.8550512445095167e-07, + "loss": 1.0082, + "mean_token_accuracy": 0.6944681406021118, + "num_tokens": 20205777.0, + "step": 781 + }, + { + "epoch": 0.08587744344388315, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.2360806465148926, + "learning_rate": 2.858711566617862e-07, + "loss": 1.0532, + "mean_token_accuracy": 0.6897708177566528, + "num_tokens": 20237386.0, + "step": 782 + }, + { + "epoch": 0.08598726114649681, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.57865309715271, + "learning_rate": 2.8623718887262077e-07, + "loss": 1.111, + "mean_token_accuracy": 0.6782652139663696, + "num_tokens": 20261696.0, + "step": 783 + }, + { + "epoch": 0.08609707884911047, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.4561574459075928, + "learning_rate": 2.866032210834553e-07, + "loss": 1.0786, + "mean_token_accuracy": 0.6817777752876282, + "num_tokens": 20287351.0, + "step": 784 + }, + { + "epoch": 0.08620689655172414, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.656750202178955, + "learning_rate": 2.8696925329428987e-07, + "loss": 1.0004, + "mean_token_accuracy": 0.7029145956039429, + "num_tokens": 20309391.0, + "step": 785 + }, + { + "epoch": 0.0863167142543378, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.7206432819366455, + "learning_rate": 2.873352855051244e-07, + "loss": 1.0596, + "mean_token_accuracy": 0.6817185282707214, + "num_tokens": 20332515.0, + "step": 786 + }, + { + "epoch": 0.08642653195695146, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.5700244903564453, + "learning_rate": 2.87701317715959e-07, + "loss": 1.0973, + "mean_token_accuracy": 0.6854998469352722, + "num_tokens": 20359379.0, + "step": 787 + }, + { + "epoch": 0.08653634965956512, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.6431291103363037, + "learning_rate": 2.8806734992679357e-07, + "loss": 1.1266, + "mean_token_accuracy": 0.6669526100158691, + "num_tokens": 20383436.0, + "step": 788 + }, + { + "epoch": 0.08664616736217878, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.2224414348602295, + "learning_rate": 2.8843338213762807e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.6765379309654236, + "num_tokens": 20414443.0, + "step": 789 + }, + { + "epoch": 0.08675598506479244, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.766042709350586, + "learning_rate": 2.8879941434846267e-07, + "loss": 1.0649, + "mean_token_accuracy": 0.6809260845184326, + "num_tokens": 20437080.0, + "step": 790 + }, + { + "epoch": 0.0868658027674061, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.54178786277771, + "learning_rate": 2.891654465592972e-07, + "loss": 0.962, + "mean_token_accuracy": 0.7146302461624146, + "num_tokens": 20461294.0, + "step": 791 + }, + { + "epoch": 0.08697562047001976, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.555973529815674, + "learning_rate": 2.895314787701317e-07, + "loss": 1.1005, + "mean_token_accuracy": 0.6724787950515747, + "num_tokens": 20486377.0, + "step": 792 + }, + { + "epoch": 0.08708543817263342, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.782360315322876, + "learning_rate": 2.898975109809663e-07, + "loss": 1.0532, + "mean_token_accuracy": 0.6905127763748169, + "num_tokens": 20508898.0, + "step": 793 + }, + { + "epoch": 0.0871952558752471, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2635786533355713, + "learning_rate": 2.9026354319180086e-07, + "loss": 1.1311, + "mean_token_accuracy": 0.6699651479721069, + "num_tokens": 20538441.0, + "step": 794 + }, + { + "epoch": 0.08730507357786076, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.5338730812072754, + "learning_rate": 2.9062957540263547e-07, + "loss": 1.078, + "mean_token_accuracy": 0.6839621067047119, + "num_tokens": 20561412.0, + "step": 795 + }, + { + "epoch": 0.08741489128047442, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.684417247772217, + "learning_rate": 2.9099560761346996e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.6822245717048645, + "num_tokens": 20586226.0, + "step": 796 + }, + { + "epoch": 0.08752470898308808, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.250908136367798, + "learning_rate": 2.913616398243045e-07, + "loss": 1.201, + "mean_token_accuracy": 0.654447615146637, + "num_tokens": 20618182.0, + "step": 797 + }, + { + "epoch": 0.08763452668570174, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.3878695964813232, + "learning_rate": 2.917276720351391e-07, + "loss": 1.1652, + "mean_token_accuracy": 0.671093225479126, + "num_tokens": 20647672.0, + "step": 798 + }, + { + "epoch": 0.0877443443883154, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.5982322692871094, + "learning_rate": 2.920937042459736e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.6895591020584106, + "num_tokens": 20671724.0, + "step": 799 + }, + { + "epoch": 0.08785416209092906, + "ewc_loss": 3.6656856536865234e-06, + "grad_norm": 2.293107032775879, + "learning_rate": 2.9245973645680816e-07, + "loss": 1.145, + "mean_token_accuracy": 0.6832610964775085, + "num_tokens": 20700193.0, + "step": 800 + }, + { + "epoch": 0.08796397979354272, + "ewc_loss": 3.6656856536865234e-06, + "grad_norm": 2.521988868713379, + "learning_rate": 2.9282576866764276e-07, + "loss": 1.0704, + "mean_token_accuracy": 0.6814356446266174, + "num_tokens": 20724905.0, + "step": 801 + }, + { + "epoch": 0.08807379749615638, + "ewc_loss": 3.680586814880371e-06, + "grad_norm": 2.5125956535339355, + "learning_rate": 2.9319180087847726e-07, + "loss": 1.1293, + "mean_token_accuracy": 0.6708621978759766, + "num_tokens": 20749749.0, + "step": 802 + }, + { + "epoch": 0.08818361519877004, + "ewc_loss": 3.680586814880371e-06, + "grad_norm": 2.3151865005493164, + "learning_rate": 2.9355783308931186e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.6984109878540039, + "num_tokens": 20778139.0, + "step": 803 + }, + { + "epoch": 0.08829343290138371, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.5054359436035156, + "learning_rate": 2.939238653001464e-07, + "loss": 1.1151, + "mean_token_accuracy": 0.687379002571106, + "num_tokens": 20801934.0, + "step": 804 + }, + { + "epoch": 0.08840325060399737, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.3367912769317627, + "learning_rate": 2.9428989751098096e-07, + "loss": 1.1127, + "mean_token_accuracy": 0.6767265200614929, + "num_tokens": 20832336.0, + "step": 805 + }, + { + "epoch": 0.08851306830661103, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.4251692295074463, + "learning_rate": 2.946559297218155e-07, + "loss": 1.1169, + "mean_token_accuracy": 0.6699598431587219, + "num_tokens": 20858646.0, + "step": 806 + }, + { + "epoch": 0.08862288600922469, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.353773355484009, + "learning_rate": 2.9502196193265006e-07, + "loss": 1.1414, + "mean_token_accuracy": 0.6671633124351501, + "num_tokens": 20887602.0, + "step": 807 + }, + { + "epoch": 0.08873270371183835, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.6585886478424072, + "learning_rate": 2.953879941434846e-07, + "loss": 1.1666, + "mean_token_accuracy": 0.6588118076324463, + "num_tokens": 20912488.0, + "step": 808 + }, + { + "epoch": 0.08884252141445201, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.3169686794281006, + "learning_rate": 2.9575402635431915e-07, + "loss": 1.1893, + "mean_token_accuracy": 0.6529806852340698, + "num_tokens": 20940624.0, + "step": 809 + }, + { + "epoch": 0.08895233911706567, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.2049167156219482, + "learning_rate": 2.961200585651537e-07, + "loss": 1.0838, + "mean_token_accuracy": 0.6824722290039062, + "num_tokens": 20970873.0, + "step": 810 + }, + { + "epoch": 0.08906215681967933, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.6664745807647705, + "learning_rate": 2.964860907759883e-07, + "loss": 1.0235, + "mean_token_accuracy": 0.6948245763778687, + "num_tokens": 20993869.0, + "step": 811 + }, + { + "epoch": 0.08917197452229299, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.4334030151367188, + "learning_rate": 2.9685212298682285e-07, + "loss": 1.0289, + "mean_token_accuracy": 0.6936309337615967, + "num_tokens": 21019651.0, + "step": 812 + }, + { + "epoch": 0.08928179222490666, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.3105850219726562, + "learning_rate": 2.9721815519765735e-07, + "loss": 1.0602, + "mean_token_accuracy": 0.6886669993400574, + "num_tokens": 21048736.0, + "step": 813 + }, + { + "epoch": 0.08939160992752032, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.695263385772705, + "learning_rate": 2.9758418740849195e-07, + "loss": 1.1906, + "mean_token_accuracy": 0.6675076484680176, + "num_tokens": 21071505.0, + "step": 814 + }, + { + "epoch": 0.08950142763013398, + "ewc_loss": 3.725290298461914e-06, + "grad_norm": 2.6565449237823486, + "learning_rate": 2.979502196193265e-07, + "loss": 1.0225, + "mean_token_accuracy": 0.6978007555007935, + "num_tokens": 21094803.0, + "step": 815 + }, + { + "epoch": 0.08961124533274764, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.53244686126709, + "learning_rate": 2.98316251830161e-07, + "loss": 1.098, + "mean_token_accuracy": 0.6810460686683655, + "num_tokens": 21120328.0, + "step": 816 + }, + { + "epoch": 0.0897210630353613, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.4287219047546387, + "learning_rate": 2.986822840409956e-07, + "loss": 1.1129, + "mean_token_accuracy": 0.6717947721481323, + "num_tokens": 21146162.0, + "step": 817 + }, + { + "epoch": 0.08983088073797496, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.5221402645111084, + "learning_rate": 2.9904831625183015e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6883341073989868, + "num_tokens": 21169041.0, + "step": 818 + }, + { + "epoch": 0.08994069844058862, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.4378881454467773, + "learning_rate": 2.994143484626647e-07, + "loss": 1.0972, + "mean_token_accuracy": 0.6818088293075562, + "num_tokens": 21195693.0, + "step": 819 + }, + { + "epoch": 0.09005051614320228, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.751384973526001, + "learning_rate": 2.9978038067349925e-07, + "loss": 1.0606, + "mean_token_accuracy": 0.6914728283882141, + "num_tokens": 21217125.0, + "step": 820 + }, + { + "epoch": 0.09016033384581594, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.4615986347198486, + "learning_rate": 3.001464128843338e-07, + "loss": 1.1611, + "mean_token_accuracy": 0.6597617268562317, + "num_tokens": 21243201.0, + "step": 821 + }, + { + "epoch": 0.0902701515484296, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.2875070571899414, + "learning_rate": 3.005124450951684e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.6900665760040283, + "num_tokens": 21272069.0, + "step": 822 + }, + { + "epoch": 0.09037996925104327, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.4142065048217773, + "learning_rate": 3.008784773060029e-07, + "loss": 1.1496, + "mean_token_accuracy": 0.6648386120796204, + "num_tokens": 21299739.0, + "step": 823 + }, + { + "epoch": 0.09048978695365693, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.9530558586120605, + "learning_rate": 3.0124450951683744e-07, + "loss": 1.0628, + "mean_token_accuracy": 0.6773849725723267, + "num_tokens": 21318516.0, + "step": 824 + }, + { + "epoch": 0.0905996046562706, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.836505889892578, + "learning_rate": 3.0161054172767204e-07, + "loss": 1.0798, + "mean_token_accuracy": 0.675198495388031, + "num_tokens": 21337893.0, + "step": 825 + }, + { + "epoch": 0.09070942235888425, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.3581652641296387, + "learning_rate": 3.0197657393850654e-07, + "loss": 1.195, + "mean_token_accuracy": 0.64875328540802, + "num_tokens": 21366728.0, + "step": 826 + }, + { + "epoch": 0.09081924006149791, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.2046220302581787, + "learning_rate": 3.0234260614934114e-07, + "loss": 1.1862, + "mean_token_accuracy": 0.6518399119377136, + "num_tokens": 21398742.0, + "step": 827 + }, + { + "epoch": 0.09092905776411157, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.562671184539795, + "learning_rate": 3.027086383601757e-07, + "loss": 0.9935, + "mean_token_accuracy": 0.7041919231414795, + "num_tokens": 21422256.0, + "step": 828 + }, + { + "epoch": 0.09103887546672523, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.496730089187622, + "learning_rate": 3.0307467057101024e-07, + "loss": 1.1259, + "mean_token_accuracy": 0.6718701124191284, + "num_tokens": 21445882.0, + "step": 829 + }, + { + "epoch": 0.09114869316933889, + "ewc_loss": 3.814697265625e-06, + "grad_norm": 2.415560483932495, + "learning_rate": 3.034407027818448e-07, + "loss": 1.1032, + "mean_token_accuracy": 0.6735970377922058, + "num_tokens": 21472863.0, + "step": 830 + }, + { + "epoch": 0.09125851087195255, + "ewc_loss": 3.814697265625e-06, + "grad_norm": 2.741645097732544, + "learning_rate": 3.0380673499267934e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6908828020095825, + "num_tokens": 21496899.0, + "step": 831 + }, + { + "epoch": 0.09136832857456623, + "ewc_loss": 3.814697265625e-06, + "grad_norm": 2.7547433376312256, + "learning_rate": 3.041727672035139e-07, + "loss": 1.1146, + "mean_token_accuracy": 0.676164984703064, + "num_tokens": 21519107.0, + "step": 832 + }, + { + "epoch": 0.09147814627717989, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.592557430267334, + "learning_rate": 3.0453879941434844e-07, + "loss": 1.1502, + "mean_token_accuracy": 0.6648604273796082, + "num_tokens": 21544059.0, + "step": 833 + }, + { + "epoch": 0.09158796397979355, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.583017110824585, + "learning_rate": 3.04904831625183e-07, + "loss": 1.074, + "mean_token_accuracy": 0.6874346733093262, + "num_tokens": 21568468.0, + "step": 834 + }, + { + "epoch": 0.0916977816824072, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.509265184402466, + "learning_rate": 3.052708638360176e-07, + "loss": 1.0918, + "mean_token_accuracy": 0.6740951538085938, + "num_tokens": 21592891.0, + "step": 835 + }, + { + "epoch": 0.09180759938502087, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.8203682899475098, + "learning_rate": 3.056368960468521e-07, + "loss": 1.0847, + "mean_token_accuracy": 0.6830224394798279, + "num_tokens": 21614569.0, + "step": 836 + }, + { + "epoch": 0.09191741708763453, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.5733726024627686, + "learning_rate": 3.0600292825768663e-07, + "loss": 1.1508, + "mean_token_accuracy": 0.658331573009491, + "num_tokens": 21638089.0, + "step": 837 + }, + { + "epoch": 0.09202723479024819, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.1038904190063477, + "learning_rate": 3.0636896046852124e-07, + "loss": 1.0964, + "mean_token_accuracy": 0.6880975365638733, + "num_tokens": 21668891.0, + "step": 838 + }, + { + "epoch": 0.09213705249286185, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.395914077758789, + "learning_rate": 3.067349926793558e-07, + "loss": 1.1473, + "mean_token_accuracy": 0.6684695482254028, + "num_tokens": 21696713.0, + "step": 839 + }, + { + "epoch": 0.0922468701954755, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.261251449584961, + "learning_rate": 3.071010248901903e-07, + "loss": 1.0091, + "mean_token_accuracy": 0.7034391164779663, + "num_tokens": 21724588.0, + "step": 840 + }, + { + "epoch": 0.09235668789808917, + "ewc_loss": 3.904104232788086e-06, + "grad_norm": 2.884678602218628, + "learning_rate": 3.074670571010249e-07, + "loss": 1.0721, + "mean_token_accuracy": 0.6934733390808105, + "num_tokens": 21745178.0, + "step": 841 + }, + { + "epoch": 0.09246650560070284, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.6157073974609375, + "learning_rate": 3.0783308931185943e-07, + "loss": 1.0718, + "mean_token_accuracy": 0.6882334351539612, + "num_tokens": 21768870.0, + "step": 842 + }, + { + "epoch": 0.0925763233033165, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3619847297668457, + "learning_rate": 3.08199121522694e-07, + "loss": 1.0571, + "mean_token_accuracy": 0.6867208480834961, + "num_tokens": 21796552.0, + "step": 843 + }, + { + "epoch": 0.09268614100593016, + "ewc_loss": 3.9637088775634766e-06, + "grad_norm": 2.398449182510376, + "learning_rate": 3.0856515373352853e-07, + "loss": 1.1292, + "mean_token_accuracy": 0.6732214689254761, + "num_tokens": 21825989.0, + "step": 844 + }, + { + "epoch": 0.09279595870854382, + "ewc_loss": 3.9637088775634766e-06, + "grad_norm": 2.639873743057251, + "learning_rate": 3.089311859443631e-07, + "loss": 1.0362, + "mean_token_accuracy": 0.6916487216949463, + "num_tokens": 21848075.0, + "step": 845 + }, + { + "epoch": 0.09290577641115748, + "ewc_loss": 3.993511199951172e-06, + "grad_norm": 2.3201892375946045, + "learning_rate": 3.092972181551977e-07, + "loss": 1.1333, + "mean_token_accuracy": 0.673363208770752, + "num_tokens": 21877666.0, + "step": 846 + }, + { + "epoch": 0.09301559411377114, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.272193670272827, + "learning_rate": 3.096632503660322e-07, + "loss": 1.0819, + "mean_token_accuracy": 0.6778210997581482, + "num_tokens": 21904463.0, + "step": 847 + }, + { + "epoch": 0.0931254118163848, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.527507781982422, + "learning_rate": 3.1002928257686673e-07, + "loss": 1.0573, + "mean_token_accuracy": 0.6822757124900818, + "num_tokens": 21927852.0, + "step": 848 + }, + { + "epoch": 0.09323522951899846, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.325789451599121, + "learning_rate": 3.1039531478770133e-07, + "loss": 1.053, + "mean_token_accuracy": 0.6950393915176392, + "num_tokens": 21956317.0, + "step": 849 + }, + { + "epoch": 0.09334504722161212, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.401587963104248, + "learning_rate": 3.107613469985358e-07, + "loss": 1.0536, + "mean_token_accuracy": 0.6845046281814575, + "num_tokens": 21981920.0, + "step": 850 + }, + { + "epoch": 0.09345486492422579, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.754838228225708, + "learning_rate": 3.1112737920937043e-07, + "loss": 1.1163, + "mean_token_accuracy": 0.6714121699333191, + "num_tokens": 22006163.0, + "step": 851 + }, + { + "epoch": 0.09356468262683945, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.535212755203247, + "learning_rate": 3.11493411420205e-07, + "loss": 1.0873, + "mean_token_accuracy": 0.6768754720687866, + "num_tokens": 22029906.0, + "step": 852 + }, + { + "epoch": 0.09367450032945311, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5109033584594727, + "learning_rate": 3.1185944363103947e-07, + "loss": 1.1266, + "mean_token_accuracy": 0.674166738986969, + "num_tokens": 22055611.0, + "step": 853 + }, + { + "epoch": 0.09378431803206677, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.391601085662842, + "learning_rate": 3.122254758418741e-07, + "loss": 1.0751, + "mean_token_accuracy": 0.6749110221862793, + "num_tokens": 22081614.0, + "step": 854 + }, + { + "epoch": 0.09389413573468043, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5837790966033936, + "learning_rate": 3.125915080527086e-07, + "loss": 1.0587, + "mean_token_accuracy": 0.6812784671783447, + "num_tokens": 22104956.0, + "step": 855 + }, + { + "epoch": 0.09400395343729409, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.239994525909424, + "learning_rate": 3.1295754026354317e-07, + "loss": 1.106, + "mean_token_accuracy": 0.6723328232765198, + "num_tokens": 22135119.0, + "step": 856 + }, + { + "epoch": 0.09411377113990775, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.7001469135284424, + "learning_rate": 3.133235724743777e-07, + "loss": 1.12, + "mean_token_accuracy": 0.6767785549163818, + "num_tokens": 22157625.0, + "step": 857 + }, + { + "epoch": 0.09422358884252141, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2444581985473633, + "learning_rate": 3.1368960468521227e-07, + "loss": 1.0946, + "mean_token_accuracy": 0.6816983222961426, + "num_tokens": 22188178.0, + "step": 858 + }, + { + "epoch": 0.09433340654513507, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3175790309906006, + "learning_rate": 3.1405563689604687e-07, + "loss": 1.0974, + "mean_token_accuracy": 0.6745567321777344, + "num_tokens": 22213762.0, + "step": 859 + }, + { + "epoch": 0.09444322424774873, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.697458028793335, + "learning_rate": 3.1442166910688137e-07, + "loss": 1.104, + "mean_token_accuracy": 0.6766191720962524, + "num_tokens": 22235680.0, + "step": 860 + }, + { + "epoch": 0.0945530419503624, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6425349712371826, + "learning_rate": 3.147877013177159e-07, + "loss": 1.0337, + "mean_token_accuracy": 0.6928590536117554, + "num_tokens": 22258638.0, + "step": 861 + }, + { + "epoch": 0.09466285965297606, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5268285274505615, + "learning_rate": 3.151537335285505e-07, + "loss": 1.0763, + "mean_token_accuracy": 0.6831841468811035, + "num_tokens": 22282129.0, + "step": 862 + }, + { + "epoch": 0.09477267735558972, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1671602725982666, + "learning_rate": 3.1551976573938507e-07, + "loss": 0.9964, + "mean_token_accuracy": 0.7027417421340942, + "num_tokens": 22312222.0, + "step": 863 + }, + { + "epoch": 0.09488249505820338, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.678809881210327, + "learning_rate": 3.1588579795021956e-07, + "loss": 1.0898, + "mean_token_accuracy": 0.6744571328163147, + "num_tokens": 22335476.0, + "step": 864 + }, + { + "epoch": 0.09499231276081704, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4800503253936768, + "learning_rate": 3.1625183016105417e-07, + "loss": 1.143, + "mean_token_accuracy": 0.6853702068328857, + "num_tokens": 22362628.0, + "step": 865 + }, + { + "epoch": 0.0951021304634307, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.632063865661621, + "learning_rate": 3.166178623718887e-07, + "loss": 1.1435, + "mean_token_accuracy": 0.6706566214561462, + "num_tokens": 22388402.0, + "step": 866 + }, + { + "epoch": 0.09521194816604436, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3411691188812256, + "learning_rate": 3.1698389458272327e-07, + "loss": 1.1233, + "mean_token_accuracy": 0.6762693524360657, + "num_tokens": 22417436.0, + "step": 867 + }, + { + "epoch": 0.09532176586865802, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.710785388946533, + "learning_rate": 3.173499267935578e-07, + "loss": 1.1508, + "mean_token_accuracy": 0.6636658906936646, + "num_tokens": 22440356.0, + "step": 868 + }, + { + "epoch": 0.09543158357127168, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.7830514907836914, + "learning_rate": 3.1771595900439236e-07, + "loss": 1.146, + "mean_token_accuracy": 0.6725117564201355, + "num_tokens": 22462445.0, + "step": 869 + }, + { + "epoch": 0.09554140127388536, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.170642852783203, + "learning_rate": 3.1808199121522697e-07, + "loss": 1.1419, + "mean_token_accuracy": 0.6690654754638672, + "num_tokens": 22495714.0, + "step": 870 + }, + { + "epoch": 0.09565121897649902, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4685871601104736, + "learning_rate": 3.1844802342606146e-07, + "loss": 1.1269, + "mean_token_accuracy": 0.6686489582061768, + "num_tokens": 22520890.0, + "step": 871 + }, + { + "epoch": 0.09576103667911268, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.25644588470459, + "learning_rate": 3.18814055636896e-07, + "loss": 1.0989, + "mean_token_accuracy": 0.683949887752533, + "num_tokens": 22548125.0, + "step": 872 + }, + { + "epoch": 0.09587085438172634, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.561833143234253, + "learning_rate": 3.191800878477306e-07, + "loss": 1.0441, + "mean_token_accuracy": 0.6961917281150818, + "num_tokens": 22571203.0, + "step": 873 + }, + { + "epoch": 0.09598067208434, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3100991249084473, + "learning_rate": 3.195461200585651e-07, + "loss": 1.0686, + "mean_token_accuracy": 0.6802868247032166, + "num_tokens": 22600078.0, + "step": 874 + }, + { + "epoch": 0.09609048978695366, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5133092403411865, + "learning_rate": 3.199121522693997e-07, + "loss": 1.1093, + "mean_token_accuracy": 0.6812517642974854, + "num_tokens": 22626236.0, + "step": 875 + }, + { + "epoch": 0.09620030748956732, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2867770195007324, + "learning_rate": 3.2027818448023426e-07, + "loss": 1.0968, + "mean_token_accuracy": 0.6740034818649292, + "num_tokens": 22653233.0, + "step": 876 + }, + { + "epoch": 0.09631012519218098, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3667948246002197, + "learning_rate": 3.2064421669106876e-07, + "loss": 1.1142, + "mean_token_accuracy": 0.6699889302253723, + "num_tokens": 22679272.0, + "step": 877 + }, + { + "epoch": 0.09641994289479464, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.507004499435425, + "learning_rate": 3.2101024890190336e-07, + "loss": 1.0517, + "mean_token_accuracy": 0.690585732460022, + "num_tokens": 22704003.0, + "step": 878 + }, + { + "epoch": 0.0965297605974083, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.455615758895874, + "learning_rate": 3.213762811127379e-07, + "loss": 1.0378, + "mean_token_accuracy": 0.6945575475692749, + "num_tokens": 22729119.0, + "step": 879 + }, + { + "epoch": 0.09663957830002197, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3820788860321045, + "learning_rate": 3.2174231332357246e-07, + "loss": 1.1159, + "mean_token_accuracy": 0.6740525960922241, + "num_tokens": 22755678.0, + "step": 880 + }, + { + "epoch": 0.09674939600263563, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1118111610412598, + "learning_rate": 3.22108345534407e-07, + "loss": 1.0925, + "mean_token_accuracy": 0.6798688173294067, + "num_tokens": 22788600.0, + "step": 881 + }, + { + "epoch": 0.09685921370524929, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.286360502243042, + "learning_rate": 3.2247437774524155e-07, + "loss": 1.0144, + "mean_token_accuracy": 0.6978654861450195, + "num_tokens": 22814636.0, + "step": 882 + }, + { + "epoch": 0.09696903140786295, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.599109649658203, + "learning_rate": 3.2284040995607616e-07, + "loss": 1.1283, + "mean_token_accuracy": 0.6703022122383118, + "num_tokens": 22839289.0, + "step": 883 + }, + { + "epoch": 0.09707884911047661, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.456636667251587, + "learning_rate": 3.2320644216691065e-07, + "loss": 1.0783, + "mean_token_accuracy": 0.6779966354370117, + "num_tokens": 22864607.0, + "step": 884 + }, + { + "epoch": 0.09718866681309027, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6046347618103027, + "learning_rate": 3.235724743777452e-07, + "loss": 1.1279, + "mean_token_accuracy": 0.6814446449279785, + "num_tokens": 22892406.0, + "step": 885 + }, + { + "epoch": 0.09729848451570393, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.525146245956421, + "learning_rate": 3.239385065885798e-07, + "loss": 1.1019, + "mean_token_accuracy": 0.6727612018585205, + "num_tokens": 22917563.0, + "step": 886 + }, + { + "epoch": 0.09740830221831759, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.350694179534912, + "learning_rate": 3.2430453879941435e-07, + "loss": 1.0781, + "mean_token_accuracy": 0.6822987794876099, + "num_tokens": 22944690.0, + "step": 887 + }, + { + "epoch": 0.09751811992093125, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.526992082595825, + "learning_rate": 3.2467057101024885e-07, + "loss": 1.0665, + "mean_token_accuracy": 0.6862941980361938, + "num_tokens": 22970435.0, + "step": 888 + }, + { + "epoch": 0.09762793762354492, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.40901255607605, + "learning_rate": 3.2503660322108345e-07, + "loss": 1.0367, + "mean_token_accuracy": 0.6954866051673889, + "num_tokens": 22996469.0, + "step": 889 + }, + { + "epoch": 0.09773775532615858, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.605721950531006, + "learning_rate": 3.25402635431918e-07, + "loss": 1.1584, + "mean_token_accuracy": 0.6700096726417542, + "num_tokens": 23019491.0, + "step": 890 + }, + { + "epoch": 0.09784757302877224, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2608566284179688, + "learning_rate": 3.2576866764275255e-07, + "loss": 1.076, + "mean_token_accuracy": 0.6881059408187866, + "num_tokens": 23048481.0, + "step": 891 + }, + { + "epoch": 0.0979573907313859, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.459350347518921, + "learning_rate": 3.261346998535871e-07, + "loss": 1.0782, + "mean_token_accuracy": 0.6812594532966614, + "num_tokens": 23074453.0, + "step": 892 + }, + { + "epoch": 0.09806720843399956, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.211095094680786, + "learning_rate": 3.2650073206442165e-07, + "loss": 0.989, + "mean_token_accuracy": 0.7053597569465637, + "num_tokens": 23103151.0, + "step": 893 + }, + { + "epoch": 0.09817702613661322, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1891908645629883, + "learning_rate": 3.2686676427525625e-07, + "loss": 1.0994, + "mean_token_accuracy": 0.6748390793800354, + "num_tokens": 23134289.0, + "step": 894 + }, + { + "epoch": 0.09828684383922688, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6461098194122314, + "learning_rate": 3.2723279648609075e-07, + "loss": 1.0288, + "mean_token_accuracy": 0.7006077766418457, + "num_tokens": 23156101.0, + "step": 895 + }, + { + "epoch": 0.09839666154184054, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5144894123077393, + "learning_rate": 3.275988286969253e-07, + "loss": 1.0026, + "mean_token_accuracy": 0.700435996055603, + "num_tokens": 23179178.0, + "step": 896 + }, + { + "epoch": 0.0985064792444542, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6402647495269775, + "learning_rate": 3.279648609077599e-07, + "loss": 1.1031, + "mean_token_accuracy": 0.6776789426803589, + "num_tokens": 23200455.0, + "step": 897 + }, + { + "epoch": 0.09861629694706786, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4716787338256836, + "learning_rate": 3.283308931185944e-07, + "loss": 1.0961, + "mean_token_accuracy": 0.6930606365203857, + "num_tokens": 23226237.0, + "step": 898 + }, + { + "epoch": 0.09872611464968153, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3220200538635254, + "learning_rate": 3.28696925329429e-07, + "loss": 1.0831, + "mean_token_accuracy": 0.6874324679374695, + "num_tokens": 23251813.0, + "step": 899 + }, + { + "epoch": 0.0988359323522952, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.350649118423462, + "learning_rate": 3.2906295754026354e-07, + "loss": 1.146, + "mean_token_accuracy": 0.6590631008148193, + "num_tokens": 23278810.0, + "step": 900 + }, + { + "epoch": 0.09894575005490885, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.215975522994995, + "learning_rate": 3.2942898975109804e-07, + "loss": 1.1325, + "mean_token_accuracy": 0.6728538274765015, + "num_tokens": 23307162.0, + "step": 901 + }, + { + "epoch": 0.09905556775752251, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2411434650421143, + "learning_rate": 3.2979502196193264e-07, + "loss": 1.144, + "mean_token_accuracy": 0.665176272392273, + "num_tokens": 23335308.0, + "step": 902 + }, + { + "epoch": 0.09916538546013617, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.473494052886963, + "learning_rate": 3.301610541727672e-07, + "loss": 1.0447, + "mean_token_accuracy": 0.6939002275466919, + "num_tokens": 23359068.0, + "step": 903 + }, + { + "epoch": 0.09927520316274983, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4793832302093506, + "learning_rate": 3.3052708638360174e-07, + "loss": 1.0084, + "mean_token_accuracy": 0.7060315012931824, + "num_tokens": 23381987.0, + "step": 904 + }, + { + "epoch": 0.0993850208653635, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9891853332519531, + "learning_rate": 3.308931185944363e-07, + "loss": 1.1214, + "mean_token_accuracy": 0.6716824769973755, + "num_tokens": 23420367.0, + "step": 905 + }, + { + "epoch": 0.09949483856797715, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.637824296951294, + "learning_rate": 3.3125915080527084e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.688784122467041, + "num_tokens": 23442587.0, + "step": 906 + }, + { + "epoch": 0.09960465627059081, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3905627727508545, + "learning_rate": 3.3162518301610544e-07, + "loss": 1.1289, + "mean_token_accuracy": 0.6739425659179688, + "num_tokens": 23469268.0, + "step": 907 + }, + { + "epoch": 0.09971447397320449, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.180527925491333, + "learning_rate": 3.3199121522693994e-07, + "loss": 1.1906, + "mean_token_accuracy": 0.6459702253341675, + "num_tokens": 23500886.0, + "step": 908 + }, + { + "epoch": 0.09982429167581815, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6311495304107666, + "learning_rate": 3.323572474377745e-07, + "loss": 1.0929, + "mean_token_accuracy": 0.6823486685752869, + "num_tokens": 23523429.0, + "step": 909 + }, + { + "epoch": 0.0999341093784318, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5105810165405273, + "learning_rate": 3.327232796486091e-07, + "loss": 1.042, + "mean_token_accuracy": 0.6976488828659058, + "num_tokens": 23545940.0, + "step": 910 + }, + { + "epoch": 0.10004392708104547, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5306973457336426, + "learning_rate": 3.3308931185944364e-07, + "loss": 1.1862, + "mean_token_accuracy": 0.6596856117248535, + "num_tokens": 23572000.0, + "step": 911 + }, + { + "epoch": 0.10015374478365913, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3908777236938477, + "learning_rate": 3.3345534407027813e-07, + "loss": 1.1095, + "mean_token_accuracy": 0.6782066822052002, + "num_tokens": 23600798.0, + "step": 912 + }, + { + "epoch": 0.10026356248627279, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4930169582366943, + "learning_rate": 3.3382137628111273e-07, + "loss": 1.0239, + "mean_token_accuracy": 0.6955089569091797, + "num_tokens": 23624264.0, + "step": 913 + }, + { + "epoch": 0.10037338018888645, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1533265113830566, + "learning_rate": 3.341874084919473e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.6848756074905396, + "num_tokens": 23655171.0, + "step": 914 + }, + { + "epoch": 0.1004831978915001, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2580807209014893, + "learning_rate": 3.3455344070278183e-07, + "loss": 1.1238, + "mean_token_accuracy": 0.6720633506774902, + "num_tokens": 23683903.0, + "step": 915 + }, + { + "epoch": 0.10059301559411377, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.540973424911499, + "learning_rate": 3.349194729136164e-07, + "loss": 1.058, + "mean_token_accuracy": 0.692379355430603, + "num_tokens": 23706406.0, + "step": 916 + }, + { + "epoch": 0.10070283329672743, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3901615142822266, + "learning_rate": 3.3528550512445093e-07, + "loss": 1.1266, + "mean_token_accuracy": 0.6671941876411438, + "num_tokens": 23732433.0, + "step": 917 + }, + { + "epoch": 0.1008126509993411, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2078216075897217, + "learning_rate": 3.356515373352855e-07, + "loss": 1.0317, + "mean_token_accuracy": 0.692297101020813, + "num_tokens": 23759236.0, + "step": 918 + }, + { + "epoch": 0.10092246870195476, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4500787258148193, + "learning_rate": 3.3601756954612003e-07, + "loss": 1.0222, + "mean_token_accuracy": 0.698104977607727, + "num_tokens": 23784398.0, + "step": 919 + }, + { + "epoch": 0.10103228640456842, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.337333917617798, + "learning_rate": 3.363836017569546e-07, + "loss": 1.0053, + "mean_token_accuracy": 0.6993992328643799, + "num_tokens": 23812650.0, + "step": 920 + }, + { + "epoch": 0.10114210410718208, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.188096761703491, + "learning_rate": 3.367496339677892e-07, + "loss": 1.093, + "mean_token_accuracy": 0.680715799331665, + "num_tokens": 23841776.0, + "step": 921 + }, + { + "epoch": 0.10125192180979574, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.5316920280456543, + "learning_rate": 3.371156661786237e-07, + "loss": 1.0788, + "mean_token_accuracy": 0.6778004169464111, + "num_tokens": 23865229.0, + "step": 922 + }, + { + "epoch": 0.1013617395124094, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.4319608211517334, + "learning_rate": 3.374816983894583e-07, + "loss": 1.0825, + "mean_token_accuracy": 0.6746349930763245, + "num_tokens": 23889843.0, + "step": 923 + }, + { + "epoch": 0.10147155721502306, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.4227895736694336, + "learning_rate": 3.3784773060029283e-07, + "loss": 1.161, + "mean_token_accuracy": 0.6650419235229492, + "num_tokens": 23917810.0, + "step": 924 + }, + { + "epoch": 0.10158137491763672, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.6551148891448975, + "learning_rate": 3.382137628111273e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.679029643535614, + "num_tokens": 23940531.0, + "step": 925 + }, + { + "epoch": 0.10169119262025038, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.2656002044677734, + "learning_rate": 3.385797950219619e-07, + "loss": 1.0743, + "mean_token_accuracy": 0.6824926137924194, + "num_tokens": 23968313.0, + "step": 926 + }, + { + "epoch": 0.10180101032286405, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.240962028503418, + "learning_rate": 3.389458272327965e-07, + "loss": 1.1177, + "mean_token_accuracy": 0.6758521795272827, + "num_tokens": 23999841.0, + "step": 927 + }, + { + "epoch": 0.10191082802547771, + "ewc_loss": 4.1425228118896484e-06, + "grad_norm": 2.1334927082061768, + "learning_rate": 3.39311859443631e-07, + "loss": 1.1396, + "mean_token_accuracy": 0.6639518737792969, + "num_tokens": 24031207.0, + "step": 928 + }, + { + "epoch": 0.10202064572809137, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.167400598526001, + "learning_rate": 3.3967789165446557e-07, + "loss": 1.1252, + "mean_token_accuracy": 0.6723071336746216, + "num_tokens": 24060419.0, + "step": 929 + }, + { + "epoch": 0.10213046343070503, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3052496910095215, + "learning_rate": 3.400439238653001e-07, + "loss": 1.0917, + "mean_token_accuracy": 0.6859890222549438, + "num_tokens": 24086752.0, + "step": 930 + }, + { + "epoch": 0.10224028113331869, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.502917766571045, + "learning_rate": 3.404099560761347e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.6874755024909973, + "num_tokens": 24110671.0, + "step": 931 + }, + { + "epoch": 0.10235009883593235, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.2221457958221436, + "learning_rate": 3.407759882869692e-07, + "loss": 1.0852, + "mean_token_accuracy": 0.6784112453460693, + "num_tokens": 24142021.0, + "step": 932 + }, + { + "epoch": 0.10245991653854601, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.764949321746826, + "learning_rate": 3.4114202049780377e-07, + "loss": 1.0894, + "mean_token_accuracy": 0.6792221069335938, + "num_tokens": 24163153.0, + "step": 933 + }, + { + "epoch": 0.10256973424115967, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1997032165527344, + "learning_rate": 3.4150805270863837e-07, + "loss": 1.1532, + "mean_token_accuracy": 0.6620455980300903, + "num_tokens": 24194395.0, + "step": 934 + }, + { + "epoch": 0.10267955194377333, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3425416946411133, + "learning_rate": 3.4187408491947287e-07, + "loss": 1.1373, + "mean_token_accuracy": 0.6649024486541748, + "num_tokens": 24220717.0, + "step": 935 + }, + { + "epoch": 0.10278936964638699, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.438521146774292, + "learning_rate": 3.422401171303074e-07, + "loss": 1.0988, + "mean_token_accuracy": 0.6820824146270752, + "num_tokens": 24245109.0, + "step": 936 + }, + { + "epoch": 0.10289918734900066, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.848909378051758, + "learning_rate": 3.42606149341142e-07, + "loss": 1.1676, + "mean_token_accuracy": 0.6677666902542114, + "num_tokens": 24264983.0, + "step": 937 + }, + { + "epoch": 0.10300900505161432, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.579694986343384, + "learning_rate": 3.4297218155197657e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6884739398956299, + "num_tokens": 24288101.0, + "step": 938 + }, + { + "epoch": 0.10311882275422798, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2474477291107178, + "learning_rate": 3.433382137628111e-07, + "loss": 1.0604, + "mean_token_accuracy": 0.6856790781021118, + "num_tokens": 24314589.0, + "step": 939 + }, + { + "epoch": 0.10322864045684164, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.579615592956543, + "learning_rate": 3.4370424597364567e-07, + "loss": 1.1418, + "mean_token_accuracy": 0.6753277778625488, + "num_tokens": 24337109.0, + "step": 940 + }, + { + "epoch": 0.1033384581594553, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.254171848297119, + "learning_rate": 3.440702781844802e-07, + "loss": 1.1283, + "mean_token_accuracy": 0.6746256351470947, + "num_tokens": 24367221.0, + "step": 941 + }, + { + "epoch": 0.10344827586206896, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3492496013641357, + "learning_rate": 3.4443631039531476e-07, + "loss": 1.1477, + "mean_token_accuracy": 0.6640753746032715, + "num_tokens": 24394210.0, + "step": 942 + }, + { + "epoch": 0.10355809356468262, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.35190749168396, + "learning_rate": 3.448023426061493e-07, + "loss": 1.0136, + "mean_token_accuracy": 0.695926308631897, + "num_tokens": 24419005.0, + "step": 943 + }, + { + "epoch": 0.10366791126729628, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.261080265045166, + "learning_rate": 3.4516837481698386e-07, + "loss": 1.0966, + "mean_token_accuracy": 0.6803848147392273, + "num_tokens": 24448427.0, + "step": 944 + }, + { + "epoch": 0.10377772896990994, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1435272693634033, + "learning_rate": 3.4553440702781846e-07, + "loss": 1.201, + "mean_token_accuracy": 0.6516863703727722, + "num_tokens": 24480209.0, + "step": 945 + }, + { + "epoch": 0.10388754667252362, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.5358200073242188, + "learning_rate": 3.4590043923865296e-07, + "loss": 1.0264, + "mean_token_accuracy": 0.6966376900672913, + "num_tokens": 24503484.0, + "step": 946 + }, + { + "epoch": 0.10399736437513728, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3932273387908936, + "learning_rate": 3.4626647144948756e-07, + "loss": 1.1009, + "mean_token_accuracy": 0.6735131740570068, + "num_tokens": 24528827.0, + "step": 947 + }, + { + "epoch": 0.10410718207775094, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.7312991619110107, + "learning_rate": 3.466325036603221e-07, + "loss": 1.0336, + "mean_token_accuracy": 0.697418749332428, + "num_tokens": 24550278.0, + "step": 948 + }, + { + "epoch": 0.1042169997803646, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.4458608627319336, + "learning_rate": 3.469985358711566e-07, + "loss": 1.0061, + "mean_token_accuracy": 0.697100818157196, + "num_tokens": 24573147.0, + "step": 949 + }, + { + "epoch": 0.10432681748297826, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.509629726409912, + "learning_rate": 3.473645680819912e-07, + "loss": 1.0734, + "mean_token_accuracy": 0.676019012928009, + "num_tokens": 24597279.0, + "step": 950 + }, + { + "epoch": 0.10443663518559192, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.515599012374878, + "learning_rate": 3.4773060029282576e-07, + "loss": 1.1226, + "mean_token_accuracy": 0.6617810726165771, + "num_tokens": 24623328.0, + "step": 951 + }, + { + "epoch": 0.10454645288820558, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.730372190475464, + "learning_rate": 3.480966325036603e-07, + "loss": 1.0707, + "mean_token_accuracy": 0.6941739320755005, + "num_tokens": 24644301.0, + "step": 952 + }, + { + "epoch": 0.10465627059081924, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0701606273651123, + "learning_rate": 3.4846266471449486e-07, + "loss": 1.1344, + "mean_token_accuracy": 0.6696408987045288, + "num_tokens": 24678615.0, + "step": 953 + }, + { + "epoch": 0.1047660882934329, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.59839129447937, + "learning_rate": 3.488286969253294e-07, + "loss": 0.9736, + "mean_token_accuracy": 0.7071595788002014, + "num_tokens": 24701336.0, + "step": 954 + }, + { + "epoch": 0.10487590599604656, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2902050018310547, + "learning_rate": 3.49194729136164e-07, + "loss": 1.0319, + "mean_token_accuracy": 0.6941426396369934, + "num_tokens": 24727917.0, + "step": 955 + }, + { + "epoch": 0.10498572369866023, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.724726676940918, + "learning_rate": 3.495607613469985e-07, + "loss": 1.0262, + "mean_token_accuracy": 0.6911413669586182, + "num_tokens": 24748147.0, + "step": 956 + }, + { + "epoch": 0.10509554140127389, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.4413013458251953, + "learning_rate": 3.4992679355783305e-07, + "loss": 1.0655, + "mean_token_accuracy": 0.6889024972915649, + "num_tokens": 24773384.0, + "step": 957 + }, + { + "epoch": 0.10520535910388755, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2824814319610596, + "learning_rate": 3.5029282576866766e-07, + "loss": 1.1, + "mean_token_accuracy": 0.6738194227218628, + "num_tokens": 24801521.0, + "step": 958 + }, + { + "epoch": 0.10531517680650121, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.302248001098633, + "learning_rate": 3.5065885797950215e-07, + "loss": 1.1217, + "mean_token_accuracy": 0.6745860576629639, + "num_tokens": 24830015.0, + "step": 959 + }, + { + "epoch": 0.10542499450911487, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.5803942680358887, + "learning_rate": 3.510248901903367e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6769839525222778, + "num_tokens": 24853232.0, + "step": 960 + }, + { + "epoch": 0.10553481221172853, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.4446845054626465, + "learning_rate": 3.513909224011713e-07, + "loss": 1.1425, + "mean_token_accuracy": 0.6687743663787842, + "num_tokens": 24877653.0, + "step": 961 + }, + { + "epoch": 0.10564462991434219, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.4589405059814453, + "learning_rate": 3.5175695461200585e-07, + "loss": 1.124, + "mean_token_accuracy": 0.6699550151824951, + "num_tokens": 24903434.0, + "step": 962 + }, + { + "epoch": 0.10575444761695585, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.407829999923706, + "learning_rate": 3.521229868228404e-07, + "loss": 1.1589, + "mean_token_accuracy": 0.668522834777832, + "num_tokens": 24931347.0, + "step": 963 + }, + { + "epoch": 0.10586426531956951, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.6010754108428955, + "learning_rate": 3.5248901903367495e-07, + "loss": 1.0594, + "mean_token_accuracy": 0.689085841178894, + "num_tokens": 24955365.0, + "step": 964 + }, + { + "epoch": 0.10597408302218318, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.665177345275879, + "learning_rate": 3.528550512445095e-07, + "loss": 1.1472, + "mean_token_accuracy": 0.6630651950836182, + "num_tokens": 24977249.0, + "step": 965 + }, + { + "epoch": 0.10608390072479684, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.220764636993408, + "learning_rate": 3.5322108345534405e-07, + "loss": 1.158, + "mean_token_accuracy": 0.6659284830093384, + "num_tokens": 25008578.0, + "step": 966 + }, + { + "epoch": 0.1061937184274105, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.7172327041625977, + "learning_rate": 3.535871156661786e-07, + "loss": 1.0207, + "mean_token_accuracy": 0.6904747486114502, + "num_tokens": 25028258.0, + "step": 967 + }, + { + "epoch": 0.10630353613002416, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.4978179931640625, + "learning_rate": 3.5395314787701315e-07, + "loss": 1.0497, + "mean_token_accuracy": 0.6895196437835693, + "num_tokens": 25052108.0, + "step": 968 + }, + { + "epoch": 0.10641335383263782, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0381412506103516, + "learning_rate": 3.5431918008784775e-07, + "loss": 1.1113, + "mean_token_accuracy": 0.6775628328323364, + "num_tokens": 25085379.0, + "step": 969 + }, + { + "epoch": 0.10652317153525148, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.561109781265259, + "learning_rate": 3.5468521229868224e-07, + "loss": 1.0892, + "mean_token_accuracy": 0.681745707988739, + "num_tokens": 25108319.0, + "step": 970 + }, + { + "epoch": 0.10663298923786514, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.284475803375244, + "learning_rate": 3.5505124450951685e-07, + "loss": 1.1965, + "mean_token_accuracy": 0.6574986577033997, + "num_tokens": 25139744.0, + "step": 971 + }, + { + "epoch": 0.1067428069404788, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3301937580108643, + "learning_rate": 3.554172767203514e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6809495091438293, + "num_tokens": 25166647.0, + "step": 972 + }, + { + "epoch": 0.10685262464309246, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.464287519454956, + "learning_rate": 3.557833089311859e-07, + "loss": 1.0639, + "mean_token_accuracy": 0.6880868673324585, + "num_tokens": 25195503.0, + "step": 973 + }, + { + "epoch": 0.10696244234570612, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.45564341545105, + "learning_rate": 3.561493411420205e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6918039917945862, + "num_tokens": 25220138.0, + "step": 974 + }, + { + "epoch": 0.1070722600483198, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.449021100997925, + "learning_rate": 3.5651537335285504e-07, + "loss": 1.0583, + "mean_token_accuracy": 0.6847231388092041, + "num_tokens": 25244278.0, + "step": 975 + }, + { + "epoch": 0.10718207775093345, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.158281087875366, + "learning_rate": 3.5688140556368954e-07, + "loss": 1.0975, + "mean_token_accuracy": 0.6706647872924805, + "num_tokens": 25275205.0, + "step": 976 + }, + { + "epoch": 0.10729189545354711, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.611358642578125, + "learning_rate": 3.5724743777452414e-07, + "loss": 1.0337, + "mean_token_accuracy": 0.6912858486175537, + "num_tokens": 25296551.0, + "step": 977 + }, + { + "epoch": 0.10740171315616077, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.498448133468628, + "learning_rate": 3.576134699853587e-07, + "loss": 0.9434, + "mean_token_accuracy": 0.718456506729126, + "num_tokens": 25319702.0, + "step": 978 + }, + { + "epoch": 0.10751153085877443, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.4001266956329346, + "learning_rate": 3.579795021961933e-07, + "loss": 1.1334, + "mean_token_accuracy": 0.6674226522445679, + "num_tokens": 25345390.0, + "step": 979 + }, + { + "epoch": 0.1076213485613881, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.364147186279297, + "learning_rate": 3.583455344070278e-07, + "loss": 1.0585, + "mean_token_accuracy": 0.6848574876785278, + "num_tokens": 25371403.0, + "step": 980 + }, + { + "epoch": 0.10773116626400175, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.7647101879119873, + "learning_rate": 3.5871156661786234e-07, + "loss": 1.0104, + "mean_token_accuracy": 0.696853756904602, + "num_tokens": 25389771.0, + "step": 981 + }, + { + "epoch": 0.10784098396661541, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.624602794647217, + "learning_rate": 3.5907759882869694e-07, + "loss": 1.0459, + "mean_token_accuracy": 0.6874682903289795, + "num_tokens": 25410123.0, + "step": 982 + }, + { + "epoch": 0.10795080166922907, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.703233242034912, + "learning_rate": 3.5944363103953144e-07, + "loss": 1.0567, + "mean_token_accuracy": 0.6824468970298767, + "num_tokens": 25432542.0, + "step": 983 + }, + { + "epoch": 0.10806061937184275, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.349255323410034, + "learning_rate": 3.59809663250366e-07, + "loss": 1.1062, + "mean_token_accuracy": 0.671094536781311, + "num_tokens": 25458481.0, + "step": 984 + }, + { + "epoch": 0.10817043707445641, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.4069020748138428, + "learning_rate": 3.601756954612006e-07, + "loss": 1.024, + "mean_token_accuracy": 0.7012657523155212, + "num_tokens": 25483261.0, + "step": 985 + }, + { + "epoch": 0.10828025477707007, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.3729000091552734, + "learning_rate": 3.6054172767203514e-07, + "loss": 1.0268, + "mean_token_accuracy": 0.6979021430015564, + "num_tokens": 25511391.0, + "step": 986 + }, + { + "epoch": 0.10839007247968373, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1167333126068115, + "learning_rate": 3.609077598828697e-07, + "loss": 1.1026, + "mean_token_accuracy": 0.6758627891540527, + "num_tokens": 25541179.0, + "step": 987 + }, + { + "epoch": 0.10849989018229739, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.557624340057373, + "learning_rate": 3.6127379209370423e-07, + "loss": 1.0127, + "mean_token_accuracy": 0.7007427215576172, + "num_tokens": 25564057.0, + "step": 988 + }, + { + "epoch": 0.10860970788491105, + "ewc_loss": 4.4405460357666016e-06, + "grad_norm": 2.357776165008545, + "learning_rate": 3.616398243045388e-07, + "loss": 1.0504, + "mean_token_accuracy": 0.6952011585235596, + "num_tokens": 25590617.0, + "step": 989 + }, + { + "epoch": 0.1087195255875247, + "ewc_loss": 4.4405460357666016e-06, + "grad_norm": 2.8732943534851074, + "learning_rate": 3.6200585651537333e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6898860931396484, + "num_tokens": 25610733.0, + "step": 990 + }, + { + "epoch": 0.10882934329013837, + "ewc_loss": 4.4405460357666016e-06, + "grad_norm": 2.4690492153167725, + "learning_rate": 3.623718887262079e-07, + "loss": 1.0848, + "mean_token_accuracy": 0.6807260513305664, + "num_tokens": 25636827.0, + "step": 991 + }, + { + "epoch": 0.10893916099275203, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.5698251724243164, + "learning_rate": 3.6273792093704243e-07, + "loss": 1.1178, + "mean_token_accuracy": 0.6758798360824585, + "num_tokens": 25659218.0, + "step": 992 + }, + { + "epoch": 0.10904897869536569, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.5729217529296875, + "learning_rate": 3.6310395314787703e-07, + "loss": 1.0711, + "mean_token_accuracy": 0.6827998757362366, + "num_tokens": 25682816.0, + "step": 993 + }, + { + "epoch": 0.10915879639797936, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.172579288482666, + "learning_rate": 3.6346998535871153e-07, + "loss": 1.0911, + "mean_token_accuracy": 0.6790474057197571, + "num_tokens": 25712238.0, + "step": 994 + }, + { + "epoch": 0.10926861410059302, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.391098976135254, + "learning_rate": 3.6383601756954613e-07, + "loss": 0.9853, + "mean_token_accuracy": 0.697907567024231, + "num_tokens": 25737964.0, + "step": 995 + }, + { + "epoch": 0.10937843180320668, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.275111675262451, + "learning_rate": 3.642020497803807e-07, + "loss": 1.0117, + "mean_token_accuracy": 0.6985941529273987, + "num_tokens": 25765993.0, + "step": 996 + }, + { + "epoch": 0.10948824950582034, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.6196341514587402, + "learning_rate": 3.645680819912152e-07, + "loss": 1.0563, + "mean_token_accuracy": 0.6894100904464722, + "num_tokens": 25789007.0, + "step": 997 + }, + { + "epoch": 0.109598067208434, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.372011661529541, + "learning_rate": 3.649341142020498e-07, + "loss": 0.9868, + "mean_token_accuracy": 0.7064016461372375, + "num_tokens": 25814004.0, + "step": 998 + }, + { + "epoch": 0.10970788491104766, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.626072645187378, + "learning_rate": 3.653001464128843e-07, + "loss": 1.0605, + "mean_token_accuracy": 0.6876218318939209, + "num_tokens": 25835515.0, + "step": 999 + }, + { + "epoch": 0.10981770261366132, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.481210708618164, + "learning_rate": 3.656661786237188e-07, + "loss": 1.1008, + "mean_token_accuracy": 0.6818015575408936, + "num_tokens": 25859300.0, + "step": 1000 + }, + { + "epoch": 0.10992752031627498, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.3214004039764404, + "learning_rate": 3.660322108345534e-07, + "loss": 1.0384, + "mean_token_accuracy": 0.6898066997528076, + "num_tokens": 25888053.0, + "step": 1001 + }, + { + "epoch": 0.11003733801888864, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.4027583599090576, + "learning_rate": 3.66398243045388e-07, + "loss": 1.1539, + "mean_token_accuracy": 0.6605489253997803, + "num_tokens": 25913789.0, + "step": 1002 + }, + { + "epoch": 0.11014715572150231, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.562622547149658, + "learning_rate": 3.667642752562226e-07, + "loss": 1.1635, + "mean_token_accuracy": 0.6702340841293335, + "num_tokens": 25936587.0, + "step": 1003 + }, + { + "epoch": 0.11025697342411597, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1727049350738525, + "learning_rate": 3.6713030746705707e-07, + "loss": 1.048, + "mean_token_accuracy": 0.6892145872116089, + "num_tokens": 25965940.0, + "step": 1004 + }, + { + "epoch": 0.11036679112672963, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.261361837387085, + "learning_rate": 3.674963396778916e-07, + "loss": 0.9899, + "mean_token_accuracy": 0.7073208093643188, + "num_tokens": 25993188.0, + "step": 1005 + }, + { + "epoch": 0.11047660882934329, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.4495630264282227, + "learning_rate": 3.678623718887262e-07, + "loss": 1.0271, + "mean_token_accuracy": 0.6922038793563843, + "num_tokens": 26016160.0, + "step": 1006 + }, + { + "epoch": 0.11058642653195695, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.4422225952148438, + "learning_rate": 3.682284040995607e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.6898666620254517, + "num_tokens": 26039892.0, + "step": 1007 + }, + { + "epoch": 0.11069624423457061, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.566420793533325, + "learning_rate": 3.6859443631039527e-07, + "loss": 1.0678, + "mean_token_accuracy": 0.6826772093772888, + "num_tokens": 26061680.0, + "step": 1008 + }, + { + "epoch": 0.11080606193718427, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.3188443183898926, + "learning_rate": 3.6896046852122987e-07, + "loss": 1.1629, + "mean_token_accuracy": 0.6576874256134033, + "num_tokens": 26088811.0, + "step": 1009 + }, + { + "epoch": 0.11091587963979793, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.4931211471557617, + "learning_rate": 3.693265007320644e-07, + "loss": 1.1236, + "mean_token_accuracy": 0.6702043414115906, + "num_tokens": 26113462.0, + "step": 1010 + }, + { + "epoch": 0.11102569734241159, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.334768295288086, + "learning_rate": 3.6969253294289897e-07, + "loss": 1.15, + "mean_token_accuracy": 0.6638098955154419, + "num_tokens": 26140867.0, + "step": 1011 + }, + { + "epoch": 0.11113551504502525, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3115487098693848, + "learning_rate": 3.700585651537335e-07, + "loss": 1.0702, + "mean_token_accuracy": 0.6785274744033813, + "num_tokens": 26168121.0, + "step": 1012 + }, + { + "epoch": 0.11124533274763893, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.7236809730529785, + "learning_rate": 3.7042459736456807e-07, + "loss": 1.0296, + "mean_token_accuracy": 0.6908428072929382, + "num_tokens": 26189626.0, + "step": 1013 + }, + { + "epoch": 0.11135515045025259, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.5661351680755615, + "learning_rate": 3.707906295754026e-07, + "loss": 1.0989, + "mean_token_accuracy": 0.6768123507499695, + "num_tokens": 26213658.0, + "step": 1014 + }, + { + "epoch": 0.11146496815286625, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1691396236419678, + "learning_rate": 3.7115666178623716e-07, + "loss": 1.0387, + "mean_token_accuracy": 0.6895956993103027, + "num_tokens": 26243731.0, + "step": 1015 + }, + { + "epoch": 0.1115747858554799, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.6164121627807617, + "learning_rate": 3.715226939970717e-07, + "loss": 1.0613, + "mean_token_accuracy": 0.6831705570220947, + "num_tokens": 26266656.0, + "step": 1016 + }, + { + "epoch": 0.11168460355809356, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.9287335872650146, + "learning_rate": 3.7188872620790626e-07, + "loss": 1.0299, + "mean_token_accuracy": 0.6938362717628479, + "num_tokens": 26285162.0, + "step": 1017 + }, + { + "epoch": 0.11179442126070722, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.4814462661743164, + "learning_rate": 3.722547584187408e-07, + "loss": 1.0415, + "mean_token_accuracy": 0.6898465156555176, + "num_tokens": 26309229.0, + "step": 1018 + }, + { + "epoch": 0.11190423896332088, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.4329516887664795, + "learning_rate": 3.726207906295754e-07, + "loss": 1.1021, + "mean_token_accuracy": 0.6780632734298706, + "num_tokens": 26334702.0, + "step": 1019 + }, + { + "epoch": 0.11201405666593454, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.533963918685913, + "learning_rate": 3.7298682284040996e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6936767101287842, + "num_tokens": 26359954.0, + "step": 1020 + }, + { + "epoch": 0.1121238743685482, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.4207417964935303, + "learning_rate": 3.7335285505124446e-07, + "loss": 1.0198, + "mean_token_accuracy": 0.696880042552948, + "num_tokens": 26383642.0, + "step": 1021 + }, + { + "epoch": 0.11223369207116188, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1561245918273926, + "learning_rate": 3.7371888726207906e-07, + "loss": 1.0324, + "mean_token_accuracy": 0.6979791522026062, + "num_tokens": 26413718.0, + "step": 1022 + }, + { + "epoch": 0.11234350977377554, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0277204513549805, + "learning_rate": 3.740849194729136e-07, + "loss": 1.1481, + "mean_token_accuracy": 0.6667289733886719, + "num_tokens": 26448070.0, + "step": 1023 + }, + { + "epoch": 0.1124533274763892, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.261573314666748, + "learning_rate": 3.744509516837481e-07, + "loss": 1.1559, + "mean_token_accuracy": 0.66387540102005, + "num_tokens": 26477876.0, + "step": 1024 + }, + { + "epoch": 0.11256314517900286, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.487412214279175, + "learning_rate": 3.748169838945827e-07, + "loss": 1.1652, + "mean_token_accuracy": 0.6562126874923706, + "num_tokens": 26503463.0, + "step": 1025 + }, + { + "epoch": 0.11267296288161652, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.340391159057617, + "learning_rate": 3.7518301610541726e-07, + "loss": 1.063, + "mean_token_accuracy": 0.6857701539993286, + "num_tokens": 26530910.0, + "step": 1026 + }, + { + "epoch": 0.11278278058423018, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3300321102142334, + "learning_rate": 3.7554904831625186e-07, + "loss": 1.1273, + "mean_token_accuracy": 0.6656935214996338, + "num_tokens": 26556511.0, + "step": 1027 + }, + { + "epoch": 0.11289259828684384, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.719304084777832, + "learning_rate": 3.7591508052708636e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6736255288124084, + "num_tokens": 26577702.0, + "step": 1028 + }, + { + "epoch": 0.1130024159894575, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.430187702178955, + "learning_rate": 3.762811127379209e-07, + "loss": 1.156, + "mean_token_accuracy": 0.6696035265922546, + "num_tokens": 26602465.0, + "step": 1029 + }, + { + "epoch": 0.11311223369207116, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.7546164989471436, + "learning_rate": 3.766471449487555e-07, + "loss": 1.0556, + "mean_token_accuracy": 0.6836317777633667, + "num_tokens": 26622436.0, + "step": 1030 + }, + { + "epoch": 0.11322205139468482, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4126341342926025, + "learning_rate": 3.7701317715959e-07, + "loss": 1.0661, + "mean_token_accuracy": 0.6871599555015564, + "num_tokens": 26646386.0, + "step": 1031 + }, + { + "epoch": 0.11333186909729849, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.430412530899048, + "learning_rate": 3.7737920937042455e-07, + "loss": 1.1619, + "mean_token_accuracy": 0.6602365970611572, + "num_tokens": 26674495.0, + "step": 1032 + }, + { + "epoch": 0.11344168679991215, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2689003944396973, + "learning_rate": 3.7774524158125915e-07, + "loss": 1.0394, + "mean_token_accuracy": 0.6923834085464478, + "num_tokens": 26701395.0, + "step": 1033 + }, + { + "epoch": 0.11355150450252581, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.605912923812866, + "learning_rate": 3.781112737920937e-07, + "loss": 1.1267, + "mean_token_accuracy": 0.6770709753036499, + "num_tokens": 26724537.0, + "step": 1034 + }, + { + "epoch": 0.11366132220513947, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4556541442871094, + "learning_rate": 3.7847730600292825e-07, + "loss": 1.1026, + "mean_token_accuracy": 0.6782934665679932, + "num_tokens": 26747889.0, + "step": 1035 + }, + { + "epoch": 0.11377113990775313, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.949782371520996, + "learning_rate": 3.788433382137628e-07, + "loss": 1.0116, + "mean_token_accuracy": 0.6971799731254578, + "num_tokens": 26767707.0, + "step": 1036 + }, + { + "epoch": 0.11388095761036679, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.341477632522583, + "learning_rate": 3.7920937042459735e-07, + "loss": 1.128, + "mean_token_accuracy": 0.6811498403549194, + "num_tokens": 26793032.0, + "step": 1037 + }, + { + "epoch": 0.11399077531298045, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.4219627380371094, + "learning_rate": 3.795754026354319e-07, + "loss": 1.039, + "mean_token_accuracy": 0.6892315745353699, + "num_tokens": 26819051.0, + "step": 1038 + }, + { + "epoch": 0.11410059301559411, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.444431781768799, + "learning_rate": 3.7994143484626645e-07, + "loss": 1.0434, + "mean_token_accuracy": 0.6931326389312744, + "num_tokens": 26842616.0, + "step": 1039 + }, + { + "epoch": 0.11421041071820777, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.517549753189087, + "learning_rate": 3.80307467057101e-07, + "loss": 1.1135, + "mean_token_accuracy": 0.6828684210777283, + "num_tokens": 26865703.0, + "step": 1040 + }, + { + "epoch": 0.11432022842082144, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.8737316131591797, + "learning_rate": 3.8067349926793555e-07, + "loss": 1.1051, + "mean_token_accuracy": 0.6686022877693176, + "num_tokens": 26884984.0, + "step": 1041 + }, + { + "epoch": 0.1144300461234351, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.3375773429870605, + "learning_rate": 3.810395314787701e-07, + "loss": 1.0821, + "mean_token_accuracy": 0.6832998991012573, + "num_tokens": 26912097.0, + "step": 1042 + }, + { + "epoch": 0.11453986382604876, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.7273433208465576, + "learning_rate": 3.814055636896047e-07, + "loss": 1.0163, + "mean_token_accuracy": 0.6983283758163452, + "num_tokens": 26932033.0, + "step": 1043 + }, + { + "epoch": 0.11464968152866242, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.5329091548919678, + "learning_rate": 3.8177159590043925e-07, + "loss": 1.1009, + "mean_token_accuracy": 0.6752402186393738, + "num_tokens": 26954418.0, + "step": 1044 + }, + { + "epoch": 0.11475949923127608, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.14396071434021, + "learning_rate": 3.8213762811127374e-07, + "loss": 1.0653, + "mean_token_accuracy": 0.6864652633666992, + "num_tokens": 26984406.0, + "step": 1045 + }, + { + "epoch": 0.11486931693388974, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.381382942199707, + "learning_rate": 3.8250366032210835e-07, + "loss": 1.0654, + "mean_token_accuracy": 0.689302921295166, + "num_tokens": 27012181.0, + "step": 1046 + }, + { + "epoch": 0.1149791346365034, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.226778030395508, + "learning_rate": 3.828696925329429e-07, + "loss": 1.09, + "mean_token_accuracy": 0.6780107021331787, + "num_tokens": 27040817.0, + "step": 1047 + }, + { + "epoch": 0.11508895233911706, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.4750478267669678, + "learning_rate": 3.832357247437774e-07, + "loss": 1.0026, + "mean_token_accuracy": 0.7110284566879272, + "num_tokens": 27063531.0, + "step": 1048 + }, + { + "epoch": 0.11519877004173072, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.5072314739227295, + "learning_rate": 3.83601756954612e-07, + "loss": 1.0399, + "mean_token_accuracy": 0.6869692206382751, + "num_tokens": 27085193.0, + "step": 1049 + }, + { + "epoch": 0.11530858774434438, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.297799825668335, + "learning_rate": 3.8396778916544654e-07, + "loss": 1.084, + "mean_token_accuracy": 0.6748404502868652, + "num_tokens": 27112026.0, + "step": 1050 + }, + { + "epoch": 0.11541840544695806, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.561458110809326, + "learning_rate": 3.8433382137628114e-07, + "loss": 1.0801, + "mean_token_accuracy": 0.685985267162323, + "num_tokens": 27134736.0, + "step": 1051 + }, + { + "epoch": 0.11552822314957172, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.10457706451416, + "learning_rate": 3.8469985358711564e-07, + "loss": 1.1912, + "mean_token_accuracy": 0.6513544917106628, + "num_tokens": 27167154.0, + "step": 1052 + }, + { + "epoch": 0.11563804085218538, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.346928119659424, + "learning_rate": 3.850658857979502e-07, + "loss": 1.0511, + "mean_token_accuracy": 0.6927826404571533, + "num_tokens": 27191779.0, + "step": 1053 + }, + { + "epoch": 0.11574785855479904, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.670595645904541, + "learning_rate": 3.854319180087848e-07, + "loss": 1.0675, + "mean_token_accuracy": 0.6809955835342407, + "num_tokens": 27214207.0, + "step": 1054 + }, + { + "epoch": 0.1158576762574127, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.3964169025421143, + "learning_rate": 3.857979502196193e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.6805869340896606, + "num_tokens": 27240519.0, + "step": 1055 + }, + { + "epoch": 0.11596749396002635, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.381615400314331, + "learning_rate": 3.8616398243045384e-07, + "loss": 1.0279, + "mean_token_accuracy": 0.6979767084121704, + "num_tokens": 27265773.0, + "step": 1056 + }, + { + "epoch": 0.11607731166264001, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.271404504776001, + "learning_rate": 3.8653001464128844e-07, + "loss": 1.1267, + "mean_token_accuracy": 0.6668492555618286, + "num_tokens": 27294062.0, + "step": 1057 + }, + { + "epoch": 0.11618712936525367, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.5516183376312256, + "learning_rate": 3.8689604685212293e-07, + "loss": 1.0596, + "mean_token_accuracy": 0.6869334578514099, + "num_tokens": 27317200.0, + "step": 1058 + }, + { + "epoch": 0.11629694706786733, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.248713970184326, + "learning_rate": 3.8726207906295754e-07, + "loss": 1.1155, + "mean_token_accuracy": 0.6731058359146118, + "num_tokens": 27345129.0, + "step": 1059 + }, + { + "epoch": 0.11640676477048101, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.1678805351257324, + "learning_rate": 3.876281112737921e-07, + "loss": 1.1305, + "mean_token_accuracy": 0.6689473390579224, + "num_tokens": 27376500.0, + "step": 1060 + }, + { + "epoch": 0.11651658247309467, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.4154632091522217, + "learning_rate": 3.8799414348462663e-07, + "loss": 1.1114, + "mean_token_accuracy": 0.6684179306030273, + "num_tokens": 27405136.0, + "step": 1061 + }, + { + "epoch": 0.11662640017570833, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.4038236141204834, + "learning_rate": 3.883601756954612e-07, + "loss": 1.0671, + "mean_token_accuracy": 0.6801676750183105, + "num_tokens": 27430922.0, + "step": 1062 + }, + { + "epoch": 0.11673621787832199, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.681574583053589, + "learning_rate": 3.8872620790629573e-07, + "loss": 1.0348, + "mean_token_accuracy": 0.6935461163520813, + "num_tokens": 27452160.0, + "step": 1063 + }, + { + "epoch": 0.11684603558093565, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.3801820278167725, + "learning_rate": 3.890922401171303e-07, + "loss": 1.0965, + "mean_token_accuracy": 0.6804496049880981, + "num_tokens": 27477265.0, + "step": 1064 + }, + { + "epoch": 0.11695585328354931, + "ewc_loss": 4.887580871582031e-06, + "grad_norm": 2.40077805519104, + "learning_rate": 3.8945827232796483e-07, + "loss": 1.1282, + "mean_token_accuracy": 0.6696544289588928, + "num_tokens": 27503052.0, + "step": 1065 + }, + { + "epoch": 0.11706567098616297, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.5249648094177246, + "learning_rate": 3.898243045387994e-07, + "loss": 1.052, + "mean_token_accuracy": 0.694390594959259, + "num_tokens": 27526173.0, + "step": 1066 + }, + { + "epoch": 0.11717548868877663, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.1602656841278076, + "learning_rate": 3.90190336749634e-07, + "loss": 1.0969, + "mean_token_accuracy": 0.6865058541297913, + "num_tokens": 27556183.0, + "step": 1067 + }, + { + "epoch": 0.11728530639139029, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.2807321548461914, + "learning_rate": 3.9055636896046853e-07, + "loss": 1.0701, + "mean_token_accuracy": 0.6846687197685242, + "num_tokens": 27583306.0, + "step": 1068 + }, + { + "epoch": 0.11739512409400395, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.391321897506714, + "learning_rate": 3.9092240117130303e-07, + "loss": 1.0285, + "mean_token_accuracy": 0.6923929452896118, + "num_tokens": 27607080.0, + "step": 1069 + }, + { + "epoch": 0.11750494179661762, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.0634841918945312, + "learning_rate": 3.9128843338213763e-07, + "loss": 1.1304, + "mean_token_accuracy": 0.6694173812866211, + "num_tokens": 27639131.0, + "step": 1070 + }, + { + "epoch": 0.11761475949923128, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.545100212097168, + "learning_rate": 3.916544655929722e-07, + "loss": 1.0567, + "mean_token_accuracy": 0.6820107698440552, + "num_tokens": 27660915.0, + "step": 1071 + }, + { + "epoch": 0.11772457720184494, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.1305294036865234, + "learning_rate": 3.920204978038067e-07, + "loss": 1.1067, + "mean_token_accuracy": 0.6736186146736145, + "num_tokens": 27693707.0, + "step": 1072 + }, + { + "epoch": 0.1178343949044586, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.3170242309570312, + "learning_rate": 3.923865300146413e-07, + "loss": 1.1867, + "mean_token_accuracy": 0.6561803817749023, + "num_tokens": 27721535.0, + "step": 1073 + }, + { + "epoch": 0.11794421260707226, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.5385537147521973, + "learning_rate": 3.927525622254758e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.7007421255111694, + "num_tokens": 27744277.0, + "step": 1074 + }, + { + "epoch": 0.11805403030968592, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.2452282905578613, + "learning_rate": 3.9311859443631043e-07, + "loss": 1.1025, + "mean_token_accuracy": 0.6755038499832153, + "num_tokens": 27773010.0, + "step": 1075 + }, + { + "epoch": 0.11816384801229958, + "ewc_loss": 4.947185516357422e-06, + "grad_norm": 2.7081496715545654, + "learning_rate": 3.934846266471449e-07, + "loss": 1.1288, + "mean_token_accuracy": 0.6808985471725464, + "num_tokens": 27794162.0, + "step": 1076 + }, + { + "epoch": 0.11827366571491324, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.4711642265319824, + "learning_rate": 3.9385065885797947e-07, + "loss": 1.085, + "mean_token_accuracy": 0.6786062717437744, + "num_tokens": 27817334.0, + "step": 1077 + }, + { + "epoch": 0.1183834834175269, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.5230958461761475, + "learning_rate": 3.942166910688141e-07, + "loss": 1.1487, + "mean_token_accuracy": 0.6592066287994385, + "num_tokens": 27841083.0, + "step": 1078 + }, + { + "epoch": 0.11849330112014057, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.6825263500213623, + "learning_rate": 3.9458272327964857e-07, + "loss": 1.014, + "mean_token_accuracy": 0.6922662258148193, + "num_tokens": 27861503.0, + "step": 1079 + }, + { + "epoch": 0.11860311882275423, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.325873613357544, + "learning_rate": 3.949487554904831e-07, + "loss": 1.0557, + "mean_token_accuracy": 0.6960035562515259, + "num_tokens": 27888130.0, + "step": 1080 + }, + { + "epoch": 0.1187129365253679, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.386793375015259, + "learning_rate": 3.953147877013177e-07, + "loss": 1.064, + "mean_token_accuracy": 0.6813356280326843, + "num_tokens": 27913796.0, + "step": 1081 + }, + { + "epoch": 0.11882275422798155, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.447402000427246, + "learning_rate": 3.956808199121522e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6807131767272949, + "num_tokens": 27938790.0, + "step": 1082 + }, + { + "epoch": 0.11893257193059521, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.296232223510742, + "learning_rate": 3.960468521229868e-07, + "loss": 1.0614, + "mean_token_accuracy": 0.6842906475067139, + "num_tokens": 27964042.0, + "step": 1083 + }, + { + "epoch": 0.11904238963320887, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.3419950008392334, + "learning_rate": 3.9641288433382137e-07, + "loss": 1.079, + "mean_token_accuracy": 0.6871333718299866, + "num_tokens": 27990871.0, + "step": 1084 + }, + { + "epoch": 0.11915220733582253, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.291600465774536, + "learning_rate": 3.967789165446559e-07, + "loss": 1.0795, + "mean_token_accuracy": 0.6728348731994629, + "num_tokens": 28019249.0, + "step": 1085 + }, + { + "epoch": 0.11926202503843619, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2891898155212402, + "learning_rate": 3.9714494875549047e-07, + "loss": 1.0798, + "mean_token_accuracy": 0.674152135848999, + "num_tokens": 28047209.0, + "step": 1086 + }, + { + "epoch": 0.11937184274104985, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.2138919830322266, + "learning_rate": 3.97510980966325e-07, + "loss": 1.1578, + "mean_token_accuracy": 0.6706081628799438, + "num_tokens": 28076909.0, + "step": 1087 + }, + { + "epoch": 0.11948166044366351, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.274089813232422, + "learning_rate": 3.9787701317715957e-07, + "loss": 1.1601, + "mean_token_accuracy": 0.6604017019271851, + "num_tokens": 28105066.0, + "step": 1088 + }, + { + "epoch": 0.11959147814627719, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.7618331909179688, + "learning_rate": 3.982430453879941e-07, + "loss": 1.063, + "mean_token_accuracy": 0.6773629188537598, + "num_tokens": 28126388.0, + "step": 1089 + }, + { + "epoch": 0.11970129584889085, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.384026288986206, + "learning_rate": 3.9860907759882866e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.6919293403625488, + "num_tokens": 28152525.0, + "step": 1090 + }, + { + "epoch": 0.1198111135515045, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.4222912788391113, + "learning_rate": 3.9897510980966327e-07, + "loss": 0.9931, + "mean_token_accuracy": 0.7046076059341431, + "num_tokens": 28176649.0, + "step": 1091 + }, + { + "epoch": 0.11992093125411817, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.8201513290405273, + "learning_rate": 3.993411420204978e-07, + "loss": 0.9767, + "mean_token_accuracy": 0.7146247029304504, + "num_tokens": 28194986.0, + "step": 1092 + }, + { + "epoch": 0.12003074895673183, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.2233262062072754, + "learning_rate": 3.997071742313323e-07, + "loss": 1.0703, + "mean_token_accuracy": 0.6845250725746155, + "num_tokens": 28225155.0, + "step": 1093 + }, + { + "epoch": 0.12014056665934549, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.2062299251556396, + "learning_rate": 4.000732064421669e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6809934377670288, + "num_tokens": 28254599.0, + "step": 1094 + }, + { + "epoch": 0.12025038436195915, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.5496387481689453, + "learning_rate": 4.0043923865300146e-07, + "loss": 1.0069, + "mean_token_accuracy": 0.6952494978904724, + "num_tokens": 28277178.0, + "step": 1095 + }, + { + "epoch": 0.1203602020645728, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.407728910446167, + "learning_rate": 4.0080527086383596e-07, + "loss": 1.0738, + "mean_token_accuracy": 0.6846597194671631, + "num_tokens": 28304301.0, + "step": 1096 + }, + { + "epoch": 0.12047001976718646, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.3138749599456787, + "learning_rate": 4.0117130307467056e-07, + "loss": 0.9607, + "mean_token_accuracy": 0.7110296487808228, + "num_tokens": 28328572.0, + "step": 1097 + }, + { + "epoch": 0.12057983746980014, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.5430731773376465, + "learning_rate": 4.015373352855051e-07, + "loss": 1.0523, + "mean_token_accuracy": 0.6879552006721497, + "num_tokens": 28352705.0, + "step": 1098 + }, + { + "epoch": 0.1206896551724138, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.3993053436279297, + "learning_rate": 4.0190336749633966e-07, + "loss": 1.1038, + "mean_token_accuracy": 0.6784429550170898, + "num_tokens": 28379284.0, + "step": 1099 + }, + { + "epoch": 0.12079947287502746, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 1.9749987125396729, + "learning_rate": 4.022693997071742e-07, + "loss": 1.0982, + "mean_token_accuracy": 0.6784901022911072, + "num_tokens": 28412910.0, + "step": 1100 + }, + { + "epoch": 0.12090929057764112, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2015774250030518, + "learning_rate": 4.0263543191800876e-07, + "loss": 1.0171, + "mean_token_accuracy": 0.7043253183364868, + "num_tokens": 28439611.0, + "step": 1101 + }, + { + "epoch": 0.12101910828025478, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.4610655307769775, + "learning_rate": 4.0300146412884336e-07, + "loss": 1.1243, + "mean_token_accuracy": 0.6745441555976868, + "num_tokens": 28464781.0, + "step": 1102 + }, + { + "epoch": 0.12112892598286844, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.6185970306396484, + "learning_rate": 4.0336749633967785e-07, + "loss": 1.0297, + "mean_token_accuracy": 0.6912183165550232, + "num_tokens": 28486031.0, + "step": 1103 + }, + { + "epoch": 0.1212387436854821, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.57126522064209, + "learning_rate": 4.037335285505124e-07, + "loss": 0.9755, + "mean_token_accuracy": 0.7056441903114319, + "num_tokens": 28507282.0, + "step": 1104 + }, + { + "epoch": 0.12134856138809576, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.4647216796875, + "learning_rate": 4.04099560761347e-07, + "loss": 1.0567, + "mean_token_accuracy": 0.6818373203277588, + "num_tokens": 28532491.0, + "step": 1105 + }, + { + "epoch": 0.12145837909070942, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.1919522285461426, + "learning_rate": 4.044655929721815e-07, + "loss": 1.0164, + "mean_token_accuracy": 0.7057768702507019, + "num_tokens": 28559439.0, + "step": 1106 + }, + { + "epoch": 0.12156819679332308, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3201687335968018, + "learning_rate": 4.048316251830161e-07, + "loss": 1.1148, + "mean_token_accuracy": 0.6709851026535034, + "num_tokens": 28584914.0, + "step": 1107 + }, + { + "epoch": 0.12167801449593675, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.2203662395477295, + "learning_rate": 4.0519765739385065e-07, + "loss": 1.0785, + "mean_token_accuracy": 0.6853868961334229, + "num_tokens": 28613358.0, + "step": 1108 + }, + { + "epoch": 0.12178783219855041, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.402655601501465, + "learning_rate": 4.055636896046852e-07, + "loss": 1.11, + "mean_token_accuracy": 0.6856555938720703, + "num_tokens": 28637827.0, + "step": 1109 + }, + { + "epoch": 0.12189764990116407, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3264410495758057, + "learning_rate": 4.0592972181551975e-07, + "loss": 1.1623, + "mean_token_accuracy": 0.6627057194709778, + "num_tokens": 28663784.0, + "step": 1110 + }, + { + "epoch": 0.12200746760377773, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.4550845623016357, + "learning_rate": 4.062957540263543e-07, + "loss": 1.1242, + "mean_token_accuracy": 0.6690976619720459, + "num_tokens": 28686562.0, + "step": 1111 + }, + { + "epoch": 0.12211728530639139, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.4404735565185547, + "learning_rate": 4.0666178623718885e-07, + "loss": 1.12, + "mean_token_accuracy": 0.6787700653076172, + "num_tokens": 28713361.0, + "step": 1112 + }, + { + "epoch": 0.12222710300900505, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.498600721359253, + "learning_rate": 4.070278184480234e-07, + "loss": 1.0402, + "mean_token_accuracy": 0.6898692846298218, + "num_tokens": 28736221.0, + "step": 1113 + }, + { + "epoch": 0.12233692071161871, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.058748245239258, + "learning_rate": 4.0739385065885795e-07, + "loss": 1.0203, + "mean_token_accuracy": 0.6943824887275696, + "num_tokens": 28767480.0, + "step": 1114 + }, + { + "epoch": 0.12244673841423237, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.58164644241333, + "learning_rate": 4.0775988286969255e-07, + "loss": 1.0392, + "mean_token_accuracy": 0.6956016421318054, + "num_tokens": 28790622.0, + "step": 1115 + }, + { + "epoch": 0.12255655611684603, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.3062446117401123, + "learning_rate": 4.081259150805271e-07, + "loss": 1.0799, + "mean_token_accuracy": 0.6896023750305176, + "num_tokens": 28816316.0, + "step": 1116 + }, + { + "epoch": 0.1226663738194597, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4488089084625244, + "learning_rate": 4.084919472913616e-07, + "loss": 1.0505, + "mean_token_accuracy": 0.687234103679657, + "num_tokens": 28838421.0, + "step": 1117 + }, + { + "epoch": 0.12277619152207336, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.593414068222046, + "learning_rate": 4.088579795021962e-07, + "loss": 1.1659, + "mean_token_accuracy": 0.6628551483154297, + "num_tokens": 28863527.0, + "step": 1118 + }, + { + "epoch": 0.12288600922468702, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.7058873176574707, + "learning_rate": 4.0922401171303075e-07, + "loss": 0.9155, + "mean_token_accuracy": 0.7247362732887268, + "num_tokens": 28882731.0, + "step": 1119 + }, + { + "epoch": 0.12299582692730068, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.5545926094055176, + "learning_rate": 4.0959004392386524e-07, + "loss": 1.0841, + "mean_token_accuracy": 0.6783690452575684, + "num_tokens": 28906336.0, + "step": 1120 + }, + { + "epoch": 0.12310564462991434, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.522399663925171, + "learning_rate": 4.0995607613469984e-07, + "loss": 0.99, + "mean_token_accuracy": 0.7074534893035889, + "num_tokens": 28927718.0, + "step": 1121 + }, + { + "epoch": 0.123215462332528, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3330883979797363, + "learning_rate": 4.103221083455344e-07, + "loss": 1.1507, + "mean_token_accuracy": 0.660772442817688, + "num_tokens": 28956062.0, + "step": 1122 + }, + { + "epoch": 0.12332528003514166, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4203498363494873, + "learning_rate": 4.1068814055636894e-07, + "loss": 1.0915, + "mean_token_accuracy": 0.6816691160202026, + "num_tokens": 28980636.0, + "step": 1123 + }, + { + "epoch": 0.12343509773775532, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.416219711303711, + "learning_rate": 4.110541727672035e-07, + "loss": 1.1208, + "mean_token_accuracy": 0.6632670760154724, + "num_tokens": 29005838.0, + "step": 1124 + }, + { + "epoch": 0.12354491544036898, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3118515014648438, + "learning_rate": 4.1142020497803804e-07, + "loss": 1.1356, + "mean_token_accuracy": 0.6637134552001953, + "num_tokens": 29032584.0, + "step": 1125 + }, + { + "epoch": 0.12365473314298264, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.5475547313690186, + "learning_rate": 4.1178623718887264e-07, + "loss": 1.0704, + "mean_token_accuracy": 0.6844526529312134, + "num_tokens": 29054249.0, + "step": 1126 + }, + { + "epoch": 0.12376455084559632, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.165576934814453, + "learning_rate": 4.1215226939970714e-07, + "loss": 1.0877, + "mean_token_accuracy": 0.6779792904853821, + "num_tokens": 29083825.0, + "step": 1127 + }, + { + "epoch": 0.12387436854820998, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.1760358810424805, + "learning_rate": 4.125183016105417e-07, + "loss": 1.0943, + "mean_token_accuracy": 0.6859252452850342, + "num_tokens": 29113619.0, + "step": 1128 + }, + { + "epoch": 0.12398418625082364, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.2697038650512695, + "learning_rate": 4.128843338213763e-07, + "loss": 1.1347, + "mean_token_accuracy": 0.6717960238456726, + "num_tokens": 29141354.0, + "step": 1129 + }, + { + "epoch": 0.1240940039534373, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.522057294845581, + "learning_rate": 4.132503660322108e-07, + "loss": 1.0745, + "mean_token_accuracy": 0.6784918308258057, + "num_tokens": 29163113.0, + "step": 1130 + }, + { + "epoch": 0.12420382165605096, + "ewc_loss": 5.424022674560547e-06, + "grad_norm": 2.2308883666992188, + "learning_rate": 4.136163982430454e-07, + "loss": 1.0308, + "mean_token_accuracy": 0.6932639479637146, + "num_tokens": 29190986.0, + "step": 1131 + }, + { + "epoch": 0.12431363935866462, + "ewc_loss": 5.424022674560547e-06, + "grad_norm": 2.1699674129486084, + "learning_rate": 4.1398243045387994e-07, + "loss": 0.963, + "mean_token_accuracy": 0.7108134627342224, + "num_tokens": 29218126.0, + "step": 1132 + }, + { + "epoch": 0.12442345706127828, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.2798616886138916, + "learning_rate": 4.143484626647145e-07, + "loss": 1.0819, + "mean_token_accuracy": 0.677129864692688, + "num_tokens": 29246177.0, + "step": 1133 + }, + { + "epoch": 0.12453327476389194, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.4309279918670654, + "learning_rate": 4.1471449487554904e-07, + "loss": 1.0472, + "mean_token_accuracy": 0.6971656680107117, + "num_tokens": 29270132.0, + "step": 1134 + }, + { + "epoch": 0.1246430924665056, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.2189841270446777, + "learning_rate": 4.150805270863836e-07, + "loss": 1.1176, + "mean_token_accuracy": 0.6688819527626038, + "num_tokens": 29298680.0, + "step": 1135 + }, + { + "epoch": 0.12475291016911927, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.59822678565979, + "learning_rate": 4.1544655929721813e-07, + "loss": 1.0216, + "mean_token_accuracy": 0.6973454356193542, + "num_tokens": 29319341.0, + "step": 1136 + }, + { + "epoch": 0.12486272787173293, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.343414545059204, + "learning_rate": 4.158125915080527e-07, + "loss": 1.0652, + "mean_token_accuracy": 0.6971006989479065, + "num_tokens": 29344645.0, + "step": 1137 + }, + { + "epoch": 0.12497254557434659, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.241722822189331, + "learning_rate": 4.1617862371888723e-07, + "loss": 1.0364, + "mean_token_accuracy": 0.6915065050125122, + "num_tokens": 29370995.0, + "step": 1138 + }, + { + "epoch": 0.12508236327696023, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.438566207885742, + "learning_rate": 4.1654465592972183e-07, + "loss": 1.1479, + "mean_token_accuracy": 0.6666138768196106, + "num_tokens": 29396692.0, + "step": 1139 + }, + { + "epoch": 0.1251921809795739, + "ewc_loss": 5.513429641723633e-06, + "grad_norm": 2.21034574508667, + "learning_rate": 4.1691068814055633e-07, + "loss": 1.0965, + "mean_token_accuracy": 0.6735478639602661, + "num_tokens": 29426669.0, + "step": 1140 + }, + { + "epoch": 0.12530199868218758, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.395427942276001, + "learning_rate": 4.172767203513909e-07, + "loss": 1.0045, + "mean_token_accuracy": 0.699373722076416, + "num_tokens": 29451681.0, + "step": 1141 + }, + { + "epoch": 0.12541181638480123, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.0413413047790527, + "learning_rate": 4.176427525622255e-07, + "loss": 1.1308, + "mean_token_accuracy": 0.6729551553726196, + "num_tokens": 29484702.0, + "step": 1142 + }, + { + "epoch": 0.1255216340874149, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.6457481384277344, + "learning_rate": 4.1800878477306003e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7055937051773071, + "num_tokens": 29504341.0, + "step": 1143 + }, + { + "epoch": 0.12563145179002855, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.296276807785034, + "learning_rate": 4.183748169838945e-07, + "loss": 1.0094, + "mean_token_accuracy": 0.6912451386451721, + "num_tokens": 29530130.0, + "step": 1144 + }, + { + "epoch": 0.12574126949264222, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.2020297050476074, + "learning_rate": 4.1874084919472913e-07, + "loss": 1.1088, + "mean_token_accuracy": 0.6724516749382019, + "num_tokens": 29559283.0, + "step": 1145 + }, + { + "epoch": 0.12585108719525587, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.3867714405059814, + "learning_rate": 4.191068814055637e-07, + "loss": 1.1018, + "mean_token_accuracy": 0.6756142377853394, + "num_tokens": 29584911.0, + "step": 1146 + }, + { + "epoch": 0.12596090489786954, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.291407585144043, + "learning_rate": 4.194729136163982e-07, + "loss": 0.9873, + "mean_token_accuracy": 0.702142596244812, + "num_tokens": 29610399.0, + "step": 1147 + }, + { + "epoch": 0.1260707226004832, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.4479498863220215, + "learning_rate": 4.198389458272328e-07, + "loss": 0.9576, + "mean_token_accuracy": 0.708172619342804, + "num_tokens": 29633032.0, + "step": 1148 + }, + { + "epoch": 0.12618054030309686, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.3615007400512695, + "learning_rate": 4.202049780380673e-07, + "loss": 1.0671, + "mean_token_accuracy": 0.6804220080375671, + "num_tokens": 29657214.0, + "step": 1149 + }, + { + "epoch": 0.1262903580057105, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.0890419483184814, + "learning_rate": 4.205710102489019e-07, + "loss": 1.0746, + "mean_token_accuracy": 0.6816178560256958, + "num_tokens": 29688013.0, + "step": 1150 + }, + { + "epoch": 0.12640017570832418, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.5519611835479736, + "learning_rate": 4.209370424597364e-07, + "loss": 1.0131, + "mean_token_accuracy": 0.6957585215568542, + "num_tokens": 29710451.0, + "step": 1151 + }, + { + "epoch": 0.12650999341093785, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3892276287078857, + "learning_rate": 4.2130307467057097e-07, + "loss": 1.0873, + "mean_token_accuracy": 0.6785133481025696, + "num_tokens": 29735654.0, + "step": 1152 + }, + { + "epoch": 0.1266198111135515, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.4263432025909424, + "learning_rate": 4.2166910688140557e-07, + "loss": 1.1356, + "mean_token_accuracy": 0.6612073183059692, + "num_tokens": 29764641.0, + "step": 1153 + }, + { + "epoch": 0.12672962881616517, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.468820571899414, + "learning_rate": 4.2203513909224007e-07, + "loss": 1.1458, + "mean_token_accuracy": 0.6690158843994141, + "num_tokens": 29790458.0, + "step": 1154 + }, + { + "epoch": 0.12683944651877882, + "ewc_loss": 5.692243576049805e-06, + "grad_norm": 2.390190839767456, + "learning_rate": 4.2240117130307467e-07, + "loss": 1.0513, + "mean_token_accuracy": 0.6874101161956787, + "num_tokens": 29816208.0, + "step": 1155 + }, + { + "epoch": 0.1269492642213925, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.3943493366241455, + "learning_rate": 4.227672035139092e-07, + "loss": 1.0833, + "mean_token_accuracy": 0.6871442794799805, + "num_tokens": 29840722.0, + "step": 1156 + }, + { + "epoch": 0.12705908192400614, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.285916805267334, + "learning_rate": 4.231332357247437e-07, + "loss": 1.1737, + "mean_token_accuracy": 0.6572602987289429, + "num_tokens": 29869138.0, + "step": 1157 + }, + { + "epoch": 0.1271688996266198, + "ewc_loss": 5.692243576049805e-06, + "grad_norm": 2.5273258686065674, + "learning_rate": 4.234992679355783e-07, + "loss": 1.0456, + "mean_token_accuracy": 0.6857936382293701, + "num_tokens": 29891804.0, + "step": 1158 + }, + { + "epoch": 0.12727871732923346, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.311389446258545, + "learning_rate": 4.2386530014641287e-07, + "loss": 1.1326, + "mean_token_accuracy": 0.6735397577285767, + "num_tokens": 29918368.0, + "step": 1159 + }, + { + "epoch": 0.12738853503184713, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.2557711601257324, + "learning_rate": 4.242313323572474e-07, + "loss": 1.0937, + "mean_token_accuracy": 0.6830952763557434, + "num_tokens": 29947243.0, + "step": 1160 + }, + { + "epoch": 0.1274983527344608, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.2041337490081787, + "learning_rate": 4.2459736456808197e-07, + "loss": 1.1251, + "mean_token_accuracy": 0.6719652414321899, + "num_tokens": 29976185.0, + "step": 1161 + }, + { + "epoch": 0.12760817043707445, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.237879753112793, + "learning_rate": 4.249633967789165e-07, + "loss": 1.0915, + "mean_token_accuracy": 0.6753647923469543, + "num_tokens": 30004893.0, + "step": 1162 + }, + { + "epoch": 0.12771798813968813, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.1233832836151123, + "learning_rate": 4.253294289897511e-07, + "loss": 1.1824, + "mean_token_accuracy": 0.6573510766029358, + "num_tokens": 30033645.0, + "step": 1163 + }, + { + "epoch": 0.12782780584230177, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.6149849891662598, + "learning_rate": 4.256954612005856e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6854760646820068, + "num_tokens": 30055734.0, + "step": 1164 + }, + { + "epoch": 0.12793762354491545, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.5412843227386475, + "learning_rate": 4.2606149341142016e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.6955307722091675, + "num_tokens": 30076609.0, + "step": 1165 + }, + { + "epoch": 0.1280474412475291, + "ewc_loss": 5.692243576049805e-06, + "grad_norm": 2.5152666568756104, + "learning_rate": 4.2642752562225476e-07, + "loss": 1.1211, + "mean_token_accuracy": 0.6634707450866699, + "num_tokens": 30101695.0, + "step": 1166 + }, + { + "epoch": 0.12815725895014277, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.386629343032837, + "learning_rate": 4.267935578330893e-07, + "loss": 1.128, + "mean_token_accuracy": 0.6729167699813843, + "num_tokens": 30128385.0, + "step": 1167 + }, + { + "epoch": 0.1282670766527564, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.5719072818756104, + "learning_rate": 4.271595900439238e-07, + "loss": 0.9097, + "mean_token_accuracy": 0.7281264662742615, + "num_tokens": 30149434.0, + "step": 1168 + }, + { + "epoch": 0.12837689435537009, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2933053970336914, + "learning_rate": 4.275256222547584e-07, + "loss": 1.0202, + "mean_token_accuracy": 0.6969645023345947, + "num_tokens": 30175072.0, + "step": 1169 + }, + { + "epoch": 0.12848671205798376, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.173701524734497, + "learning_rate": 4.2789165446559296e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6846688389778137, + "num_tokens": 30204221.0, + "step": 1170 + }, + { + "epoch": 0.1285965297605974, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.1859145164489746, + "learning_rate": 4.282576866764275e-07, + "loss": 1.0996, + "mean_token_accuracy": 0.6761796474456787, + "num_tokens": 30232624.0, + "step": 1171 + }, + { + "epoch": 0.12870634746321108, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.270979404449463, + "learning_rate": 4.2862371888726206e-07, + "loss": 1.0837, + "mean_token_accuracy": 0.6852242946624756, + "num_tokens": 30259309.0, + "step": 1172 + }, + { + "epoch": 0.12881616516582473, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.3751015663146973, + "learning_rate": 4.289897510980966e-07, + "loss": 1.0785, + "mean_token_accuracy": 0.68570476770401, + "num_tokens": 30286533.0, + "step": 1173 + }, + { + "epoch": 0.1289259828684384, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.440880298614502, + "learning_rate": 4.293557833089312e-07, + "loss": 1.0402, + "mean_token_accuracy": 0.6878630518913269, + "num_tokens": 30311150.0, + "step": 1174 + }, + { + "epoch": 0.12903580057105205, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.4873204231262207, + "learning_rate": 4.297218155197657e-07, + "loss": 1.0059, + "mean_token_accuracy": 0.6965562701225281, + "num_tokens": 30334858.0, + "step": 1175 + }, + { + "epoch": 0.12914561827366572, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.5704996585845947, + "learning_rate": 4.3008784773060026e-07, + "loss": 0.9826, + "mean_token_accuracy": 0.7097850441932678, + "num_tokens": 30358502.0, + "step": 1176 + }, + { + "epoch": 0.12925543597627936, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.3340959548950195, + "learning_rate": 4.3045387994143486e-07, + "loss": 1.114, + "mean_token_accuracy": 0.6700533628463745, + "num_tokens": 30384783.0, + "step": 1177 + }, + { + "epoch": 0.12936525367889304, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.4645936489105225, + "learning_rate": 4.3081991215226935e-07, + "loss": 1.0848, + "mean_token_accuracy": 0.6898571252822876, + "num_tokens": 30408913.0, + "step": 1178 + }, + { + "epoch": 0.1294750713815067, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.443284511566162, + "learning_rate": 4.3118594436310396e-07, + "loss": 1.0384, + "mean_token_accuracy": 0.694433331489563, + "num_tokens": 30433900.0, + "step": 1179 + }, + { + "epoch": 0.12958488908412036, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.512249708175659, + "learning_rate": 4.315519765739385e-07, + "loss": 1.0018, + "mean_token_accuracy": 0.7037067413330078, + "num_tokens": 30461190.0, + "step": 1180 + }, + { + "epoch": 0.12969470678673403, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.362351894378662, + "learning_rate": 4.31918008784773e-07, + "loss": 1.1012, + "mean_token_accuracy": 0.6755901575088501, + "num_tokens": 30489651.0, + "step": 1181 + }, + { + "epoch": 0.12980452448934768, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.3238589763641357, + "learning_rate": 4.322840409956076e-07, + "loss": 1.0773, + "mean_token_accuracy": 0.6942591071128845, + "num_tokens": 30519052.0, + "step": 1182 + }, + { + "epoch": 0.12991434219196135, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.305281639099121, + "learning_rate": 4.3265007320644215e-07, + "loss": 0.9857, + "mean_token_accuracy": 0.7037215232849121, + "num_tokens": 30543441.0, + "step": 1183 + }, + { + "epoch": 0.130024159894575, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.4749855995178223, + "learning_rate": 4.330161054172767e-07, + "loss": 1.1334, + "mean_token_accuracy": 0.6705968379974365, + "num_tokens": 30567401.0, + "step": 1184 + }, + { + "epoch": 0.13013397759718867, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.3402581214904785, + "learning_rate": 4.3338213762811125e-07, + "loss": 1.1316, + "mean_token_accuracy": 0.6677943468093872, + "num_tokens": 30594937.0, + "step": 1185 + }, + { + "epoch": 0.13024379529980232, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.1840193271636963, + "learning_rate": 4.337481698389458e-07, + "loss": 0.9847, + "mean_token_accuracy": 0.7086130976676941, + "num_tokens": 30623647.0, + "step": 1186 + }, + { + "epoch": 0.130353613002416, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.3742222785949707, + "learning_rate": 4.341142020497804e-07, + "loss": 1.0937, + "mean_token_accuracy": 0.6827487945556641, + "num_tokens": 30650386.0, + "step": 1187 + }, + { + "epoch": 0.13046343070502964, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.574493169784546, + "learning_rate": 4.344802342606149e-07, + "loss": 1.0676, + "mean_token_accuracy": 0.6817660331726074, + "num_tokens": 30673298.0, + "step": 1188 + }, + { + "epoch": 0.1305732484076433, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1806800365448, + "learning_rate": 4.3484626647144945e-07, + "loss": 1.0922, + "mean_token_accuracy": 0.6849042773246765, + "num_tokens": 30702162.0, + "step": 1189 + }, + { + "epoch": 0.13068306611025698, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.372633218765259, + "learning_rate": 4.3521229868228405e-07, + "loss": 1.1217, + "mean_token_accuracy": 0.6757669448852539, + "num_tokens": 30728125.0, + "step": 1190 + }, + { + "epoch": 0.13079288381287063, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.420048713684082, + "learning_rate": 4.355783308931186e-07, + "loss": 1.0021, + "mean_token_accuracy": 0.7021862268447876, + "num_tokens": 30753050.0, + "step": 1191 + }, + { + "epoch": 0.1309027015154843, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.397819995880127, + "learning_rate": 4.359443631039531e-07, + "loss": 1.0522, + "mean_token_accuracy": 0.6874030232429504, + "num_tokens": 30778703.0, + "step": 1192 + }, + { + "epoch": 0.13101251921809795, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.0592288970947266, + "learning_rate": 4.363103953147877e-07, + "loss": 1.0472, + "mean_token_accuracy": 0.6875321269035339, + "num_tokens": 30810814.0, + "step": 1193 + }, + { + "epoch": 0.13112233692071162, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.304478168487549, + "learning_rate": 4.3667642752562224e-07, + "loss": 1.058, + "mean_token_accuracy": 0.6988401412963867, + "num_tokens": 30836427.0, + "step": 1194 + }, + { + "epoch": 0.13123215462332527, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.1840691566467285, + "learning_rate": 4.370424597364568e-07, + "loss": 1.1233, + "mean_token_accuracy": 0.6685236692428589, + "num_tokens": 30865532.0, + "step": 1195 + }, + { + "epoch": 0.13134197232593894, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.600440263748169, + "learning_rate": 4.3740849194729134e-07, + "loss": 1.0461, + "mean_token_accuracy": 0.6863497495651245, + "num_tokens": 30885416.0, + "step": 1196 + }, + { + "epoch": 0.1314517900285526, + "ewc_loss": 5.900859832763672e-06, + "grad_norm": 2.4075918197631836, + "learning_rate": 4.377745241581259e-07, + "loss": 1.0202, + "mean_token_accuracy": 0.6911956071853638, + "num_tokens": 30909752.0, + "step": 1197 + }, + { + "epoch": 0.13156160773116626, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.788252592086792, + "learning_rate": 4.381405563689605e-07, + "loss": 1.0022, + "mean_token_accuracy": 0.7016655206680298, + "num_tokens": 30928463.0, + "step": 1198 + }, + { + "epoch": 0.13167142543377994, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.6116251945495605, + "learning_rate": 4.38506588579795e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6902748942375183, + "num_tokens": 30950612.0, + "step": 1199 + }, + { + "epoch": 0.13178124313639358, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.299980878829956, + "learning_rate": 4.3887262079062954e-07, + "loss": 1.0416, + "mean_token_accuracy": 0.6975101232528687, + "num_tokens": 30975408.0, + "step": 1200 + }, + { + "epoch": 0.13189106083900726, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.4713525772094727, + "learning_rate": 4.3923865300146414e-07, + "loss": 0.9497, + "mean_token_accuracy": 0.7131340503692627, + "num_tokens": 30998382.0, + "step": 1201 + }, + { + "epoch": 0.1320008785416209, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.2022922039031982, + "learning_rate": 4.3960468521229864e-07, + "loss": 1.0035, + "mean_token_accuracy": 0.7026125192642212, + "num_tokens": 31024645.0, + "step": 1202 + }, + { + "epoch": 0.13211069624423458, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.4134011268615723, + "learning_rate": 4.3997071742313324e-07, + "loss": 1.0604, + "mean_token_accuracy": 0.6931407451629639, + "num_tokens": 31050392.0, + "step": 1203 + }, + { + "epoch": 0.13222051394684822, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.457913875579834, + "learning_rate": 4.403367496339678e-07, + "loss": 0.9953, + "mean_token_accuracy": 0.703004002571106, + "num_tokens": 31072099.0, + "step": 1204 + }, + { + "epoch": 0.1323303316494619, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.2189571857452393, + "learning_rate": 4.407027818448023e-07, + "loss": 1.0309, + "mean_token_accuracy": 0.6895698308944702, + "num_tokens": 31099348.0, + "step": 1205 + }, + { + "epoch": 0.13244014935207554, + "ewc_loss": 5.990266799926758e-06, + "grad_norm": 2.232414722442627, + "learning_rate": 4.410688140556369e-07, + "loss": 1.0303, + "mean_token_accuracy": 0.6907385587692261, + "num_tokens": 31127674.0, + "step": 1206 + }, + { + "epoch": 0.13254996705468922, + "ewc_loss": 5.990266799926758e-06, + "grad_norm": 2.681739568710327, + "learning_rate": 4.4143484626647144e-07, + "loss": 1.0232, + "mean_token_accuracy": 0.7034245133399963, + "num_tokens": 31147826.0, + "step": 1207 + }, + { + "epoch": 0.1326597847573029, + "ewc_loss": 5.990266799926758e-06, + "grad_norm": 2.666684865951538, + "learning_rate": 4.41800878477306e-07, + "loss": 0.9709, + "mean_token_accuracy": 0.7054340839385986, + "num_tokens": 31168192.0, + "step": 1208 + }, + { + "epoch": 0.13276960245991654, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.772428274154663, + "learning_rate": 4.4216691068814053e-07, + "loss": 0.8648, + "mean_token_accuracy": 0.7245863080024719, + "num_tokens": 31186404.0, + "step": 1209 + }, + { + "epoch": 0.1328794201625302, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.14573073387146, + "learning_rate": 4.425329428989751e-07, + "loss": 1.1183, + "mean_token_accuracy": 0.6654367446899414, + "num_tokens": 31216377.0, + "step": 1210 + }, + { + "epoch": 0.13298923786514386, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.465938091278076, + "learning_rate": 4.428989751098097e-07, + "loss": 1.071, + "mean_token_accuracy": 0.6817848682403564, + "num_tokens": 31241083.0, + "step": 1211 + }, + { + "epoch": 0.13309905556775753, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3230316638946533, + "learning_rate": 4.432650073206442e-07, + "loss": 1.1045, + "mean_token_accuracy": 0.676366925239563, + "num_tokens": 31269362.0, + "step": 1212 + }, + { + "epoch": 0.13320887327037118, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 1.996837854385376, + "learning_rate": 4.4363103953147873e-07, + "loss": 1.0073, + "mean_token_accuracy": 0.6972862482070923, + "num_tokens": 31301121.0, + "step": 1213 + }, + { + "epoch": 0.13331869097298485, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.6356699466705322, + "learning_rate": 4.4399707174231333e-07, + "loss": 1.1051, + "mean_token_accuracy": 0.6692697405815125, + "num_tokens": 31323850.0, + "step": 1214 + }, + { + "epoch": 0.1334285086755985, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.33418607711792, + "learning_rate": 4.443631039531479e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6859018802642822, + "num_tokens": 31349231.0, + "step": 1215 + }, + { + "epoch": 0.13353832637821217, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.718947649002075, + "learning_rate": 4.447291361639824e-07, + "loss": 1.1014, + "mean_token_accuracy": 0.6782727837562561, + "num_tokens": 31368491.0, + "step": 1216 + }, + { + "epoch": 0.13364814408082584, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.3896758556365967, + "learning_rate": 4.45095168374817e-07, + "loss": 0.9686, + "mean_token_accuracy": 0.7048010230064392, + "num_tokens": 31392465.0, + "step": 1217 + }, + { + "epoch": 0.1337579617834395, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.439581871032715, + "learning_rate": 4.4546120058565153e-07, + "loss": 1.0733, + "mean_token_accuracy": 0.6765322685241699, + "num_tokens": 31417587.0, + "step": 1218 + }, + { + "epoch": 0.13386777948605316, + "ewc_loss": 6.0498714447021484e-06, + "grad_norm": 2.094156503677368, + "learning_rate": 4.458272327964861e-07, + "loss": 1.0267, + "mean_token_accuracy": 0.6963791847229004, + "num_tokens": 31448313.0, + "step": 1219 + }, + { + "epoch": 0.1339775971886668, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.415787935256958, + "learning_rate": 4.4619326500732063e-07, + "loss": 1.0304, + "mean_token_accuracy": 0.6891929507255554, + "num_tokens": 31472179.0, + "step": 1220 + }, + { + "epoch": 0.13408741489128048, + "ewc_loss": 6.109476089477539e-06, + "grad_norm": 2.1817824840545654, + "learning_rate": 4.465592972181552e-07, + "loss": 1.091, + "mean_token_accuracy": 0.6847102642059326, + "num_tokens": 31501837.0, + "step": 1221 + }, + { + "epoch": 0.13419723259389413, + "ewc_loss": 6.109476089477539e-06, + "grad_norm": 2.4756288528442383, + "learning_rate": 4.469253294289897e-07, + "loss": 1.0116, + "mean_token_accuracy": 0.6982775926589966, + "num_tokens": 31523276.0, + "step": 1222 + }, + { + "epoch": 0.1343070502965078, + "ewc_loss": 6.109476089477539e-06, + "grad_norm": 2.490648031234741, + "learning_rate": 4.472913616398243e-07, + "loss": 1.1596, + "mean_token_accuracy": 0.6637140512466431, + "num_tokens": 31547514.0, + "step": 1223 + }, + { + "epoch": 0.13441686799912145, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.1663618087768555, + "learning_rate": 4.476573938506588e-07, + "loss": 0.9933, + "mean_token_accuracy": 0.7007508873939514, + "num_tokens": 31575007.0, + "step": 1224 + }, + { + "epoch": 0.13452668570173512, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4030563831329346, + "learning_rate": 4.480234260614934e-07, + "loss": 1.0373, + "mean_token_accuracy": 0.6886074542999268, + "num_tokens": 31600483.0, + "step": 1225 + }, + { + "epoch": 0.13463650340434877, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4926488399505615, + "learning_rate": 4.483894582723279e-07, + "loss": 1.0256, + "mean_token_accuracy": 0.6987686157226562, + "num_tokens": 31622639.0, + "step": 1226 + }, + { + "epoch": 0.13474632110696244, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4808125495910645, + "learning_rate": 4.487554904831625e-07, + "loss": 1.0466, + "mean_token_accuracy": 0.6998229026794434, + "num_tokens": 31647351.0, + "step": 1227 + }, + { + "epoch": 0.13485613880957611, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.7499125003814697, + "learning_rate": 4.4912152269399707e-07, + "loss": 1.1055, + "mean_token_accuracy": 0.6758615970611572, + "num_tokens": 31667220.0, + "step": 1228 + }, + { + "epoch": 0.13496595651218976, + "ewc_loss": 6.22868537902832e-06, + "grad_norm": 2.5572330951690674, + "learning_rate": 4.4948755490483157e-07, + "loss": 1.1167, + "mean_token_accuracy": 0.6790781021118164, + "num_tokens": 31689044.0, + "step": 1229 + }, + { + "epoch": 0.13507577421480343, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3235063552856445, + "learning_rate": 4.4985358711566617e-07, + "loss": 1.2479, + "mean_token_accuracy": 0.64618980884552, + "num_tokens": 31717794.0, + "step": 1230 + }, + { + "epoch": 0.13518559191741708, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3561763763427734, + "learning_rate": 4.502196193265007e-07, + "loss": 1.1123, + "mean_token_accuracy": 0.6721302270889282, + "num_tokens": 31743349.0, + "step": 1231 + }, + { + "epoch": 0.13529540962003075, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.8217875957489014, + "learning_rate": 4.5058565153733527e-07, + "loss": 0.9949, + "mean_token_accuracy": 0.6962504386901855, + "num_tokens": 31761698.0, + "step": 1232 + }, + { + "epoch": 0.1354052273226444, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.869464635848999, + "learning_rate": 4.509516837481698e-07, + "loss": 0.9901, + "mean_token_accuracy": 0.7067229747772217, + "num_tokens": 31780060.0, + "step": 1233 + }, + { + "epoch": 0.13551504502525807, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.116384506225586, + "learning_rate": 4.5131771595900437e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.7014289498329163, + "num_tokens": 31810081.0, + "step": 1234 + }, + { + "epoch": 0.13562486272787172, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.2236766815185547, + "learning_rate": 4.5168374816983897e-07, + "loss": 1.0723, + "mean_token_accuracy": 0.6818452477455139, + "num_tokens": 31838983.0, + "step": 1235 + }, + { + "epoch": 0.1357346804304854, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.2937891483306885, + "learning_rate": 4.5204978038067347e-07, + "loss": 1.0802, + "mean_token_accuracy": 0.6877676248550415, + "num_tokens": 31864058.0, + "step": 1236 + }, + { + "epoch": 0.13584449813309907, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.4040749073028564, + "learning_rate": 4.52415812591508e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6838188171386719, + "num_tokens": 31889614.0, + "step": 1237 + }, + { + "epoch": 0.1359543158357127, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.3602840900421143, + "learning_rate": 4.527818448023426e-07, + "loss": 1.162, + "mean_token_accuracy": 0.6707057952880859, + "num_tokens": 31915897.0, + "step": 1238 + }, + { + "epoch": 0.1360641335383264, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.2176475524902344, + "learning_rate": 4.531478770131771e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.691948413848877, + "num_tokens": 31943588.0, + "step": 1239 + }, + { + "epoch": 0.13617395124094003, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.299502372741699, + "learning_rate": 4.5351390922401166e-07, + "loss": 1.0775, + "mean_token_accuracy": 0.6842089295387268, + "num_tokens": 31970532.0, + "step": 1240 + }, + { + "epoch": 0.1362837689435537, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3255980014801025, + "learning_rate": 4.5387994143484626e-07, + "loss": 1.0424, + "mean_token_accuracy": 0.6963999271392822, + "num_tokens": 31996697.0, + "step": 1241 + }, + { + "epoch": 0.13639358664616735, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.6162567138671875, + "learning_rate": 4.542459736456808e-07, + "loss": 0.9954, + "mean_token_accuracy": 0.7053240537643433, + "num_tokens": 32018303.0, + "step": 1242 + }, + { + "epoch": 0.13650340434878103, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1498634815216064, + "learning_rate": 4.5461200585651536e-07, + "loss": 1.1183, + "mean_token_accuracy": 0.6670007705688477, + "num_tokens": 32046571.0, + "step": 1243 + }, + { + "epoch": 0.13661322205139467, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.436086416244507, + "learning_rate": 4.549780380673499e-07, + "loss": 1.1418, + "mean_token_accuracy": 0.6698042154312134, + "num_tokens": 32072281.0, + "step": 1244 + }, + { + "epoch": 0.13672303975400835, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.420928478240967, + "learning_rate": 4.5534407027818446e-07, + "loss": 1.1783, + "mean_token_accuracy": 0.6609065532684326, + "num_tokens": 32099451.0, + "step": 1245 + }, + { + "epoch": 0.13683285745662202, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3485758304595947, + "learning_rate": 4.55710102489019e-07, + "loss": 1.0435, + "mean_token_accuracy": 0.6924333572387695, + "num_tokens": 32124080.0, + "step": 1246 + }, + { + "epoch": 0.13694267515923567, + "ewc_loss": 6.377696990966797e-06, + "grad_norm": 2.2380425930023193, + "learning_rate": 4.5607613469985356e-07, + "loss": 1.1456, + "mean_token_accuracy": 0.6627746820449829, + "num_tokens": 32153434.0, + "step": 1247 + }, + { + "epoch": 0.13705249286184934, + "ewc_loss": 6.377696990966797e-06, + "grad_norm": 2.4840352535247803, + "learning_rate": 4.564421669106881e-07, + "loss": 1.0215, + "mean_token_accuracy": 0.6977400779724121, + "num_tokens": 32176510.0, + "step": 1248 + }, + { + "epoch": 0.13716231056446299, + "ewc_loss": 6.377696990966797e-06, + "grad_norm": 2.0868635177612305, + "learning_rate": 4.568081991215227e-07, + "loss": 1.0741, + "mean_token_accuracy": 0.6789819598197937, + "num_tokens": 32206030.0, + "step": 1249 + }, + { + "epoch": 0.13727212826707666, + "ewc_loss": 6.377696990966797e-06, + "grad_norm": 2.26474928855896, + "learning_rate": 4.571742313323572e-07, + "loss": 1.0518, + "mean_token_accuracy": 0.6918044090270996, + "num_tokens": 32233255.0, + "step": 1250 + }, + { + "epoch": 0.1373819459696903, + "ewc_loss": 6.377696990966797e-06, + "grad_norm": 2.3469202518463135, + "learning_rate": 4.575402635431918e-07, + "loss": 1.0409, + "mean_token_accuracy": 0.6939035654067993, + "num_tokens": 32257818.0, + "step": 1251 + }, + { + "epoch": 0.13749176367230398, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.2146413326263428, + "learning_rate": 4.5790629575402636e-07, + "loss": 1.0459, + "mean_token_accuracy": 0.6889933943748474, + "num_tokens": 32286826.0, + "step": 1252 + }, + { + "epoch": 0.13760158137491763, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.457625389099121, + "learning_rate": 4.5827232796486085e-07, + "loss": 1.0534, + "mean_token_accuracy": 0.6858755350112915, + "num_tokens": 32310447.0, + "step": 1253 + }, + { + "epoch": 0.1377113990775313, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.3343915939331055, + "learning_rate": 4.5863836017569545e-07, + "loss": 1.0961, + "mean_token_accuracy": 0.6917237639427185, + "num_tokens": 32335713.0, + "step": 1254 + }, + { + "epoch": 0.13782121678014497, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.2663087844848633, + "learning_rate": 4.5900439238653e-07, + "loss": 1.0641, + "mean_token_accuracy": 0.6884089112281799, + "num_tokens": 32361259.0, + "step": 1255 + }, + { + "epoch": 0.13793103448275862, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.61352801322937, + "learning_rate": 4.593704245973645e-07, + "loss": 1.0488, + "mean_token_accuracy": 0.6876411437988281, + "num_tokens": 32383138.0, + "step": 1256 + }, + { + "epoch": 0.1380408521853723, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.2454910278320312, + "learning_rate": 4.597364568081991e-07, + "loss": 1.0492, + "mean_token_accuracy": 0.6863127946853638, + "num_tokens": 32409879.0, + "step": 1257 + }, + { + "epoch": 0.13815066988798594, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.665121078491211, + "learning_rate": 4.6010248901903365e-07, + "loss": 1.0293, + "mean_token_accuracy": 0.6898471713066101, + "num_tokens": 32430039.0, + "step": 1258 + }, + { + "epoch": 0.1382604875905996, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.345276117324829, + "learning_rate": 4.6046852122986825e-07, + "loss": 1.0328, + "mean_token_accuracy": 0.6904488801956177, + "num_tokens": 32456043.0, + "step": 1259 + }, + { + "epoch": 0.13837030529321326, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.0338714122772217, + "learning_rate": 4.6083455344070275e-07, + "loss": 1.1411, + "mean_token_accuracy": 0.664178192615509, + "num_tokens": 32489703.0, + "step": 1260 + }, + { + "epoch": 0.13848012299582693, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 1.9538817405700684, + "learning_rate": 4.612005856515373e-07, + "loss": 1.0946, + "mean_token_accuracy": 0.6757456064224243, + "num_tokens": 32523481.0, + "step": 1261 + }, + { + "epoch": 0.13858994069844058, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.6571805477142334, + "learning_rate": 4.615666178623719e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.694894552230835, + "num_tokens": 32544230.0, + "step": 1262 + }, + { + "epoch": 0.13869975840105425, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.6846933364868164, + "learning_rate": 4.619326500732064e-07, + "loss": 1.0034, + "mean_token_accuracy": 0.696969747543335, + "num_tokens": 32563592.0, + "step": 1263 + }, + { + "epoch": 0.1388095761036679, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.3147683143615723, + "learning_rate": 4.6229868228404095e-07, + "loss": 1.1587, + "mean_token_accuracy": 0.6724444627761841, + "num_tokens": 32592096.0, + "step": 1264 + }, + { + "epoch": 0.13891939380628157, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.4246602058410645, + "learning_rate": 4.6266471449487555e-07, + "loss": 1.0281, + "mean_token_accuracy": 0.6909661293029785, + "num_tokens": 32615477.0, + "step": 1265 + }, + { + "epoch": 0.13902921150889525, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.5584592819213867, + "learning_rate": 4.630307467057101e-07, + "loss": 1.0235, + "mean_token_accuracy": 0.6923514604568481, + "num_tokens": 32637654.0, + "step": 1266 + }, + { + "epoch": 0.1391390292115089, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.3057525157928467, + "learning_rate": 4.6339677891654465e-07, + "loss": 0.9997, + "mean_token_accuracy": 0.6996855735778809, + "num_tokens": 32661199.0, + "step": 1267 + }, + { + "epoch": 0.13924884691412256, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.5238568782806396, + "learning_rate": 4.637628111273792e-07, + "loss": 1.095, + "mean_token_accuracy": 0.6758226156234741, + "num_tokens": 32684302.0, + "step": 1268 + }, + { + "epoch": 0.1393586646167362, + "ewc_loss": 6.496906280517578e-06, + "grad_norm": 2.186373233795166, + "learning_rate": 4.6412884333821374e-07, + "loss": 1.1306, + "mean_token_accuracy": 0.6666498780250549, + "num_tokens": 32711546.0, + "step": 1269 + }, + { + "epoch": 0.13946848231934988, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1539885997772217, + "learning_rate": 4.644948755490483e-07, + "loss": 1.145, + "mean_token_accuracy": 0.6644124984741211, + "num_tokens": 32741165.0, + "step": 1270 + }, + { + "epoch": 0.13957830002196353, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0925228595733643, + "learning_rate": 4.6486090775988284e-07, + "loss": 1.0584, + "mean_token_accuracy": 0.6866005659103394, + "num_tokens": 32774701.0, + "step": 1271 + }, + { + "epoch": 0.1396881177245772, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4254255294799805, + "learning_rate": 4.652269399707174e-07, + "loss": 1.1439, + "mean_token_accuracy": 0.6661326885223389, + "num_tokens": 32798129.0, + "step": 1272 + }, + { + "epoch": 0.13979793542719085, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0687766075134277, + "learning_rate": 4.65592972181552e-07, + "loss": 1.0619, + "mean_token_accuracy": 0.6900408267974854, + "num_tokens": 32829323.0, + "step": 1273 + }, + { + "epoch": 0.13990775312980452, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2655441761016846, + "learning_rate": 4.659590043923865e-07, + "loss": 0.986, + "mean_token_accuracy": 0.7057834267616272, + "num_tokens": 32856609.0, + "step": 1274 + }, + { + "epoch": 0.1400175708324182, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3883543014526367, + "learning_rate": 4.663250366032211e-07, + "loss": 1.0391, + "mean_token_accuracy": 0.6859990358352661, + "num_tokens": 32881060.0, + "step": 1275 + }, + { + "epoch": 0.14012738853503184, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.346601724624634, + "learning_rate": 4.6669106881405564e-07, + "loss": 1.0043, + "mean_token_accuracy": 0.7122567892074585, + "num_tokens": 32904678.0, + "step": 1276 + }, + { + "epoch": 0.14023720623764552, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.6273703575134277, + "learning_rate": 4.6705710102489014e-07, + "loss": 1.0145, + "mean_token_accuracy": 0.6904207468032837, + "num_tokens": 32928285.0, + "step": 1277 + }, + { + "epoch": 0.14034702394025916, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.4312477111816406, + "learning_rate": 4.6742313323572474e-07, + "loss": 1.0458, + "mean_token_accuracy": 0.691638171672821, + "num_tokens": 32953717.0, + "step": 1278 + }, + { + "epoch": 0.14045684164287284, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.020986557006836, + "learning_rate": 4.677891654465593e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.695654034614563, + "num_tokens": 32985152.0, + "step": 1279 + }, + { + "epoch": 0.14056665934548648, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.450831890106201, + "learning_rate": 4.681551976573938e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6909104585647583, + "num_tokens": 33009420.0, + "step": 1280 + }, + { + "epoch": 0.14067647704810016, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1087913513183594, + "learning_rate": 4.685212298682284e-07, + "loss": 1.1671, + "mean_token_accuracy": 0.654872477054596, + "num_tokens": 33042302.0, + "step": 1281 + }, + { + "epoch": 0.1407862947507138, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1009507179260254, + "learning_rate": 4.6888726207906293e-07, + "loss": 1.0596, + "mean_token_accuracy": 0.6825158596038818, + "num_tokens": 33072431.0, + "step": 1282 + }, + { + "epoch": 0.14089611245332748, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2688934803009033, + "learning_rate": 4.6925329428989754e-07, + "loss": 1.1046, + "mean_token_accuracy": 0.669779896736145, + "num_tokens": 33098031.0, + "step": 1283 + }, + { + "epoch": 0.14100593015594115, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.4044241905212402, + "learning_rate": 4.6961932650073203e-07, + "loss": 1.0885, + "mean_token_accuracy": 0.6852965354919434, + "num_tokens": 33121574.0, + "step": 1284 + }, + { + "epoch": 0.1411157478585548, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.546079158782959, + "learning_rate": 4.699853587115666e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.6825817823410034, + "num_tokens": 33145414.0, + "step": 1285 + }, + { + "epoch": 0.14122556556116847, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.5014455318450928, + "learning_rate": 4.703513909224012e-07, + "loss": 0.9992, + "mean_token_accuracy": 0.6990571022033691, + "num_tokens": 33166555.0, + "step": 1286 + }, + { + "epoch": 0.14133538326378212, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.00174617767334, + "learning_rate": 4.707174231332357e-07, + "loss": 1.1246, + "mean_token_accuracy": 0.6707382202148438, + "num_tokens": 33200074.0, + "step": 1287 + }, + { + "epoch": 0.1414452009663958, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.617157220840454, + "learning_rate": 4.7108345534407023e-07, + "loss": 0.9002, + "mean_token_accuracy": 0.7287955284118652, + "num_tokens": 33220090.0, + "step": 1288 + }, + { + "epoch": 0.14155501866900944, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2505953311920166, + "learning_rate": 4.7144948755490483e-07, + "loss": 1.018, + "mean_token_accuracy": 0.6895627379417419, + "num_tokens": 33247080.0, + "step": 1289 + }, + { + "epoch": 0.1416648363716231, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.0977184772491455, + "learning_rate": 4.718155197657394e-07, + "loss": 1.1157, + "mean_token_accuracy": 0.6776061654090881, + "num_tokens": 33276588.0, + "step": 1290 + }, + { + "epoch": 0.14177465407423676, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.6437008380889893, + "learning_rate": 4.7218155197657393e-07, + "loss": 0.9712, + "mean_token_accuracy": 0.7057344913482666, + "num_tokens": 33296291.0, + "step": 1291 + }, + { + "epoch": 0.14188447177685043, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.154770851135254, + "learning_rate": 4.725475841874085e-07, + "loss": 1.1215, + "mean_token_accuracy": 0.6678204536437988, + "num_tokens": 33327544.0, + "step": 1292 + }, + { + "epoch": 0.1419942894794641, + "ewc_loss": 6.616115570068359e-06, + "grad_norm": 2.006413698196411, + "learning_rate": 4.7291361639824303e-07, + "loss": 1.0469, + "mean_token_accuracy": 0.6920343637466431, + "num_tokens": 33360733.0, + "step": 1293 + }, + { + "epoch": 0.14210410718207775, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.253650188446045, + "learning_rate": 4.732796486090776e-07, + "loss": 1.0175, + "mean_token_accuracy": 0.6979045867919922, + "num_tokens": 33386974.0, + "step": 1294 + }, + { + "epoch": 0.14221392488469142, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.461980104446411, + "learning_rate": 4.736456808199121e-07, + "loss": 1.0461, + "mean_token_accuracy": 0.6989670991897583, + "num_tokens": 33410360.0, + "step": 1295 + }, + { + "epoch": 0.14232374258730507, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2970540523529053, + "learning_rate": 4.740117130307467e-07, + "loss": 1.0452, + "mean_token_accuracy": 0.6880192756652832, + "num_tokens": 33435322.0, + "step": 1296 + }, + { + "epoch": 0.14243356028991874, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.266538143157959, + "learning_rate": 4.743777452415813e-07, + "loss": 1.2061, + "mean_token_accuracy": 0.6537102460861206, + "num_tokens": 33464229.0, + "step": 1297 + }, + { + "epoch": 0.1425433779925324, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.5044116973876953, + "learning_rate": 4.7474377745241577e-07, + "loss": 0.9526, + "mean_token_accuracy": 0.718380331993103, + "num_tokens": 33485248.0, + "step": 1298 + }, + { + "epoch": 0.14265319569514606, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.225224494934082, + "learning_rate": 4.751098096632504e-07, + "loss": 1.0547, + "mean_token_accuracy": 0.6869848966598511, + "num_tokens": 33516065.0, + "step": 1299 + }, + { + "epoch": 0.1427630133977597, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.0089173316955566, + "learning_rate": 4.754758418740849e-07, + "loss": 0.997, + "mean_token_accuracy": 0.7065190076828003, + "num_tokens": 33548790.0, + "step": 1300 + }, + { + "epoch": 0.14287283110037338, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.1916987895965576, + "learning_rate": 4.758418740849194e-07, + "loss": 1.0669, + "mean_token_accuracy": 0.6875808238983154, + "num_tokens": 33577167.0, + "step": 1301 + }, + { + "epoch": 0.14298264880298703, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.424302339553833, + "learning_rate": 4.76207906295754e-07, + "loss": 1.0385, + "mean_token_accuracy": 0.6969317197799683, + "num_tokens": 33599933.0, + "step": 1302 + }, + { + "epoch": 0.1430924665056007, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.573801040649414, + "learning_rate": 4.7657393850658857e-07, + "loss": 0.9519, + "mean_token_accuracy": 0.7093126773834229, + "num_tokens": 33618648.0, + "step": 1303 + }, + { + "epoch": 0.14320228420821438, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.311896800994873, + "learning_rate": 4.769399707174231e-07, + "loss": 1.0235, + "mean_token_accuracy": 0.6935432553291321, + "num_tokens": 33643320.0, + "step": 1304 + }, + { + "epoch": 0.14331210191082802, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.3769948482513428, + "learning_rate": 4.773060029282577e-07, + "loss": 1.1093, + "mean_token_accuracy": 0.6691390872001648, + "num_tokens": 33667215.0, + "step": 1305 + }, + { + "epoch": 0.1434219196134417, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.4413022994995117, + "learning_rate": 4.776720351390922e-07, + "loss": 1.1185, + "mean_token_accuracy": 0.6673683524131775, + "num_tokens": 33691314.0, + "step": 1306 + }, + { + "epoch": 0.14353173731605534, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.442596197128296, + "learning_rate": 4.780380673499268e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.683364748954773, + "num_tokens": 33716231.0, + "step": 1307 + }, + { + "epoch": 0.14364155501866901, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 1.9547094106674194, + "learning_rate": 4.784040995607613e-07, + "loss": 1.0662, + "mean_token_accuracy": 0.6822735071182251, + "num_tokens": 33750223.0, + "step": 1308 + }, + { + "epoch": 0.14375137272128266, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.46911883354187, + "learning_rate": 4.787701317715958e-07, + "loss": 0.9965, + "mean_token_accuracy": 0.7060454487800598, + "num_tokens": 33772320.0, + "step": 1309 + }, + { + "epoch": 0.14386119042389633, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.587792158126831, + "learning_rate": 4.791361639824304e-07, + "loss": 1.0456, + "mean_token_accuracy": 0.6898016929626465, + "num_tokens": 33793342.0, + "step": 1310 + }, + { + "epoch": 0.14397100812650998, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.6079626083374023, + "learning_rate": 4.79502196193265e-07, + "loss": 0.9617, + "mean_token_accuracy": 0.705868124961853, + "num_tokens": 33813186.0, + "step": 1311 + }, + { + "epoch": 0.14408082582912365, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.283519983291626, + "learning_rate": 4.798682284040995e-07, + "loss": 1.0104, + "mean_token_accuracy": 0.6918459534645081, + "num_tokens": 33840214.0, + "step": 1312 + }, + { + "epoch": 0.14419064353173733, + "ewc_loss": 6.765127182006836e-06, + "grad_norm": 2.588059425354004, + "learning_rate": 4.802342606149341e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6889120936393738, + "num_tokens": 33863647.0, + "step": 1313 + }, + { + "epoch": 0.14430046123435097, + "ewc_loss": 6.8247318267822266e-06, + "grad_norm": 2.045402765274048, + "learning_rate": 4.806002928257686e-07, + "loss": 1.1013, + "mean_token_accuracy": 0.6790891885757446, + "num_tokens": 33896827.0, + "step": 1314 + }, + { + "epoch": 0.14441027893696465, + "ewc_loss": 6.8247318267822266e-06, + "grad_norm": 2.3970396518707275, + "learning_rate": 4.809663250366032e-07, + "loss": 1.0619, + "mean_token_accuracy": 0.6807504892349243, + "num_tokens": 33920697.0, + "step": 1315 + }, + { + "epoch": 0.1445200966395783, + "ewc_loss": 6.8247318267822266e-06, + "grad_norm": 2.2633371353149414, + "learning_rate": 4.813323572474377e-07, + "loss": 1.0188, + "mean_token_accuracy": 0.6988052725791931, + "num_tokens": 33945998.0, + "step": 1316 + }, + { + "epoch": 0.14462991434219197, + "ewc_loss": 6.8247318267822266e-06, + "grad_norm": 2.447392463684082, + "learning_rate": 4.816983894582723e-07, + "loss": 1.0379, + "mean_token_accuracy": 0.696499228477478, + "num_tokens": 33969471.0, + "step": 1317 + }, + { + "epoch": 0.1447397320448056, + "ewc_loss": 6.8247318267822266e-06, + "grad_norm": 2.287705183029175, + "learning_rate": 4.820644216691069e-07, + "loss": 1.0718, + "mean_token_accuracy": 0.6781900525093079, + "num_tokens": 33995373.0, + "step": 1318 + }, + { + "epoch": 0.1448495497474193, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.494624376296997, + "learning_rate": 4.824304538799414e-07, + "loss": 0.9736, + "mean_token_accuracy": 0.7075813412666321, + "num_tokens": 34019470.0, + "step": 1319 + }, + { + "epoch": 0.14495936745003293, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.285787582397461, + "learning_rate": 4.827964860907759e-07, + "loss": 1.093, + "mean_token_accuracy": 0.6780679821968079, + "num_tokens": 34046792.0, + "step": 1320 + }, + { + "epoch": 0.1450691851526466, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.264975070953369, + "learning_rate": 4.831625183016105e-07, + "loss": 1.071, + "mean_token_accuracy": 0.6794407963752747, + "num_tokens": 34073908.0, + "step": 1321 + }, + { + "epoch": 0.14517900285526028, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.1600983142852783, + "learning_rate": 4.835285505124451e-07, + "loss": 0.9328, + "mean_token_accuracy": 0.7146381735801697, + "num_tokens": 34101665.0, + "step": 1322 + }, + { + "epoch": 0.14528882055787393, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.3907454013824463, + "learning_rate": 4.838945827232796e-07, + "loss": 0.9593, + "mean_token_accuracy": 0.7110143899917603, + "num_tokens": 34123530.0, + "step": 1323 + }, + { + "epoch": 0.1453986382604876, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.41677188873291, + "learning_rate": 4.842606149341142e-07, + "loss": 1.0511, + "mean_token_accuracy": 0.677916407585144, + "num_tokens": 34148038.0, + "step": 1324 + }, + { + "epoch": 0.14550845596310125, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.4084408283233643, + "learning_rate": 4.846266471449487e-07, + "loss": 1.0552, + "mean_token_accuracy": 0.6820454597473145, + "num_tokens": 34171039.0, + "step": 1325 + }, + { + "epoch": 0.14561827366571492, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.2974722385406494, + "learning_rate": 4.849926793557833e-07, + "loss": 1.0586, + "mean_token_accuracy": 0.6827704906463623, + "num_tokens": 34197557.0, + "step": 1326 + }, + { + "epoch": 0.14572809136832857, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4802134037017822, + "learning_rate": 4.853587115666178e-07, + "loss": 0.9981, + "mean_token_accuracy": 0.6962355375289917, + "num_tokens": 34219659.0, + "step": 1327 + }, + { + "epoch": 0.14583790907094224, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.5531744956970215, + "learning_rate": 4.857247437774524e-07, + "loss": 1.0622, + "mean_token_accuracy": 0.68172287940979, + "num_tokens": 34241179.0, + "step": 1328 + }, + { + "epoch": 0.14594772677355589, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.545456647872925, + "learning_rate": 4.86090775988287e-07, + "loss": 0.9425, + "mean_token_accuracy": 0.7167631387710571, + "num_tokens": 34261000.0, + "step": 1329 + }, + { + "epoch": 0.14605754447616956, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4511048793792725, + "learning_rate": 4.864568081991215e-07, + "loss": 1.0494, + "mean_token_accuracy": 0.686025083065033, + "num_tokens": 34285817.0, + "step": 1330 + }, + { + "epoch": 0.14616736217878323, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.17692494392395, + "learning_rate": 4.868228404099561e-07, + "loss": 1.1331, + "mean_token_accuracy": 0.6655007600784302, + "num_tokens": 34315537.0, + "step": 1331 + }, + { + "epoch": 0.14627717988139688, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3479065895080566, + "learning_rate": 4.871888726207906e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6855816841125488, + "num_tokens": 34338741.0, + "step": 1332 + }, + { + "epoch": 0.14638699758401055, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.404768466949463, + "learning_rate": 4.875549048316251e-07, + "loss": 1.0359, + "mean_token_accuracy": 0.6921002864837646, + "num_tokens": 34362546.0, + "step": 1333 + }, + { + "epoch": 0.1464968152866242, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 1.9889787435531616, + "learning_rate": 4.879209370424597e-07, + "loss": 1.129, + "mean_token_accuracy": 0.666056752204895, + "num_tokens": 34394885.0, + "step": 1334 + }, + { + "epoch": 0.14660663298923787, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.0308756828308105, + "learning_rate": 4.882869692532943e-07, + "loss": 1.0782, + "mean_token_accuracy": 0.6732567548751831, + "num_tokens": 34425299.0, + "step": 1335 + }, + { + "epoch": 0.14671645069185152, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.3873050212860107, + "learning_rate": 4.886530014641288e-07, + "loss": 1.0358, + "mean_token_accuracy": 0.6929057240486145, + "num_tokens": 34448680.0, + "step": 1336 + }, + { + "epoch": 0.1468262683944652, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.2773730754852295, + "learning_rate": 4.890190336749634e-07, + "loss": 1.0857, + "mean_token_accuracy": 0.6819236278533936, + "num_tokens": 34474831.0, + "step": 1337 + }, + { + "epoch": 0.14693608609707884, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.348048210144043, + "learning_rate": 4.893850658857979e-07, + "loss": 0.9189, + "mean_token_accuracy": 0.7204455733299255, + "num_tokens": 34496472.0, + "step": 1338 + }, + { + "epoch": 0.1470459037996925, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.5571961402893066, + "learning_rate": 4.897510980966325e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.713291347026825, + "num_tokens": 34519372.0, + "step": 1339 + }, + { + "epoch": 0.14715572150230616, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.297877788543701, + "learning_rate": 4.90117130307467e-07, + "loss": 1.0803, + "mean_token_accuracy": 0.6915773153305054, + "num_tokens": 34545661.0, + "step": 1340 + }, + { + "epoch": 0.14726553920491983, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.1723198890686035, + "learning_rate": 4.904831625183016e-07, + "loss": 1.1636, + "mean_token_accuracy": 0.6640567779541016, + "num_tokens": 34576126.0, + "step": 1341 + }, + { + "epoch": 0.1473753569075335, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3928513526916504, + "learning_rate": 4.908491947291362e-07, + "loss": 1.0841, + "mean_token_accuracy": 0.6806415319442749, + "num_tokens": 34600086.0, + "step": 1342 + }, + { + "epoch": 0.14748517461014715, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2913994789123535, + "learning_rate": 4.912152269399707e-07, + "loss": 0.9962, + "mean_token_accuracy": 0.7040925621986389, + "num_tokens": 34626401.0, + "step": 1343 + }, + { + "epoch": 0.14759499231276083, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.273259401321411, + "learning_rate": 4.915812591508052e-07, + "loss": 1.0928, + "mean_token_accuracy": 0.6759775280952454, + "num_tokens": 34653201.0, + "step": 1344 + }, + { + "epoch": 0.14770481001537447, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2803564071655273, + "learning_rate": 4.919472913616398e-07, + "loss": 1.0752, + "mean_token_accuracy": 0.6791417002677917, + "num_tokens": 34680307.0, + "step": 1345 + }, + { + "epoch": 0.14781462771798815, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.486570358276367, + "learning_rate": 4.923133235724744e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.6955829858779907, + "num_tokens": 34702531.0, + "step": 1346 + }, + { + "epoch": 0.1479244454206018, + "ewc_loss": 7.12275505065918e-06, + "grad_norm": 2.0315144062042236, + "learning_rate": 4.926793557833089e-07, + "loss": 1.0603, + "mean_token_accuracy": 0.6850127577781677, + "num_tokens": 34733708.0, + "step": 1347 + }, + { + "epoch": 0.14803426312321546, + "ewc_loss": 7.12275505065918e-06, + "grad_norm": 2.431403160095215, + "learning_rate": 4.930453879941435e-07, + "loss": 1.0907, + "mean_token_accuracy": 0.676846981048584, + "num_tokens": 34758625.0, + "step": 1348 + }, + { + "epoch": 0.1481440808258291, + "ewc_loss": 7.12275505065918e-06, + "grad_norm": 2.204160213470459, + "learning_rate": 4.93411420204978e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6952851414680481, + "num_tokens": 34786952.0, + "step": 1349 + }, + { + "epoch": 0.14825389852844278, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.235877752304077, + "learning_rate": 4.937774524158126e-07, + "loss": 1.1152, + "mean_token_accuracy": 0.6716836094856262, + "num_tokens": 34813181.0, + "step": 1350 + }, + { + "epoch": 0.14836371623105646, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1294071674346924, + "learning_rate": 4.941434846266471e-07, + "loss": 1.0097, + "mean_token_accuracy": 0.6944843530654907, + "num_tokens": 34845742.0, + "step": 1351 + }, + { + "epoch": 0.1484735339336701, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.255873441696167, + "learning_rate": 4.945095168374817e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6840115785598755, + "num_tokens": 34872515.0, + "step": 1352 + }, + { + "epoch": 0.14858335163628378, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.411616086959839, + "learning_rate": 4.948755490483163e-07, + "loss": 1.1112, + "mean_token_accuracy": 0.6680577993392944, + "num_tokens": 34899778.0, + "step": 1353 + }, + { + "epoch": 0.14869316933889742, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.227792263031006, + "learning_rate": 4.952415812591508e-07, + "loss": 0.9954, + "mean_token_accuracy": 0.69472336769104, + "num_tokens": 34926103.0, + "step": 1354 + }, + { + "epoch": 0.1488029870415111, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.165418863296509, + "learning_rate": 4.956076134699854e-07, + "loss": 0.9656, + "mean_token_accuracy": 0.7079997062683105, + "num_tokens": 34952471.0, + "step": 1355 + }, + { + "epoch": 0.14891280474412474, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.0837764739990234, + "learning_rate": 4.959736456808199e-07, + "loss": 1.0944, + "mean_token_accuracy": 0.6713990569114685, + "num_tokens": 34981653.0, + "step": 1356 + }, + { + "epoch": 0.14902262244673842, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.462489128112793, + "learning_rate": 4.963396778916544e-07, + "loss": 0.9799, + "mean_token_accuracy": 0.7099179625511169, + "num_tokens": 35004160.0, + "step": 1357 + }, + { + "epoch": 0.14913244014935206, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.199028968811035, + "learning_rate": 4.96705710102489e-07, + "loss": 1.0765, + "mean_token_accuracy": 0.6853011250495911, + "num_tokens": 35033695.0, + "step": 1358 + }, + { + "epoch": 0.14924225785196574, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3113229274749756, + "learning_rate": 4.970717423133236e-07, + "loss": 1.0078, + "mean_token_accuracy": 0.691547691822052, + "num_tokens": 35058960.0, + "step": 1359 + }, + { + "epoch": 0.1493520755545794, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.287461280822754, + "learning_rate": 4.974377745241581e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.6928304433822632, + "num_tokens": 35082458.0, + "step": 1360 + }, + { + "epoch": 0.14946189325719306, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.174637794494629, + "learning_rate": 4.978038067349927e-07, + "loss": 1.0704, + "mean_token_accuracy": 0.6902957558631897, + "num_tokens": 35111320.0, + "step": 1361 + }, + { + "epoch": 0.14957171095980673, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.485502004623413, + "learning_rate": 4.981698389458272e-07, + "loss": 1.0981, + "mean_token_accuracy": 0.66917884349823, + "num_tokens": 35136034.0, + "step": 1362 + }, + { + "epoch": 0.14968152866242038, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3748042583465576, + "learning_rate": 4.985358711566618e-07, + "loss": 1.0415, + "mean_token_accuracy": 0.6919581890106201, + "num_tokens": 35160498.0, + "step": 1363 + }, + { + "epoch": 0.14979134636503405, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0686795711517334, + "learning_rate": 4.989019033674963e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6838853359222412, + "num_tokens": 35191074.0, + "step": 1364 + }, + { + "epoch": 0.1499011640676477, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.345141887664795, + "learning_rate": 4.992679355783309e-07, + "loss": 1.0501, + "mean_token_accuracy": 0.6916592717170715, + "num_tokens": 35216079.0, + "step": 1365 + }, + { + "epoch": 0.15001098177026137, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4572770595550537, + "learning_rate": 4.996339677891655e-07, + "loss": 1.0532, + "mean_token_accuracy": 0.6946589946746826, + "num_tokens": 35239149.0, + "step": 1366 + }, + { + "epoch": 0.15012079947287502, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2249596118927, + "learning_rate": 5e-07, + "loss": 1.1023, + "mean_token_accuracy": 0.6789478063583374, + "num_tokens": 35267266.0, + "step": 1367 + }, + { + "epoch": 0.1502306171754887, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.206629991531372, + "learning_rate": 5.003660322108345e-07, + "loss": 1.1089, + "mean_token_accuracy": 0.6757527589797974, + "num_tokens": 35295774.0, + "step": 1368 + }, + { + "epoch": 0.15034043487810236, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.6747512817382812, + "learning_rate": 5.007320644216691e-07, + "loss": 1.0371, + "mean_token_accuracy": 0.6882256269454956, + "num_tokens": 35315866.0, + "step": 1369 + }, + { + "epoch": 0.150450252580716, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4020333290100098, + "learning_rate": 5.010980966325037e-07, + "loss": 1.072, + "mean_token_accuracy": 0.6880179643630981, + "num_tokens": 35340461.0, + "step": 1370 + }, + { + "epoch": 0.15056007028332968, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1260695457458496, + "learning_rate": 5.014641288433382e-07, + "loss": 0.9821, + "mean_token_accuracy": 0.7030900716781616, + "num_tokens": 35367551.0, + "step": 1371 + }, + { + "epoch": 0.15066988798594333, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.146728515625, + "learning_rate": 5.018301610541728e-07, + "loss": 1.1121, + "mean_token_accuracy": 0.6658934950828552, + "num_tokens": 35397737.0, + "step": 1372 + }, + { + "epoch": 0.150779705688557, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4229586124420166, + "learning_rate": 5.021961932650073e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.6943272352218628, + "num_tokens": 35422047.0, + "step": 1373 + }, + { + "epoch": 0.15088952339117065, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.1595075130462646, + "learning_rate": 5.025622254758418e-07, + "loss": 1.0496, + "mean_token_accuracy": 0.6877328157424927, + "num_tokens": 35450974.0, + "step": 1374 + }, + { + "epoch": 0.15099934109378432, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.1966495513916016, + "learning_rate": 5.029282576866764e-07, + "loss": 0.9853, + "mean_token_accuracy": 0.7046564221382141, + "num_tokens": 35478196.0, + "step": 1375 + }, + { + "epoch": 0.15110915879639797, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.1236214637756348, + "learning_rate": 5.03294289897511e-07, + "loss": 1.1135, + "mean_token_accuracy": 0.6686729192733765, + "num_tokens": 35510077.0, + "step": 1376 + }, + { + "epoch": 0.15121897649901164, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.2756428718566895, + "learning_rate": 5.036603221083456e-07, + "loss": 1.0613, + "mean_token_accuracy": 0.6845524311065674, + "num_tokens": 35538238.0, + "step": 1377 + }, + { + "epoch": 0.1513287942016253, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.146796464920044, + "learning_rate": 5.040263543191801e-07, + "loss": 1.109, + "mean_token_accuracy": 0.6690082550048828, + "num_tokens": 35569282.0, + "step": 1378 + }, + { + "epoch": 0.15143861190423896, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.360944986343384, + "learning_rate": 5.043923865300146e-07, + "loss": 1.1512, + "mean_token_accuracy": 0.671506404876709, + "num_tokens": 35595162.0, + "step": 1379 + }, + { + "epoch": 0.15154842960685264, + "ewc_loss": 7.3015689849853516e-06, + "grad_norm": 2.3249716758728027, + "learning_rate": 5.047584187408492e-07, + "loss": 0.9933, + "mean_token_accuracy": 0.6979003548622131, + "num_tokens": 35620173.0, + "step": 1380 + }, + { + "epoch": 0.15165824730946628, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.5477612018585205, + "learning_rate": 5.051244509516838e-07, + "loss": 0.9841, + "mean_token_accuracy": 0.7039878368377686, + "num_tokens": 35641699.0, + "step": 1381 + }, + { + "epoch": 0.15176806501207996, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.1727869510650635, + "learning_rate": 5.054904831625183e-07, + "loss": 1.0934, + "mean_token_accuracy": 0.6738007068634033, + "num_tokens": 35672048.0, + "step": 1382 + }, + { + "epoch": 0.1518778827146936, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.331286907196045, + "learning_rate": 5.058565153733529e-07, + "loss": 1.0887, + "mean_token_accuracy": 0.6788663268089294, + "num_tokens": 35697483.0, + "step": 1383 + }, + { + "epoch": 0.15198770041730728, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.321134567260742, + "learning_rate": 5.062225475841874e-07, + "loss": 1.0585, + "mean_token_accuracy": 0.6972566843032837, + "num_tokens": 35723408.0, + "step": 1384 + }, + { + "epoch": 0.15209751811992092, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.32002329826355, + "learning_rate": 5.065885797950219e-07, + "loss": 1.029, + "mean_token_accuracy": 0.6911610960960388, + "num_tokens": 35747189.0, + "step": 1385 + }, + { + "epoch": 0.1522073358225346, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.117720127105713, + "learning_rate": 5.069546120058566e-07, + "loss": 1.0577, + "mean_token_accuracy": 0.6821931004524231, + "num_tokens": 35777084.0, + "step": 1386 + }, + { + "epoch": 0.15231715352514824, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.349174976348877, + "learning_rate": 5.073206442166911e-07, + "loss": 0.9718, + "mean_token_accuracy": 0.7091590762138367, + "num_tokens": 35801318.0, + "step": 1387 + }, + { + "epoch": 0.15242697122776191, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.779588460922241, + "learning_rate": 5.076866764275256e-07, + "loss": 0.9906, + "mean_token_accuracy": 0.7054678797721863, + "num_tokens": 35820852.0, + "step": 1388 + }, + { + "epoch": 0.1525367889303756, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.321803331375122, + "learning_rate": 5.080527086383602e-07, + "loss": 1.0808, + "mean_token_accuracy": 0.6767691969871521, + "num_tokens": 35846653.0, + "step": 1389 + }, + { + "epoch": 0.15264660663298923, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0766873359680176, + "learning_rate": 5.084187408491947e-07, + "loss": 1.0871, + "mean_token_accuracy": 0.6762368679046631, + "num_tokens": 35877674.0, + "step": 1390 + }, + { + "epoch": 0.1527564243356029, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1023035049438477, + "learning_rate": 5.087847730600292e-07, + "loss": 1.0804, + "mean_token_accuracy": 0.6804662942886353, + "num_tokens": 35908450.0, + "step": 1391 + }, + { + "epoch": 0.15286624203821655, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.2567148208618164, + "learning_rate": 5.091508052708639e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6931682825088501, + "num_tokens": 35934858.0, + "step": 1392 + }, + { + "epoch": 0.15297605974083023, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.262636661529541, + "learning_rate": 5.095168374816984e-07, + "loss": 1.0476, + "mean_token_accuracy": 0.6857753992080688, + "num_tokens": 35962276.0, + "step": 1393 + }, + { + "epoch": 0.15308587744344387, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.524569511413574, + "learning_rate": 5.09882869692533e-07, + "loss": 1.1717, + "mean_token_accuracy": 0.6625354290008545, + "num_tokens": 35987873.0, + "step": 1394 + }, + { + "epoch": 0.15319569514605755, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4261648654937744, + "learning_rate": 5.102489019033675e-07, + "loss": 1.046, + "mean_token_accuracy": 0.6865522861480713, + "num_tokens": 36013354.0, + "step": 1395 + }, + { + "epoch": 0.1533055128486712, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2013983726501465, + "learning_rate": 5.10614934114202e-07, + "loss": 1.063, + "mean_token_accuracy": 0.6826717853546143, + "num_tokens": 36040211.0, + "step": 1396 + }, + { + "epoch": 0.15341533055128487, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2429888248443604, + "learning_rate": 5.109809663250367e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6773141622543335, + "num_tokens": 36066374.0, + "step": 1397 + }, + { + "epoch": 0.15352514825389854, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.3564422130584717, + "learning_rate": 5.113469985358712e-07, + "loss": 0.9892, + "mean_token_accuracy": 0.699659526348114, + "num_tokens": 36092114.0, + "step": 1398 + }, + { + "epoch": 0.1536349659565122, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.441542148590088, + "learning_rate": 5.117130307467057e-07, + "loss": 1.0513, + "mean_token_accuracy": 0.6899124979972839, + "num_tokens": 36117150.0, + "step": 1399 + }, + { + "epoch": 0.15374478365912586, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.131645441055298, + "learning_rate": 5.120790629575403e-07, + "loss": 1.1445, + "mean_token_accuracy": 0.6670714616775513, + "num_tokens": 36146676.0, + "step": 1400 + }, + { + "epoch": 0.1538546013617395, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3438663482666016, + "learning_rate": 5.124450951683748e-07, + "loss": 1.0994, + "mean_token_accuracy": 0.6728843450546265, + "num_tokens": 36173244.0, + "step": 1401 + }, + { + "epoch": 0.15396441906435318, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.1604788303375244, + "learning_rate": 5.128111273792094e-07, + "loss": 1.1017, + "mean_token_accuracy": 0.6779448986053467, + "num_tokens": 36203197.0, + "step": 1402 + }, + { + "epoch": 0.15407423676696683, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.2325403690338135, + "learning_rate": 5.13177159590044e-07, + "loss": 1.1019, + "mean_token_accuracy": 0.6743413209915161, + "num_tokens": 36231276.0, + "step": 1403 + }, + { + "epoch": 0.1541840544695805, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.069096565246582, + "learning_rate": 5.135431918008785e-07, + "loss": 1.067, + "mean_token_accuracy": 0.6907994747161865, + "num_tokens": 36261923.0, + "step": 1404 + }, + { + "epoch": 0.15429387217219415, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.415792465209961, + "learning_rate": 5.13909224011713e-07, + "loss": 0.9812, + "mean_token_accuracy": 0.710612952709198, + "num_tokens": 36285200.0, + "step": 1405 + }, + { + "epoch": 0.15440368987480782, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.271547555923462, + "learning_rate": 5.142752562225476e-07, + "loss": 1.0684, + "mean_token_accuracy": 0.6800133585929871, + "num_tokens": 36310732.0, + "step": 1406 + }, + { + "epoch": 0.1545135075774215, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.4456329345703125, + "learning_rate": 5.14641288433382e-07, + "loss": 0.9957, + "mean_token_accuracy": 0.7004735469818115, + "num_tokens": 36333833.0, + "step": 1407 + }, + { + "epoch": 0.15462332528003514, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3423190116882324, + "learning_rate": 5.150073206442168e-07, + "loss": 1.0062, + "mean_token_accuracy": 0.7038743495941162, + "num_tokens": 36357548.0, + "step": 1408 + }, + { + "epoch": 0.1547331429826488, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1440720558166504, + "learning_rate": 5.153733528550513e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.707294225692749, + "num_tokens": 36386854.0, + "step": 1409 + }, + { + "epoch": 0.15484296068526246, + "ewc_loss": 7.569789886474609e-06, + "grad_norm": 2.012005567550659, + "learning_rate": 5.157393850658857e-07, + "loss": 1.1586, + "mean_token_accuracy": 0.662510871887207, + "num_tokens": 36425311.0, + "step": 1410 + }, + { + "epoch": 0.15495277838787613, + "ewc_loss": 7.569789886474609e-06, + "grad_norm": 2.0785696506500244, + "learning_rate": 5.161054172767203e-07, + "loss": 0.9749, + "mean_token_accuracy": 0.7014234066009521, + "num_tokens": 36453543.0, + "step": 1411 + }, + { + "epoch": 0.15506259609048978, + "ewc_loss": 7.569789886474609e-06, + "grad_norm": 2.414982318878174, + "learning_rate": 5.164714494875548e-07, + "loss": 1.0841, + "mean_token_accuracy": 0.6880662441253662, + "num_tokens": 36479897.0, + "step": 1412 + }, + { + "epoch": 0.15517241379310345, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3229384422302246, + "learning_rate": 5.168374816983894e-07, + "loss": 1.078, + "mean_token_accuracy": 0.6819314360618591, + "num_tokens": 36508596.0, + "step": 1413 + }, + { + "epoch": 0.1552822314957171, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.4758994579315186, + "learning_rate": 5.17203513909224e-07, + "loss": 1.0732, + "mean_token_accuracy": 0.6793698072433472, + "num_tokens": 36532995.0, + "step": 1414 + }, + { + "epoch": 0.15539204919833077, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.119961977005005, + "learning_rate": 5.175695461200585e-07, + "loss": 1.0829, + "mean_token_accuracy": 0.6790528297424316, + "num_tokens": 36565082.0, + "step": 1415 + }, + { + "epoch": 0.15550186690094442, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.227818250656128, + "learning_rate": 5.17935578330893e-07, + "loss": 1.1606, + "mean_token_accuracy": 0.6559348106384277, + "num_tokens": 36595531.0, + "step": 1416 + }, + { + "epoch": 0.1556116846035581, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.4267752170562744, + "learning_rate": 5.183016105417276e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.6991986036300659, + "num_tokens": 36618337.0, + "step": 1417 + }, + { + "epoch": 0.15572150230617177, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.570723295211792, + "learning_rate": 5.186676427525622e-07, + "loss": 1.1128, + "mean_token_accuracy": 0.6688613295555115, + "num_tokens": 36641123.0, + "step": 1418 + }, + { + "epoch": 0.1558313200087854, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3378989696502686, + "learning_rate": 5.190336749633967e-07, + "loss": 1.0599, + "mean_token_accuracy": 0.6914776563644409, + "num_tokens": 36667870.0, + "step": 1419 + }, + { + "epoch": 0.15594113771139909, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.183537006378174, + "learning_rate": 5.193997071742313e-07, + "loss": 1.0469, + "mean_token_accuracy": 0.6894981861114502, + "num_tokens": 36697458.0, + "step": 1420 + }, + { + "epoch": 0.15605095541401273, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.132988929748535, + "learning_rate": 5.197657393850658e-07, + "loss": 1.1007, + "mean_token_accuracy": 0.6784920692443848, + "num_tokens": 36728999.0, + "step": 1421 + }, + { + "epoch": 0.1561607731166264, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.8966479301452637, + "learning_rate": 5.201317715959003e-07, + "loss": 1.1379, + "mean_token_accuracy": 0.6669362783432007, + "num_tokens": 36751360.0, + "step": 1422 + }, + { + "epoch": 0.15627059081924005, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.283778190612793, + "learning_rate": 5.204978038067349e-07, + "loss": 1.0955, + "mean_token_accuracy": 0.6811428666114807, + "num_tokens": 36778609.0, + "step": 1423 + }, + { + "epoch": 0.15638040852185373, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3672194480895996, + "learning_rate": 5.208638360175695e-07, + "loss": 1.0173, + "mean_token_accuracy": 0.6928259134292603, + "num_tokens": 36801124.0, + "step": 1424 + }, + { + "epoch": 0.15649022622446737, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.345038652420044, + "learning_rate": 5.212298682284041e-07, + "loss": 1.0972, + "mean_token_accuracy": 0.6737359762191772, + "num_tokens": 36826030.0, + "step": 1425 + }, + { + "epoch": 0.15660004392708105, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.21716570854187, + "learning_rate": 5.215959004392386e-07, + "loss": 1.1207, + "mean_token_accuracy": 0.6694073677062988, + "num_tokens": 36857373.0, + "step": 1426 + }, + { + "epoch": 0.15670986162969472, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3348097801208496, + "learning_rate": 5.219619326500731e-07, + "loss": 1.0219, + "mean_token_accuracy": 0.6968725919723511, + "num_tokens": 36882743.0, + "step": 1427 + }, + { + "epoch": 0.15681967933230836, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2253811359405518, + "learning_rate": 5.223279648609077e-07, + "loss": 1.0358, + "mean_token_accuracy": 0.6906787753105164, + "num_tokens": 36908317.0, + "step": 1428 + }, + { + "epoch": 0.15692949703492204, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.367823600769043, + "learning_rate": 5.226939970717423e-07, + "loss": 1.0685, + "mean_token_accuracy": 0.678511381149292, + "num_tokens": 36933934.0, + "step": 1429 + }, + { + "epoch": 0.15703931473753568, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.205221176147461, + "learning_rate": 5.230600292825768e-07, + "loss": 1.0065, + "mean_token_accuracy": 0.6963886022567749, + "num_tokens": 36961658.0, + "step": 1430 + }, + { + "epoch": 0.15714913244014936, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2668657302856445, + "learning_rate": 5.234260614934114e-07, + "loss": 1.0354, + "mean_token_accuracy": 0.6858406066894531, + "num_tokens": 36989890.0, + "step": 1431 + }, + { + "epoch": 0.157258950142763, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.5073647499084473, + "learning_rate": 5.237920937042459e-07, + "loss": 1.128, + "mean_token_accuracy": 0.6703557968139648, + "num_tokens": 37014359.0, + "step": 1432 + }, + { + "epoch": 0.15736876784537668, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.450974702835083, + "learning_rate": 5.241581259150804e-07, + "loss": 1.0338, + "mean_token_accuracy": 0.691864013671875, + "num_tokens": 37036466.0, + "step": 1433 + }, + { + "epoch": 0.15747858554799032, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.6135847568511963, + "learning_rate": 5.245241581259151e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7056369185447693, + "num_tokens": 37056557.0, + "step": 1434 + }, + { + "epoch": 0.157588403250604, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.397486686706543, + "learning_rate": 5.248901903367496e-07, + "loss": 1.0955, + "mean_token_accuracy": 0.6729074716567993, + "num_tokens": 37079546.0, + "step": 1435 + }, + { + "epoch": 0.15769822095321767, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.522988796234131, + "learning_rate": 5.252562225475841e-07, + "loss": 1.1214, + "mean_token_accuracy": 0.6707310080528259, + "num_tokens": 37103143.0, + "step": 1436 + }, + { + "epoch": 0.15780803865583132, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3056859970092773, + "learning_rate": 5.256222547584187e-07, + "loss": 0.9866, + "mean_token_accuracy": 0.7206137180328369, + "num_tokens": 37128958.0, + "step": 1437 + }, + { + "epoch": 0.157917856358445, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.6418087482452393, + "learning_rate": 5.259882869692532e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7065063714981079, + "num_tokens": 37148862.0, + "step": 1438 + }, + { + "epoch": 0.15802767406105864, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.6106746196746826, + "learning_rate": 5.263543191800877e-07, + "loss": 0.9408, + "mean_token_accuracy": 0.7130439877510071, + "num_tokens": 37168500.0, + "step": 1439 + }, + { + "epoch": 0.1581374917636723, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.5050547122955322, + "learning_rate": 5.267203513909224e-07, + "loss": 1.0649, + "mean_token_accuracy": 0.6819671392440796, + "num_tokens": 37191202.0, + "step": 1440 + }, + { + "epoch": 0.15824730946628596, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.4613211154937744, + "learning_rate": 5.270863836017569e-07, + "loss": 1.1218, + "mean_token_accuracy": 0.664158821105957, + "num_tokens": 37215939.0, + "step": 1441 + }, + { + "epoch": 0.15835712716889963, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2060251235961914, + "learning_rate": 5.274524158125915e-07, + "loss": 1.2381, + "mean_token_accuracy": 0.6608638763427734, + "num_tokens": 37246502.0, + "step": 1442 + }, + { + "epoch": 0.15846694487151328, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.450718641281128, + "learning_rate": 5.27818448023426e-07, + "loss": 0.9436, + "mean_token_accuracy": 0.7141901850700378, + "num_tokens": 37269519.0, + "step": 1443 + }, + { + "epoch": 0.15857676257412695, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.164257764816284, + "learning_rate": 5.281844802342605e-07, + "loss": 1.0129, + "mean_token_accuracy": 0.6998938322067261, + "num_tokens": 37295472.0, + "step": 1444 + }, + { + "epoch": 0.15868658027674062, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2437117099761963, + "learning_rate": 5.285505124450952e-07, + "loss": 1.0925, + "mean_token_accuracy": 0.6790200471878052, + "num_tokens": 37323089.0, + "step": 1445 + }, + { + "epoch": 0.15879639797935427, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.41237735748291, + "learning_rate": 5.289165446559297e-07, + "loss": 1.0994, + "mean_token_accuracy": 0.6697273850440979, + "num_tokens": 37349083.0, + "step": 1446 + }, + { + "epoch": 0.15890621568196794, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.522467851638794, + "learning_rate": 5.292825768667642e-07, + "loss": 0.9147, + "mean_token_accuracy": 0.7195100784301758, + "num_tokens": 37370736.0, + "step": 1447 + }, + { + "epoch": 0.1590160333845816, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.39809250831604, + "learning_rate": 5.296486090775988e-07, + "loss": 1.0758, + "mean_token_accuracy": 0.6816275119781494, + "num_tokens": 37395554.0, + "step": 1448 + }, + { + "epoch": 0.15912585108719526, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3758649826049805, + "learning_rate": 5.300146412884333e-07, + "loss": 1.0413, + "mean_token_accuracy": 0.6882361769676208, + "num_tokens": 37418863.0, + "step": 1449 + }, + { + "epoch": 0.1592356687898089, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.9408867359161377, + "learning_rate": 5.303806734992679e-07, + "loss": 1.033, + "mean_token_accuracy": 0.6838008165359497, + "num_tokens": 37435803.0, + "step": 1450 + }, + { + "epoch": 0.15934548649242258, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3111348152160645, + "learning_rate": 5.307467057101025e-07, + "loss": 1.1502, + "mean_token_accuracy": 0.6577318906784058, + "num_tokens": 37465254.0, + "step": 1451 + }, + { + "epoch": 0.15945530419503623, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.285736083984375, + "learning_rate": 5.31112737920937e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.7071664333343506, + "num_tokens": 37490392.0, + "step": 1452 + }, + { + "epoch": 0.1595651218976499, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3053369522094727, + "learning_rate": 5.314787701317715e-07, + "loss": 1.0452, + "mean_token_accuracy": 0.6873440742492676, + "num_tokens": 37515868.0, + "step": 1453 + }, + { + "epoch": 0.15967493960026355, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.396465301513672, + "learning_rate": 5.318448023426061e-07, + "loss": 0.9898, + "mean_token_accuracy": 0.7035793662071228, + "num_tokens": 37539607.0, + "step": 1454 + }, + { + "epoch": 0.15978475730287722, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.856513738632202, + "learning_rate": 5.322108345534406e-07, + "loss": 0.9223, + "mean_token_accuracy": 0.7160091400146484, + "num_tokens": 37557033.0, + "step": 1455 + }, + { + "epoch": 0.1598945750054909, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.396895170211792, + "learning_rate": 5.325768667642752e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7004570960998535, + "num_tokens": 37581118.0, + "step": 1456 + }, + { + "epoch": 0.16000439270810454, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3511271476745605, + "learning_rate": 5.329428989751098e-07, + "loss": 1.1317, + "mean_token_accuracy": 0.6677361130714417, + "num_tokens": 37609075.0, + "step": 1457 + }, + { + "epoch": 0.16011421041071822, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.133209705352783, + "learning_rate": 5.333089311859443e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.6893669366836548, + "num_tokens": 37636878.0, + "step": 1458 + }, + { + "epoch": 0.16022402811333186, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.133084535598755, + "learning_rate": 5.336749633967789e-07, + "loss": 1.0991, + "mean_token_accuracy": 0.6684989929199219, + "num_tokens": 37667267.0, + "step": 1459 + }, + { + "epoch": 0.16033384581594554, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.6518208980560303, + "learning_rate": 5.340409956076134e-07, + "loss": 1.0077, + "mean_token_accuracy": 0.7006366848945618, + "num_tokens": 37688240.0, + "step": 1460 + }, + { + "epoch": 0.16044366351855918, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.6340270042419434, + "learning_rate": 5.34407027818448e-07, + "loss": 0.9287, + "mean_token_accuracy": 0.7141275405883789, + "num_tokens": 37707416.0, + "step": 1461 + }, + { + "epoch": 0.16055348122117286, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.087679386138916, + "learning_rate": 5.347730600292826e-07, + "loss": 1.0779, + "mean_token_accuracy": 0.6835737824440002, + "num_tokens": 37740039.0, + "step": 1462 + }, + { + "epoch": 0.1606632989237865, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0824503898620605, + "learning_rate": 5.351390922401171e-07, + "loss": 1.1352, + "mean_token_accuracy": 0.6690278053283691, + "num_tokens": 37773476.0, + "step": 1463 + }, + { + "epoch": 0.16077311662640018, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1820287704467773, + "learning_rate": 5.355051244509516e-07, + "loss": 1.058, + "mean_token_accuracy": 0.6909478902816772, + "num_tokens": 37802625.0, + "step": 1464 + }, + { + "epoch": 0.16088293432901385, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7606148719787598, + "learning_rate": 5.358711566617862e-07, + "loss": 0.9618, + "mean_token_accuracy": 0.7082850933074951, + "num_tokens": 37822336.0, + "step": 1465 + }, + { + "epoch": 0.1609927520316275, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.28778338432312, + "learning_rate": 5.362371888726208e-07, + "loss": 0.9976, + "mean_token_accuracy": 0.69901442527771, + "num_tokens": 37845801.0, + "step": 1466 + }, + { + "epoch": 0.16110256973424117, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3601152896881104, + "learning_rate": 5.366032210834553e-07, + "loss": 1.108, + "mean_token_accuracy": 0.6779302358627319, + "num_tokens": 37872498.0, + "step": 1467 + }, + { + "epoch": 0.16121238743685481, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9594475030899048, + "learning_rate": 5.369692532942899e-07, + "loss": 1.0485, + "mean_token_accuracy": 0.6871182322502136, + "num_tokens": 37904152.0, + "step": 1468 + }, + { + "epoch": 0.1613222051394685, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.122865676879883, + "learning_rate": 5.373352855051244e-07, + "loss": 0.9746, + "mean_token_accuracy": 0.7109463810920715, + "num_tokens": 37931150.0, + "step": 1469 + }, + { + "epoch": 0.16143202284208213, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.38590931892395, + "learning_rate": 5.377013177159589e-07, + "loss": 1.0483, + "mean_token_accuracy": 0.6850886940956116, + "num_tokens": 37955011.0, + "step": 1470 + }, + { + "epoch": 0.1615418405446958, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.378713369369507, + "learning_rate": 5.380673499267935e-07, + "loss": 1.1142, + "mean_token_accuracy": 0.6715977191925049, + "num_tokens": 37981518.0, + "step": 1471 + }, + { + "epoch": 0.16165165824730945, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.247276544570923, + "learning_rate": 5.384333821376281e-07, + "loss": 1.0815, + "mean_token_accuracy": 0.6802945137023926, + "num_tokens": 38008938.0, + "step": 1472 + }, + { + "epoch": 0.16176147594992313, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.307530164718628, + "learning_rate": 5.387994143484626e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6877024173736572, + "num_tokens": 38036626.0, + "step": 1473 + }, + { + "epoch": 0.1618712936525368, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 1.9884661436080933, + "learning_rate": 5.391654465592972e-07, + "loss": 1.0969, + "mean_token_accuracy": 0.6740068197250366, + "num_tokens": 38070141.0, + "step": 1474 + }, + { + "epoch": 0.16198111135515045, + "ewc_loss": 8.046627044677734e-06, + "grad_norm": 2.6588118076324463, + "learning_rate": 5.395314787701317e-07, + "loss": 1.0146, + "mean_token_accuracy": 0.6936167478561401, + "num_tokens": 38092134.0, + "step": 1475 + }, + { + "epoch": 0.16209092905776412, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.6321582794189453, + "learning_rate": 5.398975109809663e-07, + "loss": 0.9785, + "mean_token_accuracy": 0.7008341550827026, + "num_tokens": 38111775.0, + "step": 1476 + }, + { + "epoch": 0.16220074676037777, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.5709896087646484, + "learning_rate": 5.402635431918009e-07, + "loss": 1.0882, + "mean_token_accuracy": 0.6771332025527954, + "num_tokens": 38134366.0, + "step": 1477 + }, + { + "epoch": 0.16231056446299144, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.283543348312378, + "learning_rate": 5.406295754026354e-07, + "loss": 1.0772, + "mean_token_accuracy": 0.6776731014251709, + "num_tokens": 38159731.0, + "step": 1478 + }, + { + "epoch": 0.1624203821656051, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2348792552948, + "learning_rate": 5.4099560761347e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6806329488754272, + "num_tokens": 38189414.0, + "step": 1479 + }, + { + "epoch": 0.16253019986821876, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.3704886436462402, + "learning_rate": 5.413616398243045e-07, + "loss": 1.0231, + "mean_token_accuracy": 0.6914619207382202, + "num_tokens": 38213059.0, + "step": 1480 + }, + { + "epoch": 0.1626400175708324, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.1832125186920166, + "learning_rate": 5.41727672035139e-07, + "loss": 1.1132, + "mean_token_accuracy": 0.6684433221817017, + "num_tokens": 38242367.0, + "step": 1481 + }, + { + "epoch": 0.16274983527344608, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2382869720458984, + "learning_rate": 5.420937042459737e-07, + "loss": 1.1284, + "mean_token_accuracy": 0.6629818081855774, + "num_tokens": 38269270.0, + "step": 1482 + }, + { + "epoch": 0.16285965297605975, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2099666595458984, + "learning_rate": 5.424597364568082e-07, + "loss": 1.0919, + "mean_token_accuracy": 0.6868308782577515, + "num_tokens": 38297171.0, + "step": 1483 + }, + { + "epoch": 0.1629694706786734, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.6235899925231934, + "learning_rate": 5.428257686676427e-07, + "loss": 1.0919, + "mean_token_accuracy": 0.6737807393074036, + "num_tokens": 38319751.0, + "step": 1484 + }, + { + "epoch": 0.16307928838128707, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.4029464721679688, + "learning_rate": 5.431918008784773e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.6922985315322876, + "num_tokens": 38344590.0, + "step": 1485 + }, + { + "epoch": 0.16318910608390072, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.3277053833007812, + "learning_rate": 5.435578330893118e-07, + "loss": 1.0026, + "mean_token_accuracy": 0.7024587988853455, + "num_tokens": 38369984.0, + "step": 1486 + }, + { + "epoch": 0.1632989237865144, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.579857587814331, + "learning_rate": 5.439238653001463e-07, + "loss": 0.9994, + "mean_token_accuracy": 0.7007278203964233, + "num_tokens": 38392266.0, + "step": 1487 + }, + { + "epoch": 0.16340874148912804, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.0112464427948, + "learning_rate": 5.44289897510981e-07, + "loss": 1.1695, + "mean_token_accuracy": 0.6601108908653259, + "num_tokens": 38426150.0, + "step": 1488 + }, + { + "epoch": 0.1635185591917417, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.0168910026550293, + "learning_rate": 5.446559297218155e-07, + "loss": 1.0088, + "mean_token_accuracy": 0.7030531167984009, + "num_tokens": 38458153.0, + "step": 1489 + }, + { + "epoch": 0.16362837689435536, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.625417947769165, + "learning_rate": 5.4502196193265e-07, + "loss": 1.0426, + "mean_token_accuracy": 0.6946697235107422, + "num_tokens": 38478408.0, + "step": 1490 + }, + { + "epoch": 0.16373819459696903, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.1801087856292725, + "learning_rate": 5.453879941434846e-07, + "loss": 0.9613, + "mean_token_accuracy": 0.7047553062438965, + "num_tokens": 38507099.0, + "step": 1491 + }, + { + "epoch": 0.16384801229958268, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.26552414894104, + "learning_rate": 5.457540263543191e-07, + "loss": 1.0318, + "mean_token_accuracy": 0.6930294036865234, + "num_tokens": 38532923.0, + "step": 1492 + }, + { + "epoch": 0.16395783000219635, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.3723127841949463, + "learning_rate": 5.461200585651538e-07, + "loss": 1.1052, + "mean_token_accuracy": 0.6746580600738525, + "num_tokens": 38558137.0, + "step": 1493 + }, + { + "epoch": 0.16406764770481003, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.214472770690918, + "learning_rate": 5.464860907759883e-07, + "loss": 1.045, + "mean_token_accuracy": 0.6886858940124512, + "num_tokens": 38584890.0, + "step": 1494 + }, + { + "epoch": 0.16417746540742367, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2371654510498047, + "learning_rate": 5.468521229868228e-07, + "loss": 1.0486, + "mean_token_accuracy": 0.6870192289352417, + "num_tokens": 38611310.0, + "step": 1495 + }, + { + "epoch": 0.16428728311003735, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2642977237701416, + "learning_rate": 5.472181551976574e-07, + "loss": 1.0966, + "mean_token_accuracy": 0.6782433986663818, + "num_tokens": 38638118.0, + "step": 1496 + }, + { + "epoch": 0.164397100812651, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.4901061058044434, + "learning_rate": 5.475841874084919e-07, + "loss": 0.9597, + "mean_token_accuracy": 0.7090932130813599, + "num_tokens": 38660015.0, + "step": 1497 + }, + { + "epoch": 0.16450691851526467, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.208888292312622, + "learning_rate": 5.479502196193265e-07, + "loss": 1.0697, + "mean_token_accuracy": 0.681589663028717, + "num_tokens": 38688580.0, + "step": 1498 + }, + { + "epoch": 0.1646167362178783, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3733325004577637, + "learning_rate": 5.483162518301611e-07, + "loss": 1.0809, + "mean_token_accuracy": 0.679175853729248, + "num_tokens": 38714193.0, + "step": 1499 + }, + { + "epoch": 0.16472655392049199, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.212435483932495, + "learning_rate": 5.486822840409956e-07, + "loss": 1.0708, + "mean_token_accuracy": 0.680531919002533, + "num_tokens": 38742448.0, + "step": 1500 + }, + { + "epoch": 0.16483637162310563, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.4600861072540283, + "learning_rate": 5.490483162518301e-07, + "loss": 0.9776, + "mean_token_accuracy": 0.7090756297111511, + "num_tokens": 38765620.0, + "step": 1501 + }, + { + "epoch": 0.1649461893257193, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.413814067840576, + "learning_rate": 5.494143484626647e-07, + "loss": 0.952, + "mean_token_accuracy": 0.7128653526306152, + "num_tokens": 38788077.0, + "step": 1502 + }, + { + "epoch": 0.16505600702833298, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.4046578407287598, + "learning_rate": 5.497803806734992e-07, + "loss": 1.0146, + "mean_token_accuracy": 0.6998281478881836, + "num_tokens": 38811822.0, + "step": 1503 + }, + { + "epoch": 0.16516582473094663, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3835113048553467, + "learning_rate": 5.501464128843338e-07, + "loss": 1.0331, + "mean_token_accuracy": 0.6892327070236206, + "num_tokens": 38835722.0, + "step": 1504 + }, + { + "epoch": 0.1652756424335603, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.244455337524414, + "learning_rate": 5.505124450951684e-07, + "loss": 0.9878, + "mean_token_accuracy": 0.7019126415252686, + "num_tokens": 38859889.0, + "step": 1505 + }, + { + "epoch": 0.16538546013617395, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.0792958736419678, + "learning_rate": 5.508784773060029e-07, + "loss": 0.9881, + "mean_token_accuracy": 0.7162895202636719, + "num_tokens": 38890641.0, + "step": 1506 + }, + { + "epoch": 0.16549527783878762, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.359626054763794, + "learning_rate": 5.512445095168374e-07, + "loss": 1.0948, + "mean_token_accuracy": 0.6810157895088196, + "num_tokens": 38914684.0, + "step": 1507 + }, + { + "epoch": 0.16560509554140126, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.5642833709716797, + "learning_rate": 5.51610541727672e-07, + "loss": 0.9972, + "mean_token_accuracy": 0.7034811973571777, + "num_tokens": 38935499.0, + "step": 1508 + }, + { + "epoch": 0.16571491324401494, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3300485610961914, + "learning_rate": 5.519765739385066e-07, + "loss": 1.0362, + "mean_token_accuracy": 0.6872756481170654, + "num_tokens": 38960393.0, + "step": 1509 + }, + { + "epoch": 0.16582473094662858, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.5061821937561035, + "learning_rate": 5.523426061493412e-07, + "loss": 1.0653, + "mean_token_accuracy": 0.6903345584869385, + "num_tokens": 38983330.0, + "step": 1510 + }, + { + "epoch": 0.16593454864924226, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.222370147705078, + "learning_rate": 5.527086383601757e-07, + "loss": 1.1268, + "mean_token_accuracy": 0.6693298816680908, + "num_tokens": 39011686.0, + "step": 1511 + }, + { + "epoch": 0.16604436635185593, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.335415840148926, + "learning_rate": 5.530746705710102e-07, + "loss": 1.0539, + "mean_token_accuracy": 0.6884421110153198, + "num_tokens": 39037527.0, + "step": 1512 + }, + { + "epoch": 0.16615418405446958, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.52679705619812, + "learning_rate": 5.534407027818448e-07, + "loss": 1.0981, + "mean_token_accuracy": 0.673337996006012, + "num_tokens": 39061584.0, + "step": 1513 + }, + { + "epoch": 0.16626400175708325, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.369032621383667, + "learning_rate": 5.538067349926794e-07, + "loss": 1.013, + "mean_token_accuracy": 0.6957389712333679, + "num_tokens": 39085730.0, + "step": 1514 + }, + { + "epoch": 0.1663738194596969, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.8465003967285156, + "learning_rate": 5.541727672035139e-07, + "loss": 1.0251, + "mean_token_accuracy": 0.6990444660186768, + "num_tokens": 39103747.0, + "step": 1515 + }, + { + "epoch": 0.16648363716231057, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.494286298751831, + "learning_rate": 5.545387994143485e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7009382247924805, + "num_tokens": 39124300.0, + "step": 1516 + }, + { + "epoch": 0.16659345486492422, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2573633193969727, + "learning_rate": 5.54904831625183e-07, + "loss": 0.9466, + "mean_token_accuracy": 0.7097805738449097, + "num_tokens": 39148700.0, + "step": 1517 + }, + { + "epoch": 0.1667032725675379, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.599055290222168, + "learning_rate": 5.552708638360175e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.668778121471405, + "num_tokens": 39172508.0, + "step": 1518 + }, + { + "epoch": 0.16681309027015154, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2304232120513916, + "learning_rate": 5.556368960468521e-07, + "loss": 0.9545, + "mean_token_accuracy": 0.7099288702011108, + "num_tokens": 39197782.0, + "step": 1519 + }, + { + "epoch": 0.1669229079727652, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.234818458557129, + "learning_rate": 5.560029282576867e-07, + "loss": 1.0051, + "mean_token_accuracy": 0.6975874900817871, + "num_tokens": 39224616.0, + "step": 1520 + }, + { + "epoch": 0.16703272567537888, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3020684719085693, + "learning_rate": 5.563689604685212e-07, + "loss": 0.9942, + "mean_token_accuracy": 0.6998487710952759, + "num_tokens": 39249771.0, + "step": 1521 + }, + { + "epoch": 0.16714254337799253, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.416027545928955, + "learning_rate": 5.567349926793558e-07, + "loss": 0.9246, + "mean_token_accuracy": 0.7149417996406555, + "num_tokens": 39271524.0, + "step": 1522 + }, + { + "epoch": 0.1672523610806062, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.194518804550171, + "learning_rate": 5.571010248901903e-07, + "loss": 0.9947, + "mean_token_accuracy": 0.6999338865280151, + "num_tokens": 39299086.0, + "step": 1523 + }, + { + "epoch": 0.16736217878321985, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.0298314094543457, + "learning_rate": 5.574670571010248e-07, + "loss": 1.0064, + "mean_token_accuracy": 0.704623818397522, + "num_tokens": 39331423.0, + "step": 1524 + }, + { + "epoch": 0.16747199648583352, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.5450351238250732, + "learning_rate": 5.578330893118595e-07, + "loss": 1.0609, + "mean_token_accuracy": 0.6823590993881226, + "num_tokens": 39352747.0, + "step": 1525 + }, + { + "epoch": 0.16758181418844717, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3966896533966064, + "learning_rate": 5.58199121522694e-07, + "loss": 1.0745, + "mean_token_accuracy": 0.6829248070716858, + "num_tokens": 39377552.0, + "step": 1526 + }, + { + "epoch": 0.16769163189106084, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.451094627380371, + "learning_rate": 5.585651537335286e-07, + "loss": 1.0498, + "mean_token_accuracy": 0.6849659085273743, + "num_tokens": 39401174.0, + "step": 1527 + }, + { + "epoch": 0.1678014495936745, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2042975425720215, + "learning_rate": 5.589311859443631e-07, + "loss": 0.9527, + "mean_token_accuracy": 0.7113037109375, + "num_tokens": 39427179.0, + "step": 1528 + }, + { + "epoch": 0.16791126729628816, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.536472797393799, + "learning_rate": 5.592972181551976e-07, + "loss": 0.9461, + "mean_token_accuracy": 0.7079958915710449, + "num_tokens": 39446853.0, + "step": 1529 + }, + { + "epoch": 0.1680210849989018, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.5058159828186035, + "learning_rate": 5.596632503660323e-07, + "loss": 0.9574, + "mean_token_accuracy": 0.7121334075927734, + "num_tokens": 39467826.0, + "step": 1530 + }, + { + "epoch": 0.16813090270151548, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.2856411933898926, + "learning_rate": 5.600292825768668e-07, + "loss": 0.9155, + "mean_token_accuracy": 0.7186814546585083, + "num_tokens": 39492608.0, + "step": 1531 + }, + { + "epoch": 0.16824072040412916, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.5592780113220215, + "learning_rate": 5.603953147877013e-07, + "loss": 1.0586, + "mean_token_accuracy": 0.6901339292526245, + "num_tokens": 39513984.0, + "step": 1532 + }, + { + "epoch": 0.1683505381067428, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.1719090938568115, + "learning_rate": 5.607613469985359e-07, + "loss": 0.9769, + "mean_token_accuracy": 0.7036182880401611, + "num_tokens": 39541191.0, + "step": 1533 + }, + { + "epoch": 0.16846035580935648, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3493101596832275, + "learning_rate": 5.611273792093704e-07, + "loss": 0.9854, + "mean_token_accuracy": 0.715655505657196, + "num_tokens": 39565937.0, + "step": 1534 + }, + { + "epoch": 0.16857017351197012, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.6814777851104736, + "learning_rate": 5.614934114202049e-07, + "loss": 1.0269, + "mean_token_accuracy": 0.69720858335495, + "num_tokens": 39585586.0, + "step": 1535 + }, + { + "epoch": 0.1686799912145838, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 1.9945536851882935, + "learning_rate": 5.618594436310396e-07, + "loss": 1.1408, + "mean_token_accuracy": 0.6652577519416809, + "num_tokens": 39621116.0, + "step": 1536 + }, + { + "epoch": 0.16878980891719744, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.4969213008880615, + "learning_rate": 5.622254758418741e-07, + "loss": 1.0734, + "mean_token_accuracy": 0.6838943958282471, + "num_tokens": 39645545.0, + "step": 1537 + }, + { + "epoch": 0.16889962661981112, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.296663522720337, + "learning_rate": 5.625915080527086e-07, + "loss": 1.0716, + "mean_token_accuracy": 0.6877169609069824, + "num_tokens": 39671881.0, + "step": 1538 + }, + { + "epoch": 0.16900944432242476, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.420781135559082, + "learning_rate": 5.629575402635432e-07, + "loss": 1.0157, + "mean_token_accuracy": 0.6950898170471191, + "num_tokens": 39695290.0, + "step": 1539 + }, + { + "epoch": 0.16911926202503844, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.222778081893921, + "learning_rate": 5.633235724743777e-07, + "loss": 1.0214, + "mean_token_accuracy": 0.7016769647598267, + "num_tokens": 39722135.0, + "step": 1540 + }, + { + "epoch": 0.1692290797276521, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.278472423553467, + "learning_rate": 5.636896046852124e-07, + "loss": 1.1505, + "mean_token_accuracy": 0.6616908311843872, + "num_tokens": 39752087.0, + "step": 1541 + }, + { + "epoch": 0.16933889743026576, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.116119384765625, + "learning_rate": 5.640556368960469e-07, + "loss": 1.0264, + "mean_token_accuracy": 0.6936984062194824, + "num_tokens": 39782108.0, + "step": 1542 + }, + { + "epoch": 0.16944871513287943, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2285070419311523, + "learning_rate": 5.644216691068814e-07, + "loss": 1.0736, + "mean_token_accuracy": 0.6754435300827026, + "num_tokens": 39810707.0, + "step": 1543 + }, + { + "epoch": 0.16955853283549308, + "ewc_loss": 8.165836334228516e-06, + "grad_norm": 2.3383612632751465, + "learning_rate": 5.64787701317716e-07, + "loss": 0.9891, + "mean_token_accuracy": 0.7005850076675415, + "num_tokens": 39836838.0, + "step": 1544 + }, + { + "epoch": 0.16966835053810675, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4445555210113525, + "learning_rate": 5.651537335285505e-07, + "loss": 0.9688, + "mean_token_accuracy": 0.7109750509262085, + "num_tokens": 39860798.0, + "step": 1545 + }, + { + "epoch": 0.1697781682407204, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3430187702178955, + "learning_rate": 5.655197657393851e-07, + "loss": 1.1204, + "mean_token_accuracy": 0.6719301342964172, + "num_tokens": 39886558.0, + "step": 1546 + }, + { + "epoch": 0.16988798594333407, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.58811354637146, + "learning_rate": 5.658857979502197e-07, + "loss": 0.9477, + "mean_token_accuracy": 0.7074747681617737, + "num_tokens": 39907445.0, + "step": 1547 + }, + { + "epoch": 0.16999780364594771, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1177480220794678, + "learning_rate": 5.662518301610542e-07, + "loss": 0.9465, + "mean_token_accuracy": 0.7180776000022888, + "num_tokens": 39937347.0, + "step": 1548 + }, + { + "epoch": 0.1701076213485614, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.168470859527588, + "learning_rate": 5.666178623718887e-07, + "loss": 0.9847, + "mean_token_accuracy": 0.705476701259613, + "num_tokens": 39966512.0, + "step": 1549 + }, + { + "epoch": 0.17021743905117506, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3947789669036865, + "learning_rate": 5.669838945827233e-07, + "loss": 1.0689, + "mean_token_accuracy": 0.6815032958984375, + "num_tokens": 39993941.0, + "step": 1550 + }, + { + "epoch": 0.1703272567537887, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.997756004333496, + "learning_rate": 5.673499267935577e-07, + "loss": 0.8892, + "mean_token_accuracy": 0.7309010028839111, + "num_tokens": 40012552.0, + "step": 1551 + }, + { + "epoch": 0.17043707445640238, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.154972791671753, + "learning_rate": 5.677159590043924e-07, + "loss": 1.0591, + "mean_token_accuracy": 0.686836838722229, + "num_tokens": 40043398.0, + "step": 1552 + }, + { + "epoch": 0.17054689215901603, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4113733768463135, + "learning_rate": 5.68081991215227e-07, + "loss": 1.0929, + "mean_token_accuracy": 0.6781075596809387, + "num_tokens": 40070386.0, + "step": 1553 + }, + { + "epoch": 0.1706567098616297, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3946430683135986, + "learning_rate": 5.684480234260615e-07, + "loss": 1.022, + "mean_token_accuracy": 0.6950194239616394, + "num_tokens": 40094339.0, + "step": 1554 + }, + { + "epoch": 0.17076652756424335, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.285576581954956, + "learning_rate": 5.688140556368959e-07, + "loss": 0.9354, + "mean_token_accuracy": 0.7211384773254395, + "num_tokens": 40118938.0, + "step": 1555 + }, + { + "epoch": 0.17087634526685702, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.212008476257324, + "learning_rate": 5.691800878477305e-07, + "loss": 1.1116, + "mean_token_accuracy": 0.6675478219985962, + "num_tokens": 40148530.0, + "step": 1556 + }, + { + "epoch": 0.17098616296947067, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2573564052581787, + "learning_rate": 5.695461200585652e-07, + "loss": 1.0669, + "mean_token_accuracy": 0.6876928806304932, + "num_tokens": 40175048.0, + "step": 1557 + }, + { + "epoch": 0.17109598067208434, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.416945457458496, + "learning_rate": 5.699121522693998e-07, + "loss": 1.0236, + "mean_token_accuracy": 0.6940038204193115, + "num_tokens": 40198815.0, + "step": 1558 + }, + { + "epoch": 0.17120579837469801, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.428257465362549, + "learning_rate": 5.702781844802342e-07, + "loss": 0.9956, + "mean_token_accuracy": 0.7072563767433167, + "num_tokens": 40221652.0, + "step": 1559 + }, + { + "epoch": 0.17131561607731166, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2832348346710205, + "learning_rate": 5.706442166910687e-07, + "loss": 1.0121, + "mean_token_accuracy": 0.6967260837554932, + "num_tokens": 40248573.0, + "step": 1560 + }, + { + "epoch": 0.17142543377992533, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.451840400695801, + "learning_rate": 5.710102489019033e-07, + "loss": 1.0855, + "mean_token_accuracy": 0.6801333427429199, + "num_tokens": 40273242.0, + "step": 1561 + }, + { + "epoch": 0.17153525148253898, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4193289279937744, + "learning_rate": 5.71376281112738e-07, + "loss": 1.0108, + "mean_token_accuracy": 0.696841835975647, + "num_tokens": 40295379.0, + "step": 1562 + }, + { + "epoch": 0.17164506918515265, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.8185088634490967, + "learning_rate": 5.717423133235724e-07, + "loss": 1.0149, + "mean_token_accuracy": 0.6942393779754639, + "num_tokens": 40313918.0, + "step": 1563 + }, + { + "epoch": 0.1717548868877663, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3638014793395996, + "learning_rate": 5.72108345534407e-07, + "loss": 1.0327, + "mean_token_accuracy": 0.6933850646018982, + "num_tokens": 40338811.0, + "step": 1564 + }, + { + "epoch": 0.17186470459037997, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4184226989746094, + "learning_rate": 5.724743777452415e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.6898305416107178, + "num_tokens": 40361823.0, + "step": 1565 + }, + { + "epoch": 0.17197452229299362, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.356969118118286, + "learning_rate": 5.72840409956076e-07, + "loss": 1.0726, + "mean_token_accuracy": 0.6897419691085815, + "num_tokens": 40386526.0, + "step": 1566 + }, + { + "epoch": 0.1720843399956073, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2562289237976074, + "learning_rate": 5.732064421669106e-07, + "loss": 0.9986, + "mean_token_accuracy": 0.7013916969299316, + "num_tokens": 40412670.0, + "step": 1567 + }, + { + "epoch": 0.17219415769822094, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2846713066101074, + "learning_rate": 5.735724743777452e-07, + "loss": 1.0186, + "mean_token_accuracy": 0.6962847709655762, + "num_tokens": 40439359.0, + "step": 1568 + }, + { + "epoch": 0.1723039754008346, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1891028881073, + "learning_rate": 5.739385065885797e-07, + "loss": 1.0133, + "mean_token_accuracy": 0.6997495889663696, + "num_tokens": 40466494.0, + "step": 1569 + }, + { + "epoch": 0.1724137931034483, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5059115886688232, + "learning_rate": 5.743045387994143e-07, + "loss": 0.9718, + "mean_token_accuracy": 0.7045832872390747, + "num_tokens": 40488412.0, + "step": 1570 + }, + { + "epoch": 0.17252361080606193, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3558826446533203, + "learning_rate": 5.746705710102488e-07, + "loss": 1.0659, + "mean_token_accuracy": 0.6828343868255615, + "num_tokens": 40511460.0, + "step": 1571 + }, + { + "epoch": 0.1726334285086756, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2255918979644775, + "learning_rate": 5.750366032210833e-07, + "loss": 0.9843, + "mean_token_accuracy": 0.7060045003890991, + "num_tokens": 40537397.0, + "step": 1572 + }, + { + "epoch": 0.17274324621128925, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.692420721054077, + "learning_rate": 5.75402635431918e-07, + "loss": 0.9928, + "mean_token_accuracy": 0.7010341882705688, + "num_tokens": 40556838.0, + "step": 1573 + }, + { + "epoch": 0.17285306391390293, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.7982571125030518, + "learning_rate": 5.757686676427525e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.6953666806221008, + "num_tokens": 40575543.0, + "step": 1574 + }, + { + "epoch": 0.17296288161651657, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1649363040924072, + "learning_rate": 5.761346998535871e-07, + "loss": 1.057, + "mean_token_accuracy": 0.6881320476531982, + "num_tokens": 40604152.0, + "step": 1575 + }, + { + "epoch": 0.17307269931913025, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3799519538879395, + "learning_rate": 5.765007320644216e-07, + "loss": 1.0508, + "mean_token_accuracy": 0.686424970626831, + "num_tokens": 40627574.0, + "step": 1576 + }, + { + "epoch": 0.1731825170217439, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2455530166625977, + "learning_rate": 5.768667642752561e-07, + "loss": 0.991, + "mean_token_accuracy": 0.6998533606529236, + "num_tokens": 40651057.0, + "step": 1577 + }, + { + "epoch": 0.17329233472435757, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.218794107437134, + "learning_rate": 5.772327964860908e-07, + "loss": 1.0262, + "mean_token_accuracy": 0.6889650821685791, + "num_tokens": 40678662.0, + "step": 1578 + }, + { + "epoch": 0.17340215242697124, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.332672595977783, + "learning_rate": 5.775988286969253e-07, + "loss": 1.0895, + "mean_token_accuracy": 0.6839778423309326, + "num_tokens": 40704905.0, + "step": 1579 + }, + { + "epoch": 0.17351197012958489, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5859909057617188, + "learning_rate": 5.779648609077598e-07, + "loss": 0.9537, + "mean_token_accuracy": 0.7119323015213013, + "num_tokens": 40725916.0, + "step": 1580 + }, + { + "epoch": 0.17362178783219856, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.588134288787842, + "learning_rate": 5.783308931185944e-07, + "loss": 1.0519, + "mean_token_accuracy": 0.6848116517066956, + "num_tokens": 40746748.0, + "step": 1581 + }, + { + "epoch": 0.1737316055348122, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.169506072998047, + "learning_rate": 5.786969253294289e-07, + "loss": 1.1233, + "mean_token_accuracy": 0.6849774122238159, + "num_tokens": 40779091.0, + "step": 1582 + }, + { + "epoch": 0.17384142323742588, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0615572929382324, + "learning_rate": 5.790629575402634e-07, + "loss": 1.102, + "mean_token_accuracy": 0.6703960299491882, + "num_tokens": 40809761.0, + "step": 1583 + }, + { + "epoch": 0.17395124094003953, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.123969793319702, + "learning_rate": 5.794289897510981e-07, + "loss": 1.0678, + "mean_token_accuracy": 0.6866147518157959, + "num_tokens": 40841486.0, + "step": 1584 + }, + { + "epoch": 0.1740610586426532, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.116396427154541, + "learning_rate": 5.797950219619326e-07, + "loss": 1.0836, + "mean_token_accuracy": 0.6828042268753052, + "num_tokens": 40870417.0, + "step": 1585 + }, + { + "epoch": 0.17417087634526685, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1913440227508545, + "learning_rate": 5.801610541727671e-07, + "loss": 1.0121, + "mean_token_accuracy": 0.7021902799606323, + "num_tokens": 40898422.0, + "step": 1586 + }, + { + "epoch": 0.17428069404788052, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.4746973514556885, + "learning_rate": 5.805270863836017e-07, + "loss": 1.1264, + "mean_token_accuracy": 0.6695870161056519, + "num_tokens": 40921246.0, + "step": 1587 + }, + { + "epoch": 0.1743905117504942, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2080037593841553, + "learning_rate": 5.808931185944362e-07, + "loss": 1.0145, + "mean_token_accuracy": 0.6897063255310059, + "num_tokens": 40947697.0, + "step": 1588 + }, + { + "epoch": 0.17450032945310784, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2866368293762207, + "learning_rate": 5.812591508052709e-07, + "loss": 0.9403, + "mean_token_accuracy": 0.7163083553314209, + "num_tokens": 40971954.0, + "step": 1589 + }, + { + "epoch": 0.1746101471557215, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1185595989227295, + "learning_rate": 5.816251830161054e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6810484528541565, + "num_tokens": 41003026.0, + "step": 1590 + }, + { + "epoch": 0.17471996485833516, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3186774253845215, + "learning_rate": 5.819912152269399e-07, + "loss": 1.064, + "mean_token_accuracy": 0.6800464391708374, + "num_tokens": 41027889.0, + "step": 1591 + }, + { + "epoch": 0.17482978256094883, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.687417507171631, + "learning_rate": 5.823572474377745e-07, + "loss": 1.0334, + "mean_token_accuracy": 0.689611554145813, + "num_tokens": 41048983.0, + "step": 1592 + }, + { + "epoch": 0.17493960026356248, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3353374004364014, + "learning_rate": 5.82723279648609e-07, + "loss": 1.0851, + "mean_token_accuracy": 0.6730213165283203, + "num_tokens": 41074025.0, + "step": 1593 + }, + { + "epoch": 0.17504941796617615, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1859242916107178, + "learning_rate": 5.830893118594436e-07, + "loss": 1.0842, + "mean_token_accuracy": 0.6767100095748901, + "num_tokens": 41104142.0, + "step": 1594 + }, + { + "epoch": 0.1751592356687898, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3237946033477783, + "learning_rate": 5.834553440702782e-07, + "loss": 1.0502, + "mean_token_accuracy": 0.6879336833953857, + "num_tokens": 41128610.0, + "step": 1595 + }, + { + "epoch": 0.17526905337140347, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.201077938079834, + "learning_rate": 5.838213762811127e-07, + "loss": 0.9905, + "mean_token_accuracy": 0.7025249600410461, + "num_tokens": 41154215.0, + "step": 1596 + }, + { + "epoch": 0.17537887107401715, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.1190366744995117, + "learning_rate": 5.841874084919472e-07, + "loss": 1.081, + "mean_token_accuracy": 0.6744168996810913, + "num_tokens": 41186683.0, + "step": 1597 + }, + { + "epoch": 0.1754886887766308, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.3097760677337646, + "learning_rate": 5.845534407027818e-07, + "loss": 0.9997, + "mean_token_accuracy": 0.6944568157196045, + "num_tokens": 41211138.0, + "step": 1598 + }, + { + "epoch": 0.17559850647924446, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.7930355072021484, + "learning_rate": 5.849194729136163e-07, + "loss": 0.9275, + "mean_token_accuracy": 0.7173671722412109, + "num_tokens": 41229133.0, + "step": 1599 + }, + { + "epoch": 0.1757083241818581, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.4481942653656006, + "learning_rate": 5.852855051244509e-07, + "loss": 1.0607, + "mean_token_accuracy": 0.6881853342056274, + "num_tokens": 41254358.0, + "step": 1600 + }, + { + "epoch": 0.17581814188447178, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.205556631088257, + "learning_rate": 5.856515373352855e-07, + "loss": 1.0876, + "mean_token_accuracy": 0.6852130889892578, + "num_tokens": 41282152.0, + "step": 1601 + }, + { + "epoch": 0.17592795958708543, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.3575215339660645, + "learning_rate": 5.8601756954612e-07, + "loss": 0.963, + "mean_token_accuracy": 0.7141952514648438, + "num_tokens": 41307760.0, + "step": 1602 + }, + { + "epoch": 0.1760377772896991, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.255261182785034, + "learning_rate": 5.863836017569545e-07, + "loss": 0.9824, + "mean_token_accuracy": 0.7035854458808899, + "num_tokens": 41333522.0, + "step": 1603 + }, + { + "epoch": 0.17614759499231275, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.3312647342681885, + "learning_rate": 5.867496339677891e-07, + "loss": 1.0183, + "mean_token_accuracy": 0.6978260278701782, + "num_tokens": 41357614.0, + "step": 1604 + }, + { + "epoch": 0.17625741269492642, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.2335474491119385, + "learning_rate": 5.871156661786237e-07, + "loss": 0.971, + "mean_token_accuracy": 0.7098402380943298, + "num_tokens": 41383591.0, + "step": 1605 + }, + { + "epoch": 0.17636723039754007, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.317614793777466, + "learning_rate": 5.874816983894583e-07, + "loss": 1.0974, + "mean_token_accuracy": 0.6729731559753418, + "num_tokens": 41410821.0, + "step": 1606 + }, + { + "epoch": 0.17647704810015374, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.5371429920196533, + "learning_rate": 5.878477306002928e-07, + "loss": 1.0694, + "mean_token_accuracy": 0.6854631900787354, + "num_tokens": 41433212.0, + "step": 1607 + }, + { + "epoch": 0.17658686580276742, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.215691566467285, + "learning_rate": 5.882137628111273e-07, + "loss": 1.1445, + "mean_token_accuracy": 0.6595195531845093, + "num_tokens": 41461618.0, + "step": 1608 + }, + { + "epoch": 0.17669668350538106, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.1598544120788574, + "learning_rate": 5.885797950219619e-07, + "loss": 1.0561, + "mean_token_accuracy": 0.6816978454589844, + "num_tokens": 41490369.0, + "step": 1609 + }, + { + "epoch": 0.17680650120799474, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.283024787902832, + "learning_rate": 5.889458272327965e-07, + "loss": 1.2102, + "mean_token_accuracy": 0.6476158499717712, + "num_tokens": 41518493.0, + "step": 1610 + }, + { + "epoch": 0.17691631891060838, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.261524200439453, + "learning_rate": 5.89311859443631e-07, + "loss": 1.1197, + "mean_token_accuracy": 0.6743656992912292, + "num_tokens": 41545826.0, + "step": 1611 + }, + { + "epoch": 0.17702613661322206, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.4692752361297607, + "learning_rate": 5.896778916544656e-07, + "loss": 1.0141, + "mean_token_accuracy": 0.6909664869308472, + "num_tokens": 41568877.0, + "step": 1612 + }, + { + "epoch": 0.1771359543158357, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.346996784210205, + "learning_rate": 5.900439238653001e-07, + "loss": 0.9829, + "mean_token_accuracy": 0.7048208713531494, + "num_tokens": 41594228.0, + "step": 1613 + }, + { + "epoch": 0.17724577201844938, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.3599956035614014, + "learning_rate": 5.904099560761346e-07, + "loss": 1.0906, + "mean_token_accuracy": 0.6790587902069092, + "num_tokens": 41620861.0, + "step": 1614 + }, + { + "epoch": 0.17735558972106302, + "ewc_loss": 8.761882781982422e-06, + "grad_norm": 2.3072001934051514, + "learning_rate": 5.907759882869692e-07, + "loss": 1.0897, + "mean_token_accuracy": 0.6816046833992004, + "num_tokens": 41646859.0, + "step": 1615 + }, + { + "epoch": 0.1774654074236767, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.436711072921753, + "learning_rate": 5.911420204978038e-07, + "loss": 1.0509, + "mean_token_accuracy": 0.6880630850791931, + "num_tokens": 41670100.0, + "step": 1616 + }, + { + "epoch": 0.17757522512629037, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.543444871902466, + "learning_rate": 5.915080527086383e-07, + "loss": 0.9905, + "mean_token_accuracy": 0.7068248987197876, + "num_tokens": 41690417.0, + "step": 1617 + }, + { + "epoch": 0.17768504282890402, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 1.976969838142395, + "learning_rate": 5.918740849194729e-07, + "loss": 0.9558, + "mean_token_accuracy": 0.7206727862358093, + "num_tokens": 41723702.0, + "step": 1618 + }, + { + "epoch": 0.1777948605315177, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.304720163345337, + "learning_rate": 5.922401171303074e-07, + "loss": 0.9721, + "mean_token_accuracy": 0.7036530375480652, + "num_tokens": 41747292.0, + "step": 1619 + }, + { + "epoch": 0.17790467823413134, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.5636661052703857, + "learning_rate": 5.926061493411419e-07, + "loss": 1.0755, + "mean_token_accuracy": 0.6807336211204529, + "num_tokens": 41771664.0, + "step": 1620 + }, + { + "epoch": 0.178014495936745, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.1138854026794434, + "learning_rate": 5.929721815519766e-07, + "loss": 1.0847, + "mean_token_accuracy": 0.6782404184341431, + "num_tokens": 41802272.0, + "step": 1621 + }, + { + "epoch": 0.17812431363935866, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.265409231185913, + "learning_rate": 5.933382137628111e-07, + "loss": 1.0261, + "mean_token_accuracy": 0.6894536018371582, + "num_tokens": 41826323.0, + "step": 1622 + }, + { + "epoch": 0.17823413134197233, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.047386646270752, + "learning_rate": 5.937042459736457e-07, + "loss": 1.1349, + "mean_token_accuracy": 0.6661732792854309, + "num_tokens": 41859043.0, + "step": 1623 + }, + { + "epoch": 0.17834394904458598, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.491070508956909, + "learning_rate": 5.940702781844802e-07, + "loss": 0.9993, + "mean_token_accuracy": 0.6999269127845764, + "num_tokens": 41880055.0, + "step": 1624 + }, + { + "epoch": 0.17845376674719965, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.291685104370117, + "learning_rate": 5.944363103953147e-07, + "loss": 1.0657, + "mean_token_accuracy": 0.6776823997497559, + "num_tokens": 41906987.0, + "step": 1625 + }, + { + "epoch": 0.17856358444981332, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.3933398723602295, + "learning_rate": 5.948023426061494e-07, + "loss": 1.0403, + "mean_token_accuracy": 0.6996802091598511, + "num_tokens": 41929196.0, + "step": 1626 + }, + { + "epoch": 0.17867340215242697, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.331112861633301, + "learning_rate": 5.951683748169839e-07, + "loss": 1.1477, + "mean_token_accuracy": 0.6660517454147339, + "num_tokens": 41955354.0, + "step": 1627 + }, + { + "epoch": 0.17878321985504064, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.258838415145874, + "learning_rate": 5.955344070278184e-07, + "loss": 1.1463, + "mean_token_accuracy": 0.6583786606788635, + "num_tokens": 41982518.0, + "step": 1628 + }, + { + "epoch": 0.1788930375576543, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.314000368118286, + "learning_rate": 5.95900439238653e-07, + "loss": 0.9976, + "mean_token_accuracy": 0.6988988518714905, + "num_tokens": 42006184.0, + "step": 1629 + }, + { + "epoch": 0.17900285526026796, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.750415325164795, + "learning_rate": 5.962664714494875e-07, + "loss": 0.9924, + "mean_token_accuracy": 0.700863242149353, + "num_tokens": 42025392.0, + "step": 1630 + }, + { + "epoch": 0.1791126729628816, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.1357855796813965, + "learning_rate": 5.96632503660322e-07, + "loss": 1.1746, + "mean_token_accuracy": 0.6613538265228271, + "num_tokens": 42055514.0, + "step": 1631 + }, + { + "epoch": 0.17922249066549528, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.355719566345215, + "learning_rate": 5.969985358711567e-07, + "loss": 1.071, + "mean_token_accuracy": 0.6819658875465393, + "num_tokens": 42081033.0, + "step": 1632 + }, + { + "epoch": 0.17933230836810893, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3807175159454346, + "learning_rate": 5.973645680819912e-07, + "loss": 0.976, + "mean_token_accuracy": 0.7108026742935181, + "num_tokens": 42104399.0, + "step": 1633 + }, + { + "epoch": 0.1794421260707226, + "ewc_loss": 8.881092071533203e-06, + "grad_norm": 2.3485875129699707, + "learning_rate": 5.977306002928257e-07, + "loss": 1.0543, + "mean_token_accuracy": 0.6832147836685181, + "num_tokens": 42128912.0, + "step": 1634 + }, + { + "epoch": 0.17955194377333628, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.71380877494812, + "learning_rate": 5.980966325036603e-07, + "loss": 1.0908, + "mean_token_accuracy": 0.6758580803871155, + "num_tokens": 42150892.0, + "step": 1635 + }, + { + "epoch": 0.17966176147594992, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.478759288787842, + "learning_rate": 5.984626647144948e-07, + "loss": 0.9754, + "mean_token_accuracy": 0.7118122577667236, + "num_tokens": 42172510.0, + "step": 1636 + }, + { + "epoch": 0.1797715791785636, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2232234477996826, + "learning_rate": 5.988286969253294e-07, + "loss": 1.0518, + "mean_token_accuracy": 0.6845182180404663, + "num_tokens": 42199310.0, + "step": 1637 + }, + { + "epoch": 0.17988139688117724, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.131078004837036, + "learning_rate": 5.99194729136164e-07, + "loss": 1.0864, + "mean_token_accuracy": 0.674201488494873, + "num_tokens": 42229990.0, + "step": 1638 + }, + { + "epoch": 0.17999121458379091, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.561554431915283, + "learning_rate": 5.995607613469985e-07, + "loss": 0.9584, + "mean_token_accuracy": 0.712734043598175, + "num_tokens": 42249858.0, + "step": 1639 + }, + { + "epoch": 0.18010103228640456, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.631131887435913, + "learning_rate": 5.999267935578331e-07, + "loss": 0.9533, + "mean_token_accuracy": 0.7086496353149414, + "num_tokens": 42271355.0, + "step": 1640 + }, + { + "epoch": 0.18021084998901823, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2739181518554688, + "learning_rate": 6.002928257686676e-07, + "loss": 1.023, + "mean_token_accuracy": 0.6980669498443604, + "num_tokens": 42297624.0, + "step": 1641 + }, + { + "epoch": 0.18032066769163188, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2938220500946045, + "learning_rate": 6.006588579795022e-07, + "loss": 0.9629, + "mean_token_accuracy": 0.7069181203842163, + "num_tokens": 42322395.0, + "step": 1642 + }, + { + "epoch": 0.18043048539424555, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.5537397861480713, + "learning_rate": 6.010248901903368e-07, + "loss": 1.0163, + "mean_token_accuracy": 0.7095682621002197, + "num_tokens": 42344634.0, + "step": 1643 + }, + { + "epoch": 0.1805403030968592, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.369662046432495, + "learning_rate": 6.013909224011713e-07, + "loss": 1.0981, + "mean_token_accuracy": 0.6778258085250854, + "num_tokens": 42370501.0, + "step": 1644 + }, + { + "epoch": 0.18065012079947287, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.240870952606201, + "learning_rate": 6.017569546120058e-07, + "loss": 1.0468, + "mean_token_accuracy": 0.6913483142852783, + "num_tokens": 42396966.0, + "step": 1645 + }, + { + "epoch": 0.18075993850208655, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.287991762161255, + "learning_rate": 6.021229868228404e-07, + "loss": 1.0943, + "mean_token_accuracy": 0.6857826113700867, + "num_tokens": 42423033.0, + "step": 1646 + }, + { + "epoch": 0.1808697562047002, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 1.9665578603744507, + "learning_rate": 6.024890190336749e-07, + "loss": 1.0425, + "mean_token_accuracy": 0.694553017616272, + "num_tokens": 42456449.0, + "step": 1647 + }, + { + "epoch": 0.18097957390731387, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.3921613693237305, + "learning_rate": 6.028550512445095e-07, + "loss": 1.0078, + "mean_token_accuracy": 0.6994770765304565, + "num_tokens": 42480035.0, + "step": 1648 + }, + { + "epoch": 0.1810893916099275, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.376394033432007, + "learning_rate": 6.032210834553441e-07, + "loss": 1.017, + "mean_token_accuracy": 0.7008526921272278, + "num_tokens": 42503308.0, + "step": 1649 + }, + { + "epoch": 0.1811992093125412, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.274341106414795, + "learning_rate": 6.035871156661786e-07, + "loss": 1.0831, + "mean_token_accuracy": 0.6786530017852783, + "num_tokens": 42532075.0, + "step": 1650 + }, + { + "epoch": 0.18130902701515483, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3341896533966064, + "learning_rate": 6.039531478770131e-07, + "loss": 0.9943, + "mean_token_accuracy": 0.7038760185241699, + "num_tokens": 42556184.0, + "step": 1651 + }, + { + "epoch": 0.1814188447177685, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5536108016967773, + "learning_rate": 6.043191800878477e-07, + "loss": 1.0947, + "mean_token_accuracy": 0.6799367666244507, + "num_tokens": 42578435.0, + "step": 1652 + }, + { + "epoch": 0.18152866242038215, + "ewc_loss": 9.179115295410156e-06, + "grad_norm": 2.369713306427002, + "learning_rate": 6.046852122986823e-07, + "loss": 1.0493, + "mean_token_accuracy": 0.6869180202484131, + "num_tokens": 42604295.0, + "step": 1653 + }, + { + "epoch": 0.18163848012299583, + "ewc_loss": 9.179115295410156e-06, + "grad_norm": 2.262498140335083, + "learning_rate": 6.050512445095168e-07, + "loss": 1.1209, + "mean_token_accuracy": 0.6802578568458557, + "num_tokens": 42632476.0, + "step": 1654 + }, + { + "epoch": 0.1817482978256095, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.4399380683898926, + "learning_rate": 6.054172767203514e-07, + "loss": 1.0991, + "mean_token_accuracy": 0.6750161647796631, + "num_tokens": 42657921.0, + "step": 1655 + }, + { + "epoch": 0.18185811552822315, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.4688117504119873, + "learning_rate": 6.057833089311859e-07, + "loss": 1.1409, + "mean_token_accuracy": 0.6646488904953003, + "num_tokens": 42683798.0, + "step": 1656 + }, + { + "epoch": 0.18196793323083682, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.377020835876465, + "learning_rate": 6.061493411420205e-07, + "loss": 0.9941, + "mean_token_accuracy": 0.7013235688209534, + "num_tokens": 42707333.0, + "step": 1657 + }, + { + "epoch": 0.18207775093345047, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2473039627075195, + "learning_rate": 6.065153733528551e-07, + "loss": 1.0455, + "mean_token_accuracy": 0.6936679482460022, + "num_tokens": 42733776.0, + "step": 1658 + }, + { + "epoch": 0.18218756863606414, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9340274333953857, + "learning_rate": 6.068814055636896e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6776952743530273, + "num_tokens": 42769503.0, + "step": 1659 + }, + { + "epoch": 0.18229738633867779, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.5617241859436035, + "learning_rate": 6.072474377745242e-07, + "loss": 1.0046, + "mean_token_accuracy": 0.7048865556716919, + "num_tokens": 42791718.0, + "step": 1660 + }, + { + "epoch": 0.18240720404129146, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.3465991020202637, + "learning_rate": 6.076134699853587e-07, + "loss": 1.0305, + "mean_token_accuracy": 0.6919146776199341, + "num_tokens": 42816722.0, + "step": 1661 + }, + { + "epoch": 0.1825170217439051, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1183061599731445, + "learning_rate": 6.079795021961932e-07, + "loss": 0.9815, + "mean_token_accuracy": 0.7002652287483215, + "num_tokens": 42844961.0, + "step": 1662 + }, + { + "epoch": 0.18262683944651878, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.193450450897217, + "learning_rate": 6.083455344070278e-07, + "loss": 0.9392, + "mean_token_accuracy": 0.71014803647995, + "num_tokens": 42870511.0, + "step": 1663 + }, + { + "epoch": 0.18273665714913245, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1982009410858154, + "learning_rate": 6.087115666178624e-07, + "loss": 0.9664, + "mean_token_accuracy": 0.707743763923645, + "num_tokens": 42897649.0, + "step": 1664 + }, + { + "epoch": 0.1828464748517461, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.084569215774536, + "learning_rate": 6.090775988286969e-07, + "loss": 0.9766, + "mean_token_accuracy": 0.7045094966888428, + "num_tokens": 42926306.0, + "step": 1665 + }, + { + "epoch": 0.18295629255435977, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.2745285034179688, + "learning_rate": 6.094436310395315e-07, + "loss": 1.0158, + "mean_token_accuracy": 0.6942140460014343, + "num_tokens": 42951568.0, + "step": 1666 + }, + { + "epoch": 0.18306611025697342, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.3891286849975586, + "learning_rate": 6.09809663250366e-07, + "loss": 1.0958, + "mean_token_accuracy": 0.6866564750671387, + "num_tokens": 42975783.0, + "step": 1667 + }, + { + "epoch": 0.1831759279595871, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1470401287078857, + "learning_rate": 6.101756954612005e-07, + "loss": 1.1186, + "mean_token_accuracy": 0.6711210608482361, + "num_tokens": 43004209.0, + "step": 1668 + }, + { + "epoch": 0.18328574566220074, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.40856671333313, + "learning_rate": 6.105417276720352e-07, + "loss": 1.0338, + "mean_token_accuracy": 0.6876379251480103, + "num_tokens": 43028852.0, + "step": 1669 + }, + { + "epoch": 0.1833955633648144, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.402705192565918, + "learning_rate": 6.109077598828697e-07, + "loss": 1.0046, + "mean_token_accuracy": 0.6934975385665894, + "num_tokens": 43051679.0, + "step": 1670 + }, + { + "epoch": 0.18350538106742806, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.234795331954956, + "learning_rate": 6.112737920937042e-07, + "loss": 1.0933, + "mean_token_accuracy": 0.6771068572998047, + "num_tokens": 43079437.0, + "step": 1671 + }, + { + "epoch": 0.18361519877004173, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.010127067565918, + "learning_rate": 6.116398243045388e-07, + "loss": 1.0641, + "mean_token_accuracy": 0.6812338829040527, + "num_tokens": 43111882.0, + "step": 1672 + }, + { + "epoch": 0.1837250164726554, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.328385353088379, + "learning_rate": 6.120058565153733e-07, + "loss": 1.0688, + "mean_token_accuracy": 0.6919026374816895, + "num_tokens": 43138332.0, + "step": 1673 + }, + { + "epoch": 0.18383483417526905, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3129994869232178, + "learning_rate": 6.12371888726208e-07, + "loss": 1.0655, + "mean_token_accuracy": 0.6865989565849304, + "num_tokens": 43163166.0, + "step": 1674 + }, + { + "epoch": 0.18394465187788273, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0935699939727783, + "learning_rate": 6.127379209370425e-07, + "loss": 1.0185, + "mean_token_accuracy": 0.6983241438865662, + "num_tokens": 43192052.0, + "step": 1675 + }, + { + "epoch": 0.18405446958049637, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2033746242523193, + "learning_rate": 6.13103953147877e-07, + "loss": 0.9767, + "mean_token_accuracy": 0.7020092010498047, + "num_tokens": 43218645.0, + "step": 1676 + }, + { + "epoch": 0.18416428728311005, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.443624496459961, + "learning_rate": 6.134699853587116e-07, + "loss": 1.075, + "mean_token_accuracy": 0.6739616394042969, + "num_tokens": 43241509.0, + "step": 1677 + }, + { + "epoch": 0.1842741049857237, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4803085327148438, + "learning_rate": 6.138360175695461e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.6956441402435303, + "num_tokens": 43264698.0, + "step": 1678 + }, + { + "epoch": 0.18438392268833736, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.490652084350586, + "learning_rate": 6.142020497803806e-07, + "loss": 0.9877, + "mean_token_accuracy": 0.7066114544868469, + "num_tokens": 43287337.0, + "step": 1679 + }, + { + "epoch": 0.184493740390951, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4470341205596924, + "learning_rate": 6.145680819912153e-07, + "loss": 1.0734, + "mean_token_accuracy": 0.68474280834198, + "num_tokens": 43313430.0, + "step": 1680 + }, + { + "epoch": 0.18460355809356468, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.40048885345459, + "learning_rate": 6.149341142020498e-07, + "loss": 1.0472, + "mean_token_accuracy": 0.6879701614379883, + "num_tokens": 43337787.0, + "step": 1681 + }, + { + "epoch": 0.18471337579617833, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.981128215789795, + "learning_rate": 6.153001464128843e-07, + "loss": 1.099, + "mean_token_accuracy": 0.6755709052085876, + "num_tokens": 43373378.0, + "step": 1682 + }, + { + "epoch": 0.184823193498792, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9781033992767334, + "learning_rate": 6.156661786237189e-07, + "loss": 0.9567, + "mean_token_accuracy": 0.710017204284668, + "num_tokens": 43404353.0, + "step": 1683 + }, + { + "epoch": 0.18493301120140568, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1201698780059814, + "learning_rate": 6.160322108345534e-07, + "loss": 1.0886, + "mean_token_accuracy": 0.6811988949775696, + "num_tokens": 43433934.0, + "step": 1684 + }, + { + "epoch": 0.18504282890401932, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9762886762619019, + "learning_rate": 6.16398243045388e-07, + "loss": 0.9858, + "mean_token_accuracy": 0.698521077632904, + "num_tokens": 43466980.0, + "step": 1685 + }, + { + "epoch": 0.185152646606633, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.188091516494751, + "learning_rate": 6.167642752562226e-07, + "loss": 1.0163, + "mean_token_accuracy": 0.7036364674568176, + "num_tokens": 43494439.0, + "step": 1686 + }, + { + "epoch": 0.18526246430924664, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.311044692993164, + "learning_rate": 6.171303074670571e-07, + "loss": 1.1318, + "mean_token_accuracy": 0.6715076565742493, + "num_tokens": 43521827.0, + "step": 1687 + }, + { + "epoch": 0.18537228201186032, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.451606512069702, + "learning_rate": 6.174963396778916e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6745298504829407, + "num_tokens": 43546844.0, + "step": 1688 + }, + { + "epoch": 0.18548209971447396, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1434550285339355, + "learning_rate": 6.178623718887262e-07, + "loss": 1.1073, + "mean_token_accuracy": 0.6683220267295837, + "num_tokens": 43579098.0, + "step": 1689 + }, + { + "epoch": 0.18559191741708764, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.6287424564361572, + "learning_rate": 6.182284040995608e-07, + "loss": 0.9049, + "mean_token_accuracy": 0.7257931232452393, + "num_tokens": 43599155.0, + "step": 1690 + }, + { + "epoch": 0.18570173511970128, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.44008731842041, + "learning_rate": 6.185944363103954e-07, + "loss": 1.0101, + "mean_token_accuracy": 0.6983397603034973, + "num_tokens": 43622669.0, + "step": 1691 + }, + { + "epoch": 0.18581155282231496, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.375753402709961, + "learning_rate": 6.189604685212299e-07, + "loss": 0.9992, + "mean_token_accuracy": 0.6988521814346313, + "num_tokens": 43646422.0, + "step": 1692 + }, + { + "epoch": 0.18592137052492863, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0843305587768555, + "learning_rate": 6.193265007320644e-07, + "loss": 1.0009, + "mean_token_accuracy": 0.6996857523918152, + "num_tokens": 43676400.0, + "step": 1693 + }, + { + "epoch": 0.18603118822754228, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.6912786960601807, + "learning_rate": 6.19692532942899e-07, + "loss": 1.0083, + "mean_token_accuracy": 0.6988992691040039, + "num_tokens": 43697193.0, + "step": 1694 + }, + { + "epoch": 0.18614100593015595, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.454721689224243, + "learning_rate": 6.200585651537335e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.6889728307723999, + "num_tokens": 43721079.0, + "step": 1695 + }, + { + "epoch": 0.1862508236327696, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3315775394439697, + "learning_rate": 6.204245973645681e-07, + "loss": 1.0396, + "mean_token_accuracy": 0.6940385103225708, + "num_tokens": 43747631.0, + "step": 1696 + }, + { + "epoch": 0.18636064133538327, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.146324634552002, + "learning_rate": 6.207906295754027e-07, + "loss": 1.0103, + "mean_token_accuracy": 0.6956826448440552, + "num_tokens": 43778341.0, + "step": 1697 + }, + { + "epoch": 0.18647045903799692, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.399505853652954, + "learning_rate": 6.211566617862372e-07, + "loss": 1.0074, + "mean_token_accuracy": 0.6984342336654663, + "num_tokens": 43802076.0, + "step": 1698 + }, + { + "epoch": 0.1865802767406106, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3117434978485107, + "learning_rate": 6.215226939970716e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6788157820701599, + "num_tokens": 43829357.0, + "step": 1699 + }, + { + "epoch": 0.18669009444322424, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4607322216033936, + "learning_rate": 6.218887262079063e-07, + "loss": 0.9577, + "mean_token_accuracy": 0.7050804495811462, + "num_tokens": 43852687.0, + "step": 1700 + }, + { + "epoch": 0.1867999121458379, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.364621639251709, + "learning_rate": 6.222547584187409e-07, + "loss": 1.0544, + "mean_token_accuracy": 0.6909194588661194, + "num_tokens": 43877297.0, + "step": 1701 + }, + { + "epoch": 0.18690972984845158, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4617834091186523, + "learning_rate": 6.226207906295753e-07, + "loss": 0.9953, + "mean_token_accuracy": 0.6966697573661804, + "num_tokens": 43899048.0, + "step": 1702 + }, + { + "epoch": 0.18701954755106523, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.416557788848877, + "learning_rate": 6.2298682284041e-07, + "loss": 0.928, + "mean_token_accuracy": 0.7171322703361511, + "num_tokens": 43921618.0, + "step": 1703 + }, + { + "epoch": 0.1871293652536789, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.977636694908142, + "learning_rate": 6.233528550512444e-07, + "loss": 1.0804, + "mean_token_accuracy": 0.6892393231391907, + "num_tokens": 43952213.0, + "step": 1704 + }, + { + "epoch": 0.18723918295629255, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.174431324005127, + "learning_rate": 6.237188872620789e-07, + "loss": 1.0088, + "mean_token_accuracy": 0.7017797827720642, + "num_tokens": 43979710.0, + "step": 1705 + }, + { + "epoch": 0.18734900065890622, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.2428596019744873, + "learning_rate": 6.240849194729137e-07, + "loss": 1.0107, + "mean_token_accuracy": 0.6958329677581787, + "num_tokens": 44005736.0, + "step": 1706 + }, + { + "epoch": 0.18745881836151987, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 1.9926947355270386, + "learning_rate": 6.244509516837481e-07, + "loss": 1.109, + "mean_token_accuracy": 0.6792625188827515, + "num_tokens": 44036907.0, + "step": 1707 + }, + { + "epoch": 0.18756863606413354, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.0023274421691895, + "learning_rate": 6.248169838945828e-07, + "loss": 1.0658, + "mean_token_accuracy": 0.6811660528182983, + "num_tokens": 44067574.0, + "step": 1708 + }, + { + "epoch": 0.1876784537667472, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.3018524646759033, + "learning_rate": 6.251830161054172e-07, + "loss": 1.099, + "mean_token_accuracy": 0.6750198006629944, + "num_tokens": 44092129.0, + "step": 1709 + }, + { + "epoch": 0.18778827146936086, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.070262908935547, + "learning_rate": 6.255490483162517e-07, + "loss": 1.1204, + "mean_token_accuracy": 0.6734856367111206, + "num_tokens": 44124520.0, + "step": 1710 + }, + { + "epoch": 0.18789808917197454, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.3226277828216553, + "learning_rate": 6.259150805270863e-07, + "loss": 1.0703, + "mean_token_accuracy": 0.6895779371261597, + "num_tokens": 44150189.0, + "step": 1711 + }, + { + "epoch": 0.18800790687458818, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.511606216430664, + "learning_rate": 6.262811127379209e-07, + "loss": 0.9838, + "mean_token_accuracy": 0.7047934532165527, + "num_tokens": 44171622.0, + "step": 1712 + }, + { + "epoch": 0.18811772457720186, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.3250513076782227, + "learning_rate": 6.266471449487554e-07, + "loss": 1.074, + "mean_token_accuracy": 0.690969705581665, + "num_tokens": 44196722.0, + "step": 1713 + }, + { + "epoch": 0.1882275422798155, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.0508322715759277, + "learning_rate": 6.2701317715959e-07, + "loss": 1.1121, + "mean_token_accuracy": 0.6697304844856262, + "num_tokens": 44228788.0, + "step": 1714 + }, + { + "epoch": 0.18833735998242918, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.488111734390259, + "learning_rate": 6.273792093704245e-07, + "loss": 1.0586, + "mean_token_accuracy": 0.6836752891540527, + "num_tokens": 44253946.0, + "step": 1715 + }, + { + "epoch": 0.18844717768504282, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.4038469791412354, + "learning_rate": 6.27745241581259e-07, + "loss": 1.0631, + "mean_token_accuracy": 0.6822866201400757, + "num_tokens": 44277666.0, + "step": 1716 + }, + { + "epoch": 0.1885569953876565, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.073162317276001, + "learning_rate": 6.281112737920937e-07, + "loss": 0.9949, + "mean_token_accuracy": 0.69987952709198, + "num_tokens": 44306333.0, + "step": 1717 + }, + { + "epoch": 0.18866681309027014, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2022604942321777, + "learning_rate": 6.284773060029282e-07, + "loss": 1.079, + "mean_token_accuracy": 0.6770700812339783, + "num_tokens": 44333933.0, + "step": 1718 + }, + { + "epoch": 0.18877663079288381, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.3119776248931885, + "learning_rate": 6.288433382137627e-07, + "loss": 0.9926, + "mean_token_accuracy": 0.6987459659576416, + "num_tokens": 44358642.0, + "step": 1719 + }, + { + "epoch": 0.18888644849549746, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.4802517890930176, + "learning_rate": 6.292093704245973e-07, + "loss": 1.0247, + "mean_token_accuracy": 0.6871346235275269, + "num_tokens": 44381083.0, + "step": 1720 + }, + { + "epoch": 0.18899626619811113, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.27305269241333, + "learning_rate": 6.295754026354318e-07, + "loss": 1.0768, + "mean_token_accuracy": 0.6849287152290344, + "num_tokens": 44406847.0, + "step": 1721 + }, + { + "epoch": 0.1891060839007248, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2238640785217285, + "learning_rate": 6.299414348462665e-07, + "loss": 1.0625, + "mean_token_accuracy": 0.67763751745224, + "num_tokens": 44433809.0, + "step": 1722 + }, + { + "epoch": 0.18921590160333845, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.408456802368164, + "learning_rate": 6.30307467057101e-07, + "loss": 1.0537, + "mean_token_accuracy": 0.6834909319877625, + "num_tokens": 44458042.0, + "step": 1723 + }, + { + "epoch": 0.18932571930595213, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.706160068511963, + "learning_rate": 6.306734992679355e-07, + "loss": 1.0057, + "mean_token_accuracy": 0.6970251798629761, + "num_tokens": 44478035.0, + "step": 1724 + }, + { + "epoch": 0.18943553700856577, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 1.960123062133789, + "learning_rate": 6.310395314787701e-07, + "loss": 1.1106, + "mean_token_accuracy": 0.6680430769920349, + "num_tokens": 44513865.0, + "step": 1725 + }, + { + "epoch": 0.18954535471117945, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.6036996841430664, + "learning_rate": 6.314055636896046e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.698746919631958, + "num_tokens": 44534733.0, + "step": 1726 + }, + { + "epoch": 0.1896551724137931, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2359678745269775, + "learning_rate": 6.317715959004391e-07, + "loss": 0.9808, + "mean_token_accuracy": 0.7033553123474121, + "num_tokens": 44560374.0, + "step": 1727 + }, + { + "epoch": 0.18976499011640677, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.739323616027832, + "learning_rate": 6.321376281112738e-07, + "loss": 1.0204, + "mean_token_accuracy": 0.7115684151649475, + "num_tokens": 44579550.0, + "step": 1728 + }, + { + "epoch": 0.1898748078190204, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.18229079246521, + "learning_rate": 6.325036603221083e-07, + "loss": 1.057, + "mean_token_accuracy": 0.6811222434043884, + "num_tokens": 44605935.0, + "step": 1729 + }, + { + "epoch": 0.1899846255216341, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.4705002307891846, + "learning_rate": 6.328696925329428e-07, + "loss": 1.0238, + "mean_token_accuracy": 0.6948506832122803, + "num_tokens": 44628919.0, + "step": 1730 + }, + { + "epoch": 0.19009444322424776, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2321617603302, + "learning_rate": 6.332357247437774e-07, + "loss": 1.0524, + "mean_token_accuracy": 0.6816362738609314, + "num_tokens": 44656105.0, + "step": 1731 + }, + { + "epoch": 0.1902042609268614, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1249117851257324, + "learning_rate": 6.336017569546119e-07, + "loss": 1.061, + "mean_token_accuracy": 0.6783463954925537, + "num_tokens": 44686900.0, + "step": 1732 + }, + { + "epoch": 0.19031407862947508, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.5491366386413574, + "learning_rate": 6.339677891654465e-07, + "loss": 1.0055, + "mean_token_accuracy": 0.6983740925788879, + "num_tokens": 44710054.0, + "step": 1733 + }, + { + "epoch": 0.19042389633208873, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.63578724861145, + "learning_rate": 6.343338213762811e-07, + "loss": 0.9877, + "mean_token_accuracy": 0.7054013013839722, + "num_tokens": 44730033.0, + "step": 1734 + }, + { + "epoch": 0.1905337140347024, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2844836711883545, + "learning_rate": 6.346998535871156e-07, + "loss": 0.9539, + "mean_token_accuracy": 0.7080437541007996, + "num_tokens": 44756831.0, + "step": 1735 + }, + { + "epoch": 0.19064353173731605, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2566511631011963, + "learning_rate": 6.350658857979501e-07, + "loss": 1.0333, + "mean_token_accuracy": 0.6885890960693359, + "num_tokens": 44782982.0, + "step": 1736 + }, + { + "epoch": 0.19075334943992972, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2642407417297363, + "learning_rate": 6.354319180087847e-07, + "loss": 1.1559, + "mean_token_accuracy": 0.6658865809440613, + "num_tokens": 44811212.0, + "step": 1737 + }, + { + "epoch": 0.19086316714254337, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.480051040649414, + "learning_rate": 6.357979502196193e-07, + "loss": 1.0086, + "mean_token_accuracy": 0.6996333599090576, + "num_tokens": 44833233.0, + "step": 1738 + }, + { + "epoch": 0.19097298484515704, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.0750105381011963, + "learning_rate": 6.361639824304539e-07, + "loss": 1.0557, + "mean_token_accuracy": 0.6880769729614258, + "num_tokens": 44861753.0, + "step": 1739 + }, + { + "epoch": 0.1910828025477707, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.200918197631836, + "learning_rate": 6.365300146412884e-07, + "loss": 1.0389, + "mean_token_accuracy": 0.6980140209197998, + "num_tokens": 44887824.0, + "step": 1740 + }, + { + "epoch": 0.19119262025038436, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.287238121032715, + "learning_rate": 6.368960468521229e-07, + "loss": 1.0245, + "mean_token_accuracy": 0.6858128309249878, + "num_tokens": 44914371.0, + "step": 1741 + }, + { + "epoch": 0.19130243795299803, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.103818416595459, + "learning_rate": 6.372620790629575e-07, + "loss": 1.0915, + "mean_token_accuracy": 0.6752873063087463, + "num_tokens": 44943379.0, + "step": 1742 + }, + { + "epoch": 0.19141225565561168, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.436122179031372, + "learning_rate": 6.37628111273792e-07, + "loss": 1.1092, + "mean_token_accuracy": 0.6793209314346313, + "num_tokens": 44967204.0, + "step": 1743 + }, + { + "epoch": 0.19152207335822535, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.231194257736206, + "learning_rate": 6.379941434846266e-07, + "loss": 0.9932, + "mean_token_accuracy": 0.6984097957611084, + "num_tokens": 44993218.0, + "step": 1744 + }, + { + "epoch": 0.191631891060839, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.4344916343688965, + "learning_rate": 6.383601756954612e-07, + "loss": 1.0152, + "mean_token_accuracy": 0.6945403814315796, + "num_tokens": 45016771.0, + "step": 1745 + }, + { + "epoch": 0.19174170876345267, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.462852716445923, + "learning_rate": 6.387262079062957e-07, + "loss": 1.057, + "mean_token_accuracy": 0.685813844203949, + "num_tokens": 45041696.0, + "step": 1746 + }, + { + "epoch": 0.19185152646606632, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.2664835453033447, + "learning_rate": 6.390922401171302e-07, + "loss": 1.0694, + "mean_token_accuracy": 0.6845830082893372, + "num_tokens": 45067615.0, + "step": 1747 + }, + { + "epoch": 0.19196134416868, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4122960567474365, + "learning_rate": 6.394582723279648e-07, + "loss": 1.0685, + "mean_token_accuracy": 0.6924625635147095, + "num_tokens": 45092065.0, + "step": 1748 + }, + { + "epoch": 0.19207116187129367, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4007699489593506, + "learning_rate": 6.398243045387994e-07, + "loss": 1.0855, + "mean_token_accuracy": 0.6821146607398987, + "num_tokens": 45116555.0, + "step": 1749 + }, + { + "epoch": 0.1921809795739073, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2338926792144775, + "learning_rate": 6.401903367496339e-07, + "loss": 1.0941, + "mean_token_accuracy": 0.6800568699836731, + "num_tokens": 45142968.0, + "step": 1750 + }, + { + "epoch": 0.19229079727652099, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.6097912788391113, + "learning_rate": 6.405563689604685e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7013169527053833, + "num_tokens": 45162421.0, + "step": 1751 + }, + { + "epoch": 0.19240061497913463, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.27032208442688, + "learning_rate": 6.40922401171303e-07, + "loss": 0.9512, + "mean_token_accuracy": 0.7115261554718018, + "num_tokens": 45188853.0, + "step": 1752 + }, + { + "epoch": 0.1925104326817483, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.355241298675537, + "learning_rate": 6.412884333821375e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.6951295137405396, + "num_tokens": 45211908.0, + "step": 1753 + }, + { + "epoch": 0.19262025038436195, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2596542835235596, + "learning_rate": 6.416544655929722e-07, + "loss": 0.9515, + "mean_token_accuracy": 0.7119032740592957, + "num_tokens": 45235897.0, + "step": 1754 + }, + { + "epoch": 0.19273006808697563, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1829559803009033, + "learning_rate": 6.420204978038067e-07, + "loss": 1.1885, + "mean_token_accuracy": 0.653622031211853, + "num_tokens": 45266668.0, + "step": 1755 + }, + { + "epoch": 0.19283988578958927, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3195717334747314, + "learning_rate": 6.423865300146413e-07, + "loss": 1.0855, + "mean_token_accuracy": 0.6726996898651123, + "num_tokens": 45291961.0, + "step": 1756 + }, + { + "epoch": 0.19294970349220295, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1738247871398926, + "learning_rate": 6.427525622254758e-07, + "loss": 1.0498, + "mean_token_accuracy": 0.6887697577476501, + "num_tokens": 45321224.0, + "step": 1757 + }, + { + "epoch": 0.1930595211948166, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2115557193756104, + "learning_rate": 6.431185944363103e-07, + "loss": 1.041, + "mean_token_accuracy": 0.6931350827217102, + "num_tokens": 45349120.0, + "step": 1758 + }, + { + "epoch": 0.19316933889743026, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4693217277526855, + "learning_rate": 6.434846266471449e-07, + "loss": 0.9592, + "mean_token_accuracy": 0.7099024057388306, + "num_tokens": 45370609.0, + "step": 1759 + }, + { + "epoch": 0.19327915660004394, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0425853729248047, + "learning_rate": 6.438506588579795e-07, + "loss": 1.0217, + "mean_token_accuracy": 0.6958690881729126, + "num_tokens": 45402004.0, + "step": 1760 + }, + { + "epoch": 0.19338897430265758, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.566129207611084, + "learning_rate": 6.44216691068814e-07, + "loss": 0.9704, + "mean_token_accuracy": 0.7087500095367432, + "num_tokens": 45423421.0, + "step": 1761 + }, + { + "epoch": 0.19349879200527126, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.149144411087036, + "learning_rate": 6.445827232796486e-07, + "loss": 0.975, + "mean_token_accuracy": 0.7090919613838196, + "num_tokens": 45449405.0, + "step": 1762 + }, + { + "epoch": 0.1936086097078849, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.505025625228882, + "learning_rate": 6.449487554904831e-07, + "loss": 0.9851, + "mean_token_accuracy": 0.6943643093109131, + "num_tokens": 45470264.0, + "step": 1763 + }, + { + "epoch": 0.19371842741049858, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.043926477432251, + "learning_rate": 6.453147877013176e-07, + "loss": 0.9835, + "mean_token_accuracy": 0.7102909088134766, + "num_tokens": 45501935.0, + "step": 1764 + }, + { + "epoch": 0.19382824511311222, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.277949094772339, + "learning_rate": 6.456808199121523e-07, + "loss": 1.0801, + "mean_token_accuracy": 0.6812334060668945, + "num_tokens": 45528933.0, + "step": 1765 + }, + { + "epoch": 0.1939380628157259, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.380723476409912, + "learning_rate": 6.460468521229868e-07, + "loss": 0.9397, + "mean_token_accuracy": 0.7136078476905823, + "num_tokens": 45551417.0, + "step": 1766 + }, + { + "epoch": 0.19404788051833954, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.0439834594726562, + "learning_rate": 6.464128843338213e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.6925650835037231, + "num_tokens": 45582175.0, + "step": 1767 + }, + { + "epoch": 0.19415769822095322, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4164178371429443, + "learning_rate": 6.467789165446559e-07, + "loss": 1.1005, + "mean_token_accuracy": 0.6718156337738037, + "num_tokens": 45606661.0, + "step": 1768 + }, + { + "epoch": 0.1942675159235669, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.371220111846924, + "learning_rate": 6.471449487554904e-07, + "loss": 1.0604, + "mean_token_accuracy": 0.6816762089729309, + "num_tokens": 45632245.0, + "step": 1769 + }, + { + "epoch": 0.19437733362618054, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.936739206314087, + "learning_rate": 6.475109809663251e-07, + "loss": 1.0776, + "mean_token_accuracy": 0.6833305358886719, + "num_tokens": 45666175.0, + "step": 1770 + }, + { + "epoch": 0.1944871513287942, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3370845317840576, + "learning_rate": 6.478770131771596e-07, + "loss": 1.0415, + "mean_token_accuracy": 0.6981831789016724, + "num_tokens": 45692819.0, + "step": 1771 + }, + { + "epoch": 0.19459696903140786, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1020255088806152, + "learning_rate": 6.482430453879941e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.6938903331756592, + "num_tokens": 45721537.0, + "step": 1772 + }, + { + "epoch": 0.19470678673402153, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4717204570770264, + "learning_rate": 6.486090775988287e-07, + "loss": 1.0795, + "mean_token_accuracy": 0.6871129870414734, + "num_tokens": 45745690.0, + "step": 1773 + }, + { + "epoch": 0.19481660443663518, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3787591457366943, + "learning_rate": 6.489751098096632e-07, + "loss": 1.1069, + "mean_token_accuracy": 0.673858642578125, + "num_tokens": 45771981.0, + "step": 1774 + }, + { + "epoch": 0.19492642213924885, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.5197243690490723, + "learning_rate": 6.493411420204977e-07, + "loss": 1.0369, + "mean_token_accuracy": 0.6947212219238281, + "num_tokens": 45796644.0, + "step": 1775 + }, + { + "epoch": 0.1950362398418625, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.8195533752441406, + "learning_rate": 6.497071742313324e-07, + "loss": 1.0753, + "mean_token_accuracy": 0.6778128147125244, + "num_tokens": 45815886.0, + "step": 1776 + }, + { + "epoch": 0.19514605754447617, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2477166652679443, + "learning_rate": 6.500732064421669e-07, + "loss": 1.0486, + "mean_token_accuracy": 0.6867901086807251, + "num_tokens": 45843574.0, + "step": 1777 + }, + { + "epoch": 0.19525587524708984, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.200855255126953, + "learning_rate": 6.504392386530014e-07, + "loss": 1.016, + "mean_token_accuracy": 0.69814532995224, + "num_tokens": 45870615.0, + "step": 1778 + }, + { + "epoch": 0.1953656929497035, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4827842712402344, + "learning_rate": 6.50805270863836e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.696321964263916, + "num_tokens": 45892574.0, + "step": 1779 + }, + { + "epoch": 0.19547551065231716, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3913493156433105, + "learning_rate": 6.511713030746705e-07, + "loss": 1.0939, + "mean_token_accuracy": 0.6762992143630981, + "num_tokens": 45917428.0, + "step": 1780 + }, + { + "epoch": 0.1955853283549308, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.072939157485962, + "learning_rate": 6.515373352855051e-07, + "loss": 1.0188, + "mean_token_accuracy": 0.6910207867622375, + "num_tokens": 45947191.0, + "step": 1781 + }, + { + "epoch": 0.19569514605754448, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2359707355499268, + "learning_rate": 6.519033674963397e-07, + "loss": 0.996, + "mean_token_accuracy": 0.7006053924560547, + "num_tokens": 45973562.0, + "step": 1782 + }, + { + "epoch": 0.19580496376015813, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.317333459854126, + "learning_rate": 6.522693997071742e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.7027888298034668, + "num_tokens": 46000779.0, + "step": 1783 + }, + { + "epoch": 0.1959147814627718, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2924892902374268, + "learning_rate": 6.526354319180087e-07, + "loss": 1.0355, + "mean_token_accuracy": 0.6929695010185242, + "num_tokens": 46025924.0, + "step": 1784 + }, + { + "epoch": 0.19602459916538545, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2524046897888184, + "learning_rate": 6.530014641288433e-07, + "loss": 1.0464, + "mean_token_accuracy": 0.6855796575546265, + "num_tokens": 46052335.0, + "step": 1785 + }, + { + "epoch": 0.19613441686799912, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.3318748474121094, + "learning_rate": 6.533674963396779e-07, + "loss": 0.9711, + "mean_token_accuracy": 0.7071787118911743, + "num_tokens": 46075495.0, + "step": 1786 + }, + { + "epoch": 0.1962442345706128, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.655519723892212, + "learning_rate": 6.537335285505125e-07, + "loss": 0.992, + "mean_token_accuracy": 0.6962170004844666, + "num_tokens": 46094772.0, + "step": 1787 + }, + { + "epoch": 0.19635405227322644, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.273876190185547, + "learning_rate": 6.54099560761347e-07, + "loss": 1.0989, + "mean_token_accuracy": 0.6730601787567139, + "num_tokens": 46123044.0, + "step": 1788 + }, + { + "epoch": 0.19646386997584012, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.329338312149048, + "learning_rate": 6.544655929721815e-07, + "loss": 1.0336, + "mean_token_accuracy": 0.6919792294502258, + "num_tokens": 46146887.0, + "step": 1789 + }, + { + "epoch": 0.19657368767845376, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.488285779953003, + "learning_rate": 6.548316251830161e-07, + "loss": 1.0263, + "mean_token_accuracy": 0.6997965574264526, + "num_tokens": 46168260.0, + "step": 1790 + }, + { + "epoch": 0.19668350538106744, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.50103759765625, + "learning_rate": 6.551976573938506e-07, + "loss": 1.0838, + "mean_token_accuracy": 0.6844652891159058, + "num_tokens": 46190266.0, + "step": 1791 + }, + { + "epoch": 0.19679332308368108, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4022815227508545, + "learning_rate": 6.555636896046852e-07, + "loss": 1.0515, + "mean_token_accuracy": 0.687410831451416, + "num_tokens": 46214139.0, + "step": 1792 + }, + { + "epoch": 0.19690314078629476, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.6452372074127197, + "learning_rate": 6.559297218155198e-07, + "loss": 1.0544, + "mean_token_accuracy": 0.6922059059143066, + "num_tokens": 46235396.0, + "step": 1793 + }, + { + "epoch": 0.1970129584889084, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2305638790130615, + "learning_rate": 6.562957540263543e-07, + "loss": 1.016, + "mean_token_accuracy": 0.6941735744476318, + "num_tokens": 46261539.0, + "step": 1794 + }, + { + "epoch": 0.19712277619152208, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.295442819595337, + "learning_rate": 6.566617862371888e-07, + "loss": 1.0203, + "mean_token_accuracy": 0.6951667070388794, + "num_tokens": 46286833.0, + "step": 1795 + }, + { + "epoch": 0.19723259389413572, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.2373523712158203, + "learning_rate": 6.570278184480234e-07, + "loss": 1.0654, + "mean_token_accuracy": 0.683392345905304, + "num_tokens": 46313733.0, + "step": 1796 + }, + { + "epoch": 0.1973424115967494, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.3224542140960693, + "learning_rate": 6.57393850658858e-07, + "loss": 1.0672, + "mean_token_accuracy": 0.6844637393951416, + "num_tokens": 46337087.0, + "step": 1797 + }, + { + "epoch": 0.19745222929936307, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.785236358642578, + "learning_rate": 6.577598828696925e-07, + "loss": 0.965, + "mean_token_accuracy": 0.7027713060379028, + "num_tokens": 46356648.0, + "step": 1798 + }, + { + "epoch": 0.19756204700197671, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.3532118797302246, + "learning_rate": 6.581259150805271e-07, + "loss": 0.9349, + "mean_token_accuracy": 0.7135041952133179, + "num_tokens": 46378939.0, + "step": 1799 + }, + { + "epoch": 0.1976718647045904, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.1250715255737305, + "learning_rate": 6.584919472913616e-07, + "loss": 1.0372, + "mean_token_accuracy": 0.6860877871513367, + "num_tokens": 46408171.0, + "step": 1800 + }, + { + "epoch": 0.19778168240720403, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4264516830444336, + "learning_rate": 6.588579795021961e-07, + "loss": 0.9941, + "mean_token_accuracy": 0.7031900882720947, + "num_tokens": 46431563.0, + "step": 1801 + }, + { + "epoch": 0.1978915001098177, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4174013137817383, + "learning_rate": 6.592240117130308e-07, + "loss": 1.0246, + "mean_token_accuracy": 0.7007296681404114, + "num_tokens": 46455897.0, + "step": 1802 + }, + { + "epoch": 0.19800131781243135, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.138324499130249, + "learning_rate": 6.595900439238653e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6750045418739319, + "num_tokens": 46485421.0, + "step": 1803 + }, + { + "epoch": 0.19811113551504503, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.7413251399993896, + "learning_rate": 6.599560761346999e-07, + "loss": 0.9988, + "mean_token_accuracy": 0.6985788345336914, + "num_tokens": 46504392.0, + "step": 1804 + }, + { + "epoch": 0.19822095321765867, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4720351696014404, + "learning_rate": 6.603221083455344e-07, + "loss": 0.9475, + "mean_token_accuracy": 0.7148571014404297, + "num_tokens": 46526987.0, + "step": 1805 + }, + { + "epoch": 0.19833077092027235, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.3252789974212646, + "learning_rate": 6.606881405563689e-07, + "loss": 1.0109, + "mean_token_accuracy": 0.6972226500511169, + "num_tokens": 46552167.0, + "step": 1806 + }, + { + "epoch": 0.19844058862288602, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.330960273742676, + "learning_rate": 6.610541727672035e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.6873450875282288, + "num_tokens": 46578431.0, + "step": 1807 + }, + { + "epoch": 0.19855040632549967, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4630868434906006, + "learning_rate": 6.614202049780381e-07, + "loss": 0.9141, + "mean_token_accuracy": 0.7196933031082153, + "num_tokens": 46597882.0, + "step": 1808 + }, + { + "epoch": 0.19866022402811334, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.1454153060913086, + "learning_rate": 6.617862371888726e-07, + "loss": 1.0884, + "mean_token_accuracy": 0.6916476488113403, + "num_tokens": 46627173.0, + "step": 1809 + }, + { + "epoch": 0.198770041730727, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.3254811763763428, + "learning_rate": 6.621522693997072e-07, + "loss": 1.043, + "mean_token_accuracy": 0.6950938701629639, + "num_tokens": 46652241.0, + "step": 1810 + }, + { + "epoch": 0.19887985943334066, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.214266061782837, + "learning_rate": 6.625183016105417e-07, + "loss": 1.0856, + "mean_token_accuracy": 0.6757978200912476, + "num_tokens": 46681172.0, + "step": 1811 + }, + { + "epoch": 0.1989896771359543, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.0648398399353027, + "learning_rate": 6.628843338213762e-07, + "loss": 1.0679, + "mean_token_accuracy": 0.6836932897567749, + "num_tokens": 46711892.0, + "step": 1812 + }, + { + "epoch": 0.19909949483856798, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.003197193145752, + "learning_rate": 6.632503660322109e-07, + "loss": 1.069, + "mean_token_accuracy": 0.6797057390213013, + "num_tokens": 46744839.0, + "step": 1813 + }, + { + "epoch": 0.19920931254118163, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.3230249881744385, + "learning_rate": 6.636163982430454e-07, + "loss": 0.9345, + "mean_token_accuracy": 0.7054651975631714, + "num_tokens": 46768706.0, + "step": 1814 + }, + { + "epoch": 0.1993191302437953, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.665834426879883, + "learning_rate": 6.639824304538799e-07, + "loss": 0.9886, + "mean_token_accuracy": 0.7017563581466675, + "num_tokens": 46789029.0, + "step": 1815 + }, + { + "epoch": 0.19942894794640897, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.1464223861694336, + "learning_rate": 6.643484626647145e-07, + "loss": 1.0698, + "mean_token_accuracy": 0.6813626289367676, + "num_tokens": 46819385.0, + "step": 1816 + }, + { + "epoch": 0.19953876564902262, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.3180761337280273, + "learning_rate": 6.64714494875549e-07, + "loss": 1.0515, + "mean_token_accuracy": 0.6858088970184326, + "num_tokens": 46845617.0, + "step": 1817 + }, + { + "epoch": 0.1996485833516363, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.492344856262207, + "learning_rate": 6.650805270863836e-07, + "loss": 0.9777, + "mean_token_accuracy": 0.7029760479927063, + "num_tokens": 46867168.0, + "step": 1818 + }, + { + "epoch": 0.19975840105424994, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.37735915184021, + "learning_rate": 6.654465592972182e-07, + "loss": 0.9792, + "mean_token_accuracy": 0.7056591510772705, + "num_tokens": 46893199.0, + "step": 1819 + }, + { + "epoch": 0.1998682187568636, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.783790349960327, + "learning_rate": 6.658125915080527e-07, + "loss": 1.0181, + "mean_token_accuracy": 0.6962101459503174, + "num_tokens": 46912484.0, + "step": 1820 + }, + { + "epoch": 0.19997803645947726, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 1.9601514339447021, + "learning_rate": 6.661786237188873e-07, + "loss": 1.1008, + "mean_token_accuracy": 0.6838592886924744, + "num_tokens": 46949071.0, + "step": 1821 + }, + { + "epoch": 0.20008785416209093, + "ewc_loss": 1.0728836059570312e-05, + "grad_norm": 2.3164560794830322, + "learning_rate": 6.665446559297218e-07, + "loss": 0.9896, + "mean_token_accuracy": 0.6973059177398682, + "num_tokens": 46976660.0, + "step": 1822 + }, + { + "epoch": 0.20019767186470458, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2299611568450928, + "learning_rate": 6.669106881405563e-07, + "loss": 0.9126, + "mean_token_accuracy": 0.7263529896736145, + "num_tokens": 47003264.0, + "step": 1823 + }, + { + "epoch": 0.20030748956731825, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.9882316589355469, + "learning_rate": 6.67276720351391e-07, + "loss": 1.0335, + "mean_token_accuracy": 0.6871451139450073, + "num_tokens": 47036317.0, + "step": 1824 + }, + { + "epoch": 0.2004173072699319, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.477187395095825, + "learning_rate": 6.676427525622255e-07, + "loss": 1.0609, + "mean_token_accuracy": 0.6861164569854736, + "num_tokens": 47059959.0, + "step": 1825 + }, + { + "epoch": 0.20052712497254557, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.391846179962158, + "learning_rate": 6.6800878477306e-07, + "loss": 1.0094, + "mean_token_accuracy": 0.6961161494255066, + "num_tokens": 47083996.0, + "step": 1826 + }, + { + "epoch": 0.20063694267515925, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1470813751220703, + "learning_rate": 6.683748169838946e-07, + "loss": 1.0131, + "mean_token_accuracy": 0.7174140214920044, + "num_tokens": 47111746.0, + "step": 1827 + }, + { + "epoch": 0.2007467603777729, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.254572868347168, + "learning_rate": 6.687408491947291e-07, + "loss": 1.0884, + "mean_token_accuracy": 0.6844346523284912, + "num_tokens": 47137621.0, + "step": 1828 + }, + { + "epoch": 0.20085657808038657, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3162052631378174, + "learning_rate": 6.691068814055637e-07, + "loss": 0.9876, + "mean_token_accuracy": 0.6990835666656494, + "num_tokens": 47162400.0, + "step": 1829 + }, + { + "epoch": 0.2009663957830002, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5764153003692627, + "learning_rate": 6.694729136163983e-07, + "loss": 0.9849, + "mean_token_accuracy": 0.7024210691452026, + "num_tokens": 47186479.0, + "step": 1830 + }, + { + "epoch": 0.20107621348561389, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0164663791656494, + "learning_rate": 6.698389458272328e-07, + "loss": 1.0794, + "mean_token_accuracy": 0.6875780820846558, + "num_tokens": 47217322.0, + "step": 1831 + }, + { + "epoch": 0.20118603118822753, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.02193284034729, + "learning_rate": 6.702049780380673e-07, + "loss": 1.021, + "mean_token_accuracy": 0.6878975629806519, + "num_tokens": 47247815.0, + "step": 1832 + }, + { + "epoch": 0.2012958488908412, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.293256998062134, + "learning_rate": 6.705710102489019e-07, + "loss": 1.0136, + "mean_token_accuracy": 0.6965388059616089, + "num_tokens": 47274328.0, + "step": 1833 + }, + { + "epoch": 0.20140566659345485, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.075998544692993, + "learning_rate": 6.709370424597365e-07, + "loss": 1.0305, + "mean_token_accuracy": 0.6866246461868286, + "num_tokens": 47303590.0, + "step": 1834 + }, + { + "epoch": 0.20151548429606853, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.141003370285034, + "learning_rate": 6.71303074670571e-07, + "loss": 0.8973, + "mean_token_accuracy": 0.7298351526260376, + "num_tokens": 47331291.0, + "step": 1835 + }, + { + "epoch": 0.2016253019986822, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.2647249698638916, + "learning_rate": 6.716691068814056e-07, + "loss": 1.0089, + "mean_token_accuracy": 0.6954572200775146, + "num_tokens": 47356358.0, + "step": 1836 + }, + { + "epoch": 0.20173511970129585, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.210754156112671, + "learning_rate": 6.720351390922401e-07, + "loss": 0.9972, + "mean_token_accuracy": 0.7001628875732422, + "num_tokens": 47383864.0, + "step": 1837 + }, + { + "epoch": 0.20184493740390952, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3517496585845947, + "learning_rate": 6.724011713030747e-07, + "loss": 1.0284, + "mean_token_accuracy": 0.6918584704399109, + "num_tokens": 47408405.0, + "step": 1838 + }, + { + "epoch": 0.20195475510652316, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.421165704727173, + "learning_rate": 6.727672035139092e-07, + "loss": 0.9988, + "mean_token_accuracy": 0.7029131054878235, + "num_tokens": 47432899.0, + "step": 1839 + }, + { + "epoch": 0.20206457280913684, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.3357455730438232, + "learning_rate": 6.731332357247438e-07, + "loss": 1.1242, + "mean_token_accuracy": 0.671689510345459, + "num_tokens": 47459200.0, + "step": 1840 + }, + { + "epoch": 0.20217439051175048, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.1013095378875732, + "learning_rate": 6.734992679355784e-07, + "loss": 1.1156, + "mean_token_accuracy": 0.6737546920776367, + "num_tokens": 47489755.0, + "step": 1841 + }, + { + "epoch": 0.20228420821436416, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 1.9689453840255737, + "learning_rate": 6.738653001464129e-07, + "loss": 1.0387, + "mean_token_accuracy": 0.6884151101112366, + "num_tokens": 47521733.0, + "step": 1842 + }, + { + "epoch": 0.2023940259169778, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.5592172145843506, + "learning_rate": 6.742313323572474e-07, + "loss": 0.9539, + "mean_token_accuracy": 0.7124350666999817, + "num_tokens": 47541998.0, + "step": 1843 + }, + { + "epoch": 0.20250384361959148, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.1335573196411133, + "learning_rate": 6.74597364568082e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.687716543674469, + "num_tokens": 47569865.0, + "step": 1844 + }, + { + "epoch": 0.20261366132220515, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.2991576194763184, + "learning_rate": 6.749633967789166e-07, + "loss": 1.1165, + "mean_token_accuracy": 0.6747463941574097, + "num_tokens": 47598795.0, + "step": 1845 + }, + { + "epoch": 0.2027234790248188, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.1601650714874268, + "learning_rate": 6.753294289897511e-07, + "loss": 1.1157, + "mean_token_accuracy": 0.6718767881393433, + "num_tokens": 47627078.0, + "step": 1846 + }, + { + "epoch": 0.20283329672743247, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 1.9717864990234375, + "learning_rate": 6.756954612005857e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.6966886520385742, + "num_tokens": 47660542.0, + "step": 1847 + }, + { + "epoch": 0.20294311443004612, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.3002302646636963, + "learning_rate": 6.760614934114202e-07, + "loss": 1.0811, + "mean_token_accuracy": 0.6808932423591614, + "num_tokens": 47687250.0, + "step": 1848 + }, + { + "epoch": 0.2030529321326598, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.5499989986419678, + "learning_rate": 6.764275256222546e-07, + "loss": 0.9777, + "mean_token_accuracy": 0.6960214376449585, + "num_tokens": 47708672.0, + "step": 1849 + }, + { + "epoch": 0.20316274983527344, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.5847129821777344, + "learning_rate": 6.767935578330894e-07, + "loss": 0.9589, + "mean_token_accuracy": 0.712659478187561, + "num_tokens": 47730052.0, + "step": 1850 + }, + { + "epoch": 0.2032725675378871, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.3065497875213623, + "learning_rate": 6.771595900439239e-07, + "loss": 1.0951, + "mean_token_accuracy": 0.6740144491195679, + "num_tokens": 47757372.0, + "step": 1851 + }, + { + "epoch": 0.20338238524050076, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.2965712547302246, + "learning_rate": 6.775256222547583e-07, + "loss": 1.0536, + "mean_token_accuracy": 0.6776584982872009, + "num_tokens": 47784543.0, + "step": 1852 + }, + { + "epoch": 0.20349220294311443, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.140946626663208, + "learning_rate": 6.77891654465593e-07, + "loss": 1.0462, + "mean_token_accuracy": 0.6855717897415161, + "num_tokens": 47813373.0, + "step": 1853 + }, + { + "epoch": 0.2036020206457281, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.3040521144866943, + "learning_rate": 6.782576866764274e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.6865501403808594, + "num_tokens": 47837686.0, + "step": 1854 + }, + { + "epoch": 0.20371183834834175, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.097982883453369, + "learning_rate": 6.78623718887262e-07, + "loss": 1.0686, + "mean_token_accuracy": 0.6891909837722778, + "num_tokens": 47870281.0, + "step": 1855 + }, + { + "epoch": 0.20382165605095542, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.161992073059082, + "learning_rate": 6.789897510980966e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6745541095733643, + "num_tokens": 47899080.0, + "step": 1856 + }, + { + "epoch": 0.20393147375356907, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.2789084911346436, + "learning_rate": 6.793557833089311e-07, + "loss": 1.0644, + "mean_token_accuracy": 0.6864356994628906, + "num_tokens": 47923965.0, + "step": 1857 + }, + { + "epoch": 0.20404129145618274, + "ewc_loss": 1.0967254638671875e-05, + "grad_norm": 2.5189459323883057, + "learning_rate": 6.797218155197657e-07, + "loss": 0.9933, + "mean_token_accuracy": 0.7005336880683899, + "num_tokens": 47945835.0, + "step": 1858 + }, + { + "epoch": 0.2041511091587964, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.400879383087158, + "learning_rate": 6.800878477306002e-07, + "loss": 1.0316, + "mean_token_accuracy": 0.6891705989837646, + "num_tokens": 47971101.0, + "step": 1859 + }, + { + "epoch": 0.20426092686141006, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.397984266281128, + "learning_rate": 6.804538799414347e-07, + "loss": 1.0897, + "mean_token_accuracy": 0.6922075748443604, + "num_tokens": 47995297.0, + "step": 1860 + }, + { + "epoch": 0.2043707445640237, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.134220838546753, + "learning_rate": 6.808199121522694e-07, + "loss": 1.0755, + "mean_token_accuracy": 0.6781335473060608, + "num_tokens": 48028033.0, + "step": 1861 + }, + { + "epoch": 0.20448056226663738, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.4171624183654785, + "learning_rate": 6.811859443631039e-07, + "loss": 1.0425, + "mean_token_accuracy": 0.686054527759552, + "num_tokens": 48050471.0, + "step": 1862 + }, + { + "epoch": 0.20459037996925103, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.439866304397583, + "learning_rate": 6.815519765739384e-07, + "loss": 1.035, + "mean_token_accuracy": 0.690883994102478, + "num_tokens": 48073222.0, + "step": 1863 + }, + { + "epoch": 0.2047001976718647, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.318333387374878, + "learning_rate": 6.81918008784773e-07, + "loss": 1.0252, + "mean_token_accuracy": 0.6928741335868835, + "num_tokens": 48097610.0, + "step": 1864 + }, + { + "epoch": 0.20481001537447838, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.554713726043701, + "learning_rate": 6.822840409956075e-07, + "loss": 0.9995, + "mean_token_accuracy": 0.7011297345161438, + "num_tokens": 48119076.0, + "step": 1865 + }, + { + "epoch": 0.20491983307709202, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.219907760620117, + "learning_rate": 6.826500732064421e-07, + "loss": 0.9891, + "mean_token_accuracy": 0.6996839642524719, + "num_tokens": 48145067.0, + "step": 1866 + }, + { + "epoch": 0.2050296507797057, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.334073305130005, + "learning_rate": 6.830161054172767e-07, + "loss": 1.0537, + "mean_token_accuracy": 0.6776328086853027, + "num_tokens": 48169280.0, + "step": 1867 + }, + { + "epoch": 0.20513946848231934, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.0613439083099365, + "learning_rate": 6.833821376281112e-07, + "loss": 0.9639, + "mean_token_accuracy": 0.7039039134979248, + "num_tokens": 48201385.0, + "step": 1868 + }, + { + "epoch": 0.20524928618493302, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.3847622871398926, + "learning_rate": 6.837481698389457e-07, + "loss": 0.9988, + "mean_token_accuracy": 0.6962894797325134, + "num_tokens": 48227893.0, + "step": 1869 + }, + { + "epoch": 0.20535910388754666, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.275651454925537, + "learning_rate": 6.841142020497803e-07, + "loss": 1.06, + "mean_token_accuracy": 0.6861473321914673, + "num_tokens": 48255573.0, + "step": 1870 + }, + { + "epoch": 0.20546892159016034, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.4374184608459473, + "learning_rate": 6.844802342606148e-07, + "loss": 0.9404, + "mean_token_accuracy": 0.720419704914093, + "num_tokens": 48276990.0, + "step": 1871 + }, + { + "epoch": 0.20557873929277398, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.4678635597229004, + "learning_rate": 6.848462664714495e-07, + "loss": 1.0309, + "mean_token_accuracy": 0.6940063834190369, + "num_tokens": 48298559.0, + "step": 1872 + }, + { + "epoch": 0.20568855699538766, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.1362078189849854, + "learning_rate": 6.85212298682284e-07, + "loss": 1.0752, + "mean_token_accuracy": 0.673508882522583, + "num_tokens": 48325901.0, + "step": 1873 + }, + { + "epoch": 0.20579837469800133, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.528733015060425, + "learning_rate": 6.855783308931185e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6848270893096924, + "num_tokens": 48347441.0, + "step": 1874 + }, + { + "epoch": 0.20590819240061498, + "ewc_loss": 1.1205673217773438e-05, + "grad_norm": 3.0041871070861816, + "learning_rate": 6.859443631039531e-07, + "loss": 1.0297, + "mean_token_accuracy": 0.6941656470298767, + "num_tokens": 48365276.0, + "step": 1875 + }, + { + "epoch": 0.20601801010322865, + "ewc_loss": 1.1205673217773438e-05, + "grad_norm": 2.625128746032715, + "learning_rate": 6.863103953147876e-07, + "loss": 0.9373, + "mean_token_accuracy": 0.7164907455444336, + "num_tokens": 48385953.0, + "step": 1876 + }, + { + "epoch": 0.2061278278058423, + "ewc_loss": 1.1205673217773438e-05, + "grad_norm": 2.3634092807769775, + "learning_rate": 6.866764275256222e-07, + "loss": 0.9881, + "mean_token_accuracy": 0.7026623487472534, + "num_tokens": 48410901.0, + "step": 1877 + }, + { + "epoch": 0.20623764550845597, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.3020076751708984, + "learning_rate": 6.870424597364568e-07, + "loss": 1.0816, + "mean_token_accuracy": 0.6812452077865601, + "num_tokens": 48438408.0, + "step": 1878 + }, + { + "epoch": 0.20634746321106961, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.674245834350586, + "learning_rate": 6.874084919472913e-07, + "loss": 0.9987, + "mean_token_accuracy": 0.7018103003501892, + "num_tokens": 48457518.0, + "step": 1879 + }, + { + "epoch": 0.2064572809136833, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.157545566558838, + "learning_rate": 6.877745241581258e-07, + "loss": 0.9857, + "mean_token_accuracy": 0.7091972827911377, + "num_tokens": 48486392.0, + "step": 1880 + }, + { + "epoch": 0.20656709861629693, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.099547863006592, + "learning_rate": 6.881405563689604e-07, + "loss": 1.1218, + "mean_token_accuracy": 0.677424430847168, + "num_tokens": 48516058.0, + "step": 1881 + }, + { + "epoch": 0.2066769163189106, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.118990421295166, + "learning_rate": 6.88506588579795e-07, + "loss": 1.0735, + "mean_token_accuracy": 0.6780732870101929, + "num_tokens": 48546658.0, + "step": 1882 + }, + { + "epoch": 0.20678673402152428, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4835400581359863, + "learning_rate": 6.888726207906295e-07, + "loss": 1.0573, + "mean_token_accuracy": 0.6809371113777161, + "num_tokens": 48569941.0, + "step": 1883 + }, + { + "epoch": 0.20689655172413793, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.621164083480835, + "learning_rate": 6.892386530014641e-07, + "loss": 0.9988, + "mean_token_accuracy": 0.6989400386810303, + "num_tokens": 48590774.0, + "step": 1884 + }, + { + "epoch": 0.2070063694267516, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.1617841720581055, + "learning_rate": 6.896046852122986e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6807833909988403, + "num_tokens": 48617652.0, + "step": 1885 + }, + { + "epoch": 0.20711618712936525, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2641425132751465, + "learning_rate": 6.899707174231331e-07, + "loss": 1.0056, + "mean_token_accuracy": 0.6908625960350037, + "num_tokens": 48642381.0, + "step": 1886 + }, + { + "epoch": 0.20722600483197892, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.0919926166534424, + "learning_rate": 6.903367496339677e-07, + "loss": 0.9998, + "mean_token_accuracy": 0.7016515731811523, + "num_tokens": 48671585.0, + "step": 1887 + }, + { + "epoch": 0.20733582253459257, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.060189962387085, + "learning_rate": 6.907027818448023e-07, + "loss": 1.068, + "mean_token_accuracy": 0.6795343160629272, + "num_tokens": 48703218.0, + "step": 1888 + }, + { + "epoch": 0.20744564023720624, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.1429197788238525, + "learning_rate": 6.910688140556369e-07, + "loss": 1.0695, + "mean_token_accuracy": 0.6795207858085632, + "num_tokens": 48733166.0, + "step": 1889 + }, + { + "epoch": 0.2075554579398199, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3619062900543213, + "learning_rate": 6.914348462664714e-07, + "loss": 1.0256, + "mean_token_accuracy": 0.6950827836990356, + "num_tokens": 48756462.0, + "step": 1890 + }, + { + "epoch": 0.20766527564243356, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.1583778858184814, + "learning_rate": 6.918008784773059e-07, + "loss": 1.0811, + "mean_token_accuracy": 0.689670979976654, + "num_tokens": 48785002.0, + "step": 1891 + }, + { + "epoch": 0.20777509334504723, + "ewc_loss": 1.150369644165039e-05, + "grad_norm": 2.218623638153076, + "learning_rate": 6.921669106881405e-07, + "loss": 0.973, + "mean_token_accuracy": 0.709272027015686, + "num_tokens": 48810763.0, + "step": 1892 + }, + { + "epoch": 0.20788491104766088, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3660597801208496, + "learning_rate": 6.925329428989751e-07, + "loss": 1.0009, + "mean_token_accuracy": 0.696772038936615, + "num_tokens": 48835009.0, + "step": 1893 + }, + { + "epoch": 0.20799472875027455, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.404910087585449, + "learning_rate": 6.928989751098096e-07, + "loss": 0.9977, + "mean_token_accuracy": 0.6982367038726807, + "num_tokens": 48858156.0, + "step": 1894 + }, + { + "epoch": 0.2081045464528882, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.477144479751587, + "learning_rate": 6.932650073206442e-07, + "loss": 1.064, + "mean_token_accuracy": 0.6810194849967957, + "num_tokens": 48882263.0, + "step": 1895 + }, + { + "epoch": 0.20821436415550187, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.487525224685669, + "learning_rate": 6.936310395314787e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6871948838233948, + "num_tokens": 48905461.0, + "step": 1896 + }, + { + "epoch": 0.20832418185811552, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.197364330291748, + "learning_rate": 6.939970717423132e-07, + "loss": 1.029, + "mean_token_accuracy": 0.690234899520874, + "num_tokens": 48931619.0, + "step": 1897 + }, + { + "epoch": 0.2084339995607292, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.298454523086548, + "learning_rate": 6.943631039531479e-07, + "loss": 1.0661, + "mean_token_accuracy": 0.6830980777740479, + "num_tokens": 48957444.0, + "step": 1898 + }, + { + "epoch": 0.20854381726334284, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.1410861015319824, + "learning_rate": 6.947291361639824e-07, + "loss": 0.967, + "mean_token_accuracy": 0.7029346227645874, + "num_tokens": 48986415.0, + "step": 1899 + }, + { + "epoch": 0.2086536349659565, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.399728536605835, + "learning_rate": 6.950951683748169e-07, + "loss": 0.997, + "mean_token_accuracy": 0.6998739838600159, + "num_tokens": 49010061.0, + "step": 1900 + }, + { + "epoch": 0.20876345266857016, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.505063772201538, + "learning_rate": 6.954612005856515e-07, + "loss": 0.9542, + "mean_token_accuracy": 0.713659405708313, + "num_tokens": 49033376.0, + "step": 1901 + }, + { + "epoch": 0.20887327037118383, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.377967357635498, + "learning_rate": 6.95827232796486e-07, + "loss": 1.0196, + "mean_token_accuracy": 0.6940273642539978, + "num_tokens": 49057563.0, + "step": 1902 + }, + { + "epoch": 0.2089830880737975, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.255167007446289, + "learning_rate": 6.961932650073206e-07, + "loss": 1.1002, + "mean_token_accuracy": 0.6691756248474121, + "num_tokens": 49085030.0, + "step": 1903 + }, + { + "epoch": 0.20909290577641115, + "ewc_loss": 1.1682510375976562e-05, + "grad_norm": 2.512727737426758, + "learning_rate": 6.965592972181552e-07, + "loss": 0.9371, + "mean_token_accuracy": 0.710719883441925, + "num_tokens": 49104353.0, + "step": 1904 + }, + { + "epoch": 0.20920272347902483, + "ewc_loss": 1.1682510375976562e-05, + "grad_norm": 2.2070767879486084, + "learning_rate": 6.969253294289897e-07, + "loss": 1.0417, + "mean_token_accuracy": 0.6953303217887878, + "num_tokens": 49134413.0, + "step": 1905 + }, + { + "epoch": 0.20931254118163847, + "ewc_loss": 1.1682510375976562e-05, + "grad_norm": 2.1815896034240723, + "learning_rate": 6.972913616398243e-07, + "loss": 1.0401, + "mean_token_accuracy": 0.6922311186790466, + "num_tokens": 49164137.0, + "step": 1906 + }, + { + "epoch": 0.20942235888425215, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1432807445526123, + "learning_rate": 6.976573938506588e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.6885197162628174, + "num_tokens": 49192401.0, + "step": 1907 + }, + { + "epoch": 0.2095321765868658, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2875397205352783, + "learning_rate": 6.980234260614933e-07, + "loss": 1.0254, + "mean_token_accuracy": 0.6908442378044128, + "num_tokens": 49218341.0, + "step": 1908 + }, + { + "epoch": 0.20964199428947947, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.6117920875549316, + "learning_rate": 6.98389458272328e-07, + "loss": 1.0105, + "mean_token_accuracy": 0.6987258195877075, + "num_tokens": 49239163.0, + "step": 1909 + }, + { + "epoch": 0.2097518119920931, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.151987075805664, + "learning_rate": 6.987554904831625e-07, + "loss": 1.0341, + "mean_token_accuracy": 0.6926530599594116, + "num_tokens": 49268705.0, + "step": 1910 + }, + { + "epoch": 0.20986162969470679, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.415116548538208, + "learning_rate": 6.99121522693997e-07, + "loss": 0.9618, + "mean_token_accuracy": 0.7090339660644531, + "num_tokens": 49293437.0, + "step": 1911 + }, + { + "epoch": 0.20997144739732046, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3678696155548096, + "learning_rate": 6.994875549048316e-07, + "loss": 1.0954, + "mean_token_accuracy": 0.6694691777229309, + "num_tokens": 49317640.0, + "step": 1912 + }, + { + "epoch": 0.2100812650999341, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.0603384971618652, + "learning_rate": 6.998535871156661e-07, + "loss": 1.037, + "mean_token_accuracy": 0.6917091608047485, + "num_tokens": 49348204.0, + "step": 1913 + }, + { + "epoch": 0.21019108280254778, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2460708618164062, + "learning_rate": 7.002196193265007e-07, + "loss": 1.0224, + "mean_token_accuracy": 0.6993314027786255, + "num_tokens": 49376892.0, + "step": 1914 + }, + { + "epoch": 0.21030090050516143, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1428115367889404, + "learning_rate": 7.005856515373353e-07, + "loss": 0.9951, + "mean_token_accuracy": 0.6914651393890381, + "num_tokens": 49405788.0, + "step": 1915 + }, + { + "epoch": 0.2104107182077751, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4061717987060547, + "learning_rate": 7.009516837481698e-07, + "loss": 1.0529, + "mean_token_accuracy": 0.6851421594619751, + "num_tokens": 49428328.0, + "step": 1916 + }, + { + "epoch": 0.21052053591038875, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.1362311840057373, + "learning_rate": 7.013177159590043e-07, + "loss": 1.0729, + "mean_token_accuracy": 0.6910926699638367, + "num_tokens": 49456756.0, + "step": 1917 + }, + { + "epoch": 0.21063035361300242, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.412820339202881, + "learning_rate": 7.016837481698389e-07, + "loss": 1.0242, + "mean_token_accuracy": 0.6938643455505371, + "num_tokens": 49481855.0, + "step": 1918 + }, + { + "epoch": 0.21074017131561606, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.3512628078460693, + "learning_rate": 7.020497803806734e-07, + "loss": 1.0541, + "mean_token_accuracy": 0.6864449977874756, + "num_tokens": 49507527.0, + "step": 1919 + }, + { + "epoch": 0.21084998901822974, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.1869559288024902, + "learning_rate": 7.024158125915081e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.6783586144447327, + "num_tokens": 49536120.0, + "step": 1920 + }, + { + "epoch": 0.2109598067208434, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5244667530059814, + "learning_rate": 7.027818448023426e-07, + "loss": 0.9945, + "mean_token_accuracy": 0.6996433734893799, + "num_tokens": 49557633.0, + "step": 1921 + }, + { + "epoch": 0.21106962442345706, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.3839688301086426, + "learning_rate": 7.031478770131771e-07, + "loss": 0.9901, + "mean_token_accuracy": 0.7042323350906372, + "num_tokens": 49581184.0, + "step": 1922 + }, + { + "epoch": 0.21117944212607073, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.1630616188049316, + "learning_rate": 7.035139092240117e-07, + "loss": 1.0547, + "mean_token_accuracy": 0.6860103607177734, + "num_tokens": 49612774.0, + "step": 1923 + }, + { + "epoch": 0.21128925982868438, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.6528842449188232, + "learning_rate": 7.038799414348462e-07, + "loss": 0.9905, + "mean_token_accuracy": 0.6987910270690918, + "num_tokens": 49632537.0, + "step": 1924 + }, + { + "epoch": 0.21139907753129805, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.0881285667419434, + "learning_rate": 7.042459736456808e-07, + "loss": 0.9424, + "mean_token_accuracy": 0.7219970226287842, + "num_tokens": 49661533.0, + "step": 1925 + }, + { + "epoch": 0.2115088952339117, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.081273078918457, + "learning_rate": 7.046120058565154e-07, + "loss": 0.992, + "mean_token_accuracy": 0.6995818614959717, + "num_tokens": 49691168.0, + "step": 1926 + }, + { + "epoch": 0.21161871293652537, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1967997550964355, + "learning_rate": 7.049780380673499e-07, + "loss": 1.1258, + "mean_token_accuracy": 0.6710789203643799, + "num_tokens": 49718846.0, + "step": 1927 + }, + { + "epoch": 0.21172853063913902, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.6543257236480713, + "learning_rate": 7.053440702781844e-07, + "loss": 1.0227, + "mean_token_accuracy": 0.6958920955657959, + "num_tokens": 49738425.0, + "step": 1928 + }, + { + "epoch": 0.2118383483417527, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.244377851486206, + "learning_rate": 7.05710102489019e-07, + "loss": 0.9691, + "mean_token_accuracy": 0.7078475952148438, + "num_tokens": 49762843.0, + "step": 1929 + }, + { + "epoch": 0.21194816604436637, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.190112829208374, + "learning_rate": 7.060761346998536e-07, + "loss": 0.9169, + "mean_token_accuracy": 0.7263680696487427, + "num_tokens": 49787535.0, + "step": 1930 + }, + { + "epoch": 0.21205798374698, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5257203578948975, + "learning_rate": 7.064421669106881e-07, + "loss": 0.9504, + "mean_token_accuracy": 0.7179152369499207, + "num_tokens": 49808954.0, + "step": 1931 + }, + { + "epoch": 0.21216780144959368, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.246549606323242, + "learning_rate": 7.068081991215227e-07, + "loss": 0.9599, + "mean_token_accuracy": 0.7134448885917664, + "num_tokens": 49833932.0, + "step": 1932 + }, + { + "epoch": 0.21227761915220733, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.266672134399414, + "learning_rate": 7.071742313323572e-07, + "loss": 1.0464, + "mean_token_accuracy": 0.6885989904403687, + "num_tokens": 49861910.0, + "step": 1933 + }, + { + "epoch": 0.212387436854821, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4569101333618164, + "learning_rate": 7.075402635431917e-07, + "loss": 1.0507, + "mean_token_accuracy": 0.6881820559501648, + "num_tokens": 49885140.0, + "step": 1934 + }, + { + "epoch": 0.21249725455743465, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.2162420749664307, + "learning_rate": 7.079062957540263e-07, + "loss": 0.9454, + "mean_token_accuracy": 0.7129776477813721, + "num_tokens": 49909653.0, + "step": 1935 + }, + { + "epoch": 0.21260707226004832, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.267857074737549, + "learning_rate": 7.082723279648609e-07, + "loss": 1.0563, + "mean_token_accuracy": 0.6869904398918152, + "num_tokens": 49936584.0, + "step": 1936 + }, + { + "epoch": 0.21271688996266197, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.572021484375, + "learning_rate": 7.086383601756955e-07, + "loss": 1.0208, + "mean_token_accuracy": 0.690048336982727, + "num_tokens": 49956702.0, + "step": 1937 + }, + { + "epoch": 0.21282670766527564, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4864306449890137, + "learning_rate": 7.0900439238653e-07, + "loss": 0.9733, + "mean_token_accuracy": 0.6953601241111755, + "num_tokens": 49977896.0, + "step": 1938 + }, + { + "epoch": 0.2129365253678893, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.2050344944000244, + "learning_rate": 7.093704245973645e-07, + "loss": 0.961, + "mean_token_accuracy": 0.7117820382118225, + "num_tokens": 50003208.0, + "step": 1939 + }, + { + "epoch": 0.21304634307050296, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.487405776977539, + "learning_rate": 7.097364568081991e-07, + "loss": 1.0253, + "mean_token_accuracy": 0.692361056804657, + "num_tokens": 50024940.0, + "step": 1940 + }, + { + "epoch": 0.21315616077311664, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1069722175598145, + "learning_rate": 7.101024890190337e-07, + "loss": 0.9156, + "mean_token_accuracy": 0.7228578329086304, + "num_tokens": 50051034.0, + "step": 1941 + }, + { + "epoch": 0.21326597847573028, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.2755870819091797, + "learning_rate": 7.104685212298682e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.6871480941772461, + "num_tokens": 50077175.0, + "step": 1942 + }, + { + "epoch": 0.21337579617834396, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 1.9225027561187744, + "learning_rate": 7.108345534407028e-07, + "loss": 1.1611, + "mean_token_accuracy": 0.6541171073913574, + "num_tokens": 50115357.0, + "step": 1943 + }, + { + "epoch": 0.2134856138809576, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.357184886932373, + "learning_rate": 7.112005856515373e-07, + "loss": 0.9509, + "mean_token_accuracy": 0.7122793197631836, + "num_tokens": 50143631.0, + "step": 1944 + }, + { + "epoch": 0.21359543158357128, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.4084582328796387, + "learning_rate": 7.115666178623718e-07, + "loss": 0.9851, + "mean_token_accuracy": 0.7072079181671143, + "num_tokens": 50167265.0, + "step": 1945 + }, + { + "epoch": 0.21370524928618492, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.443049192428589, + "learning_rate": 7.119326500732065e-07, + "loss": 1.1164, + "mean_token_accuracy": 0.6684746742248535, + "num_tokens": 50191169.0, + "step": 1946 + }, + { + "epoch": 0.2138150669887986, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.3587238788604736, + "learning_rate": 7.12298682284041e-07, + "loss": 1.0345, + "mean_token_accuracy": 0.6887713670730591, + "num_tokens": 50218920.0, + "step": 1947 + }, + { + "epoch": 0.21392488469141224, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.41884446144104, + "learning_rate": 7.126647144948755e-07, + "loss": 1.1292, + "mean_token_accuracy": 0.6660308837890625, + "num_tokens": 50244790.0, + "step": 1948 + }, + { + "epoch": 0.21403470239402592, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.249589681625366, + "learning_rate": 7.130307467057101e-07, + "loss": 0.9954, + "mean_token_accuracy": 0.7010967135429382, + "num_tokens": 50270316.0, + "step": 1949 + }, + { + "epoch": 0.2141445200966396, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.2343361377716064, + "learning_rate": 7.133967789165446e-07, + "loss": 1.0082, + "mean_token_accuracy": 0.6978472471237183, + "num_tokens": 50296368.0, + "step": 1950 + }, + { + "epoch": 0.21425433779925324, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.5158538818359375, + "learning_rate": 7.137628111273791e-07, + "loss": 0.9893, + "mean_token_accuracy": 0.6948965787887573, + "num_tokens": 50317194.0, + "step": 1951 + }, + { + "epoch": 0.2143641555018669, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.4115753173828125, + "learning_rate": 7.141288433382138e-07, + "loss": 0.9559, + "mean_token_accuracy": 0.7107090950012207, + "num_tokens": 50338475.0, + "step": 1952 + }, + { + "epoch": 0.21447397320448056, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.313432216644287, + "learning_rate": 7.144948755490483e-07, + "loss": 1.0685, + "mean_token_accuracy": 0.6762059330940247, + "num_tokens": 50363753.0, + "step": 1953 + }, + { + "epoch": 0.21458379090709423, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.341912031173706, + "learning_rate": 7.148609077598829e-07, + "loss": 0.9676, + "mean_token_accuracy": 0.7051279544830322, + "num_tokens": 50387871.0, + "step": 1954 + }, + { + "epoch": 0.21469360860970788, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.511958122253418, + "learning_rate": 7.152269399707174e-07, + "loss": 1.0114, + "mean_token_accuracy": 0.6961401700973511, + "num_tokens": 50408307.0, + "step": 1955 + }, + { + "epoch": 0.21480342631232155, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.3042707443237305, + "learning_rate": 7.155929721815519e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6898884773254395, + "num_tokens": 50434392.0, + "step": 1956 + }, + { + "epoch": 0.2149132440149352, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.0768299102783203, + "learning_rate": 7.159590043923866e-07, + "loss": 1.036, + "mean_token_accuracy": 0.6876595616340637, + "num_tokens": 50464382.0, + "step": 1957 + }, + { + "epoch": 0.21502306171754887, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.6355531215667725, + "learning_rate": 7.163250366032211e-07, + "loss": 1.1379, + "mean_token_accuracy": 0.6586993336677551, + "num_tokens": 50486233.0, + "step": 1958 + }, + { + "epoch": 0.21513287942016254, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.163440465927124, + "learning_rate": 7.166910688140556e-07, + "loss": 1.0812, + "mean_token_accuracy": 0.6728597283363342, + "num_tokens": 50516822.0, + "step": 1959 + }, + { + "epoch": 0.2152426971227762, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.4425442218780518, + "learning_rate": 7.170571010248902e-07, + "loss": 1.1092, + "mean_token_accuracy": 0.6743327975273132, + "num_tokens": 50541270.0, + "step": 1960 + }, + { + "epoch": 0.21535251482538986, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.2119433879852295, + "learning_rate": 7.174231332357247e-07, + "loss": 0.9801, + "mean_token_accuracy": 0.7003705501556396, + "num_tokens": 50566736.0, + "step": 1961 + }, + { + "epoch": 0.2154623325280035, + "ewc_loss": 1.2159347534179688e-05, + "grad_norm": 2.266648054122925, + "learning_rate": 7.177891654465593e-07, + "loss": 1.0182, + "mean_token_accuracy": 0.6919738054275513, + "num_tokens": 50591212.0, + "step": 1962 + }, + { + "epoch": 0.21557215023061718, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.532707452774048, + "learning_rate": 7.181551976573939e-07, + "loss": 0.9431, + "mean_token_accuracy": 0.7118110656738281, + "num_tokens": 50613334.0, + "step": 1963 + }, + { + "epoch": 0.21568196793323083, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.040062665939331, + "learning_rate": 7.185212298682284e-07, + "loss": 0.95, + "mean_token_accuracy": 0.7105496525764465, + "num_tokens": 50642793.0, + "step": 1964 + }, + { + "epoch": 0.2157917856358445, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2620649337768555, + "learning_rate": 7.188872620790629e-07, + "loss": 1.1189, + "mean_token_accuracy": 0.6792392730712891, + "num_tokens": 50670184.0, + "step": 1965 + }, + { + "epoch": 0.21590160333845815, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.839022636413574, + "learning_rate": 7.192532942898975e-07, + "loss": 0.9426, + "mean_token_accuracy": 0.7066987752914429, + "num_tokens": 50688140.0, + "step": 1966 + }, + { + "epoch": 0.21601142104107182, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3179638385772705, + "learning_rate": 7.19619326500732e-07, + "loss": 1.0451, + "mean_token_accuracy": 0.6841541528701782, + "num_tokens": 50712549.0, + "step": 1967 + }, + { + "epoch": 0.2161212387436855, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.409820795059204, + "learning_rate": 7.199853587115667e-07, + "loss": 1.1316, + "mean_token_accuracy": 0.6652233004570007, + "num_tokens": 50738520.0, + "step": 1968 + }, + { + "epoch": 0.21623105644629914, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2264461517333984, + "learning_rate": 7.203513909224012e-07, + "loss": 0.969, + "mean_token_accuracy": 0.7150729894638062, + "num_tokens": 50763757.0, + "step": 1969 + }, + { + "epoch": 0.21634087414891282, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.0281732082366943, + "learning_rate": 7.207174231332357e-07, + "loss": 1.0268, + "mean_token_accuracy": 0.7036532759666443, + "num_tokens": 50793832.0, + "step": 1970 + }, + { + "epoch": 0.21645069185152646, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5749971866607666, + "learning_rate": 7.210834553440703e-07, + "loss": 0.9394, + "mean_token_accuracy": 0.7098060846328735, + "num_tokens": 50814416.0, + "step": 1971 + }, + { + "epoch": 0.21656050955414013, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.7386651039123535, + "learning_rate": 7.214494875549048e-07, + "loss": 0.9443, + "mean_token_accuracy": 0.7065436840057373, + "num_tokens": 50833389.0, + "step": 1972 + }, + { + "epoch": 0.21667032725675378, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1788625717163086, + "learning_rate": 7.218155197657394e-07, + "loss": 1.1337, + "mean_token_accuracy": 0.6713329553604126, + "num_tokens": 50863979.0, + "step": 1973 + }, + { + "epoch": 0.21678014495936745, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.422914981842041, + "learning_rate": 7.22181551976574e-07, + "loss": 0.9588, + "mean_token_accuracy": 0.7048283815383911, + "num_tokens": 50886208.0, + "step": 1974 + }, + { + "epoch": 0.2168899626619811, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.471547842025757, + "learning_rate": 7.225475841874085e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6859918236732483, + "num_tokens": 50908463.0, + "step": 1975 + }, + { + "epoch": 0.21699978036459477, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4838366508483887, + "learning_rate": 7.22913616398243e-07, + "loss": 0.9835, + "mean_token_accuracy": 0.7098162174224854, + "num_tokens": 50931728.0, + "step": 1976 + }, + { + "epoch": 0.21710959806720842, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2163784503936768, + "learning_rate": 7.232796486090776e-07, + "loss": 1.0602, + "mean_token_accuracy": 0.6824284791946411, + "num_tokens": 50956836.0, + "step": 1977 + }, + { + "epoch": 0.2172194157698221, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.117570638656616, + "learning_rate": 7.236456808199122e-07, + "loss": 1.1407, + "mean_token_accuracy": 0.661088228225708, + "num_tokens": 50988874.0, + "step": 1978 + }, + { + "epoch": 0.21732923347243577, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.259669065475464, + "learning_rate": 7.240117130307467e-07, + "loss": 1.1107, + "mean_token_accuracy": 0.6842602491378784, + "num_tokens": 51017825.0, + "step": 1979 + }, + { + "epoch": 0.2174390511750494, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1284143924713135, + "learning_rate": 7.243777452415813e-07, + "loss": 0.9936, + "mean_token_accuracy": 0.7002316117286682, + "num_tokens": 51045955.0, + "step": 1980 + }, + { + "epoch": 0.2175488688776631, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4912784099578857, + "learning_rate": 7.247437774524158e-07, + "loss": 1.0392, + "mean_token_accuracy": 0.6799615025520325, + "num_tokens": 51069330.0, + "step": 1981 + }, + { + "epoch": 0.21765868658027673, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.579540491104126, + "learning_rate": 7.251098096632503e-07, + "loss": 1.0692, + "mean_token_accuracy": 0.6931359767913818, + "num_tokens": 51092178.0, + "step": 1982 + }, + { + "epoch": 0.2177685042828904, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3511030673980713, + "learning_rate": 7.254758418740849e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.6940426826477051, + "num_tokens": 51115554.0, + "step": 1983 + }, + { + "epoch": 0.21787832198550405, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.467642307281494, + "learning_rate": 7.258418740849195e-07, + "loss": 0.9799, + "mean_token_accuracy": 0.7039011716842651, + "num_tokens": 51135508.0, + "step": 1984 + }, + { + "epoch": 0.21798813968811773, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.210338592529297, + "learning_rate": 7.262079062957541e-07, + "loss": 0.9799, + "mean_token_accuracy": 0.7065954804420471, + "num_tokens": 51159938.0, + "step": 1985 + }, + { + "epoch": 0.21809795739073137, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.115558385848999, + "learning_rate": 7.265739385065886e-07, + "loss": 1.1244, + "mean_token_accuracy": 0.6689644455909729, + "num_tokens": 51189162.0, + "step": 1986 + }, + { + "epoch": 0.21820777509334505, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.353722333908081, + "learning_rate": 7.269399707174231e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6917288303375244, + "num_tokens": 51213141.0, + "step": 1987 + }, + { + "epoch": 0.21831759279595872, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.135457992553711, + "learning_rate": 7.273060029282577e-07, + "loss": 1.0373, + "mean_token_accuracy": 0.6881684064865112, + "num_tokens": 51243740.0, + "step": 1988 + }, + { + "epoch": 0.21842741049857237, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4823765754699707, + "learning_rate": 7.276720351390923e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6990185976028442, + "num_tokens": 51267656.0, + "step": 1989 + }, + { + "epoch": 0.21853722820118604, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3310372829437256, + "learning_rate": 7.280380673499268e-07, + "loss": 0.9446, + "mean_token_accuracy": 0.7126507759094238, + "num_tokens": 51291884.0, + "step": 1990 + }, + { + "epoch": 0.21864704590379969, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.462886095046997, + "learning_rate": 7.284040995607614e-07, + "loss": 1.0484, + "mean_token_accuracy": 0.6846588253974915, + "num_tokens": 51313957.0, + "step": 1991 + }, + { + "epoch": 0.21875686360641336, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.6887876987457275, + "learning_rate": 7.287701317715959e-07, + "loss": 0.9868, + "mean_token_accuracy": 0.7010301351547241, + "num_tokens": 51333146.0, + "step": 1992 + }, + { + "epoch": 0.218866681309027, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.552978277206421, + "learning_rate": 7.291361639824304e-07, + "loss": 1.0726, + "mean_token_accuracy": 0.6825467944145203, + "num_tokens": 51354676.0, + "step": 1993 + }, + { + "epoch": 0.21897649901164068, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.371634006500244, + "learning_rate": 7.295021961932651e-07, + "loss": 0.999, + "mean_token_accuracy": 0.7006664872169495, + "num_tokens": 51377314.0, + "step": 1994 + }, + { + "epoch": 0.21908631671425433, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.3189542293548584, + "learning_rate": 7.298682284040996e-07, + "loss": 1.0538, + "mean_token_accuracy": 0.686397135257721, + "num_tokens": 51404483.0, + "step": 1995 + }, + { + "epoch": 0.219196134416868, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.3784477710723877, + "learning_rate": 7.30234260614934e-07, + "loss": 1.0041, + "mean_token_accuracy": 0.6967206597328186, + "num_tokens": 51427270.0, + "step": 1996 + }, + { + "epoch": 0.21930595211948167, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.4922592639923096, + "learning_rate": 7.306002928257687e-07, + "loss": 1.069, + "mean_token_accuracy": 0.6854802966117859, + "num_tokens": 51451033.0, + "step": 1997 + }, + { + "epoch": 0.21941576982209532, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.2689833641052246, + "learning_rate": 7.309663250366031e-07, + "loss": 0.9918, + "mean_token_accuracy": 0.7028844356536865, + "num_tokens": 51477625.0, + "step": 1998 + }, + { + "epoch": 0.219525587524709, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.5666768550872803, + "learning_rate": 7.313323572474376e-07, + "loss": 1.0435, + "mean_token_accuracy": 0.6875689029693604, + "num_tokens": 51499313.0, + "step": 1999 + }, + { + "epoch": 0.21963540522732264, + "ewc_loss": 1.2516975402832031e-05, + "grad_norm": 2.439739942550659, + "learning_rate": 7.316983894582724e-07, + "loss": 1.0078, + "mean_token_accuracy": 0.7049583196640015, + "num_tokens": 51523584.0, + "step": 2000 + }, + { + "epoch": 0.2197452229299363, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.5105676651000977, + "learning_rate": 7.320644216691068e-07, + "loss": 1.0682, + "mean_token_accuracy": 0.6808122396469116, + "num_tokens": 51544804.0, + "step": 2001 + }, + { + "epoch": 0.21985504063254996, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3576931953430176, + "learning_rate": 7.324304538799415e-07, + "loss": 0.9755, + "mean_token_accuracy": 0.7032487392425537, + "num_tokens": 51568871.0, + "step": 2002 + }, + { + "epoch": 0.21996485833516363, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2062735557556152, + "learning_rate": 7.32796486090776e-07, + "loss": 1.0176, + "mean_token_accuracy": 0.7048985958099365, + "num_tokens": 51596345.0, + "step": 2003 + }, + { + "epoch": 0.22007467603777728, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1120076179504395, + "learning_rate": 7.331625183016104e-07, + "loss": 1.0619, + "mean_token_accuracy": 0.6861572861671448, + "num_tokens": 51624644.0, + "step": 2004 + }, + { + "epoch": 0.22018449374039095, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.332824468612671, + "learning_rate": 7.335285505124452e-07, + "loss": 0.9303, + "mean_token_accuracy": 0.7162779569625854, + "num_tokens": 51647338.0, + "step": 2005 + }, + { + "epoch": 0.22029431144300463, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4378774166107178, + "learning_rate": 7.338945827232796e-07, + "loss": 1.0398, + "mean_token_accuracy": 0.690812349319458, + "num_tokens": 51670588.0, + "step": 2006 + }, + { + "epoch": 0.22040412914561827, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.258114814758301, + "learning_rate": 7.342606149341141e-07, + "loss": 1.0003, + "mean_token_accuracy": 0.7095876336097717, + "num_tokens": 51694915.0, + "step": 2007 + }, + { + "epoch": 0.22051394684823195, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1679253578186035, + "learning_rate": 7.346266471449487e-07, + "loss": 1.1023, + "mean_token_accuracy": 0.6702907085418701, + "num_tokens": 51722461.0, + "step": 2008 + }, + { + "epoch": 0.2206237645508456, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.212451457977295, + "learning_rate": 7.349926793557832e-07, + "loss": 0.993, + "mean_token_accuracy": 0.6978591084480286, + "num_tokens": 51750114.0, + "step": 2009 + }, + { + "epoch": 0.22073358225345927, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5020577907562256, + "learning_rate": 7.353587115666178e-07, + "loss": 0.9728, + "mean_token_accuracy": 0.706066906452179, + "num_tokens": 51769519.0, + "step": 2010 + }, + { + "epoch": 0.2208433999560729, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.501894235610962, + "learning_rate": 7.357247437774524e-07, + "loss": 0.9565, + "mean_token_accuracy": 0.7147246599197388, + "num_tokens": 51790952.0, + "step": 2011 + }, + { + "epoch": 0.22095321765868658, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.058986186981201, + "learning_rate": 7.360907759882869e-07, + "loss": 1.0147, + "mean_token_accuracy": 0.7005774974822998, + "num_tokens": 51820766.0, + "step": 2012 + }, + { + "epoch": 0.22106303536130023, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.167109966278076, + "learning_rate": 7.364568081991214e-07, + "loss": 1.0788, + "mean_token_accuracy": 0.6819972395896912, + "num_tokens": 51848779.0, + "step": 2013 + }, + { + "epoch": 0.2211728530639139, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4648478031158447, + "learning_rate": 7.36822840409956e-07, + "loss": 1.0517, + "mean_token_accuracy": 0.6881327629089355, + "num_tokens": 51872683.0, + "step": 2014 + }, + { + "epoch": 0.22128267076652755, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.54945969581604, + "learning_rate": 7.371888726207905e-07, + "loss": 0.982, + "mean_token_accuracy": 0.7031494379043579, + "num_tokens": 51892998.0, + "step": 2015 + }, + { + "epoch": 0.22139248846914122, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3540358543395996, + "learning_rate": 7.375549048316251e-07, + "loss": 0.9966, + "mean_token_accuracy": 0.6964661478996277, + "num_tokens": 51916800.0, + "step": 2016 + }, + { + "epoch": 0.2215023061717549, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.464393138885498, + "learning_rate": 7.379209370424597e-07, + "loss": 0.9723, + "mean_token_accuracy": 0.7064465284347534, + "num_tokens": 51938004.0, + "step": 2017 + }, + { + "epoch": 0.22161212387436854, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5072360038757324, + "learning_rate": 7.382869692532942e-07, + "loss": 0.9882, + "mean_token_accuracy": 0.6981256008148193, + "num_tokens": 51957847.0, + "step": 2018 + }, + { + "epoch": 0.22172194157698222, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.458035707473755, + "learning_rate": 7.386530014641288e-07, + "loss": 0.978, + "mean_token_accuracy": 0.7049554586410522, + "num_tokens": 51980724.0, + "step": 2019 + }, + { + "epoch": 0.22183175927959586, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.267378568649292, + "learning_rate": 7.390190336749633e-07, + "loss": 0.9554, + "mean_token_accuracy": 0.704078197479248, + "num_tokens": 52004838.0, + "step": 2020 + }, + { + "epoch": 0.22194157698220954, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1626200675964355, + "learning_rate": 7.393850658857979e-07, + "loss": 1.0911, + "mean_token_accuracy": 0.6874692440032959, + "num_tokens": 52031951.0, + "step": 2021 + }, + { + "epoch": 0.22205139468482318, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5114388465881348, + "learning_rate": 7.397510980966325e-07, + "loss": 1.0388, + "mean_token_accuracy": 0.6848418116569519, + "num_tokens": 52055321.0, + "step": 2022 + }, + { + "epoch": 0.22216121238743686, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2838172912597656, + "learning_rate": 7.40117130307467e-07, + "loss": 1.0907, + "mean_token_accuracy": 0.673888087272644, + "num_tokens": 52081626.0, + "step": 2023 + }, + { + "epoch": 0.2222710300900505, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 1.8872909545898438, + "learning_rate": 7.404831625183015e-07, + "loss": 1.061, + "mean_token_accuracy": 0.6845107078552246, + "num_tokens": 52116812.0, + "step": 2024 + }, + { + "epoch": 0.22238084779266418, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2220301628112793, + "learning_rate": 7.408491947291361e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6929952502250671, + "num_tokens": 52146142.0, + "step": 2025 + }, + { + "epoch": 0.22249066549527785, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.314373016357422, + "learning_rate": 7.412152269399707e-07, + "loss": 1.0484, + "mean_token_accuracy": 0.6871784925460815, + "num_tokens": 52171070.0, + "step": 2026 + }, + { + "epoch": 0.2226004831978915, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.204176664352417, + "learning_rate": 7.415812591508052e-07, + "loss": 0.943, + "mean_token_accuracy": 0.7109309434890747, + "num_tokens": 52198273.0, + "step": 2027 + }, + { + "epoch": 0.22271030090050517, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4954357147216797, + "learning_rate": 7.419472913616398e-07, + "loss": 1.018, + "mean_token_accuracy": 0.6954523921012878, + "num_tokens": 52220792.0, + "step": 2028 + }, + { + "epoch": 0.22282011860311882, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2342112064361572, + "learning_rate": 7.423133235724743e-07, + "loss": 1.0196, + "mean_token_accuracy": 0.7065372467041016, + "num_tokens": 52248973.0, + "step": 2029 + }, + { + "epoch": 0.2229299363057325, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4537758827209473, + "learning_rate": 7.426793557833088e-07, + "loss": 1.0905, + "mean_token_accuracy": 0.6779759526252747, + "num_tokens": 52271488.0, + "step": 2030 + }, + { + "epoch": 0.22303975400834614, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2233359813690186, + "learning_rate": 7.430453879941434e-07, + "loss": 1.1305, + "mean_token_accuracy": 0.663097083568573, + "num_tokens": 52299754.0, + "step": 2031 + }, + { + "epoch": 0.2231495717109598, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4044289588928223, + "learning_rate": 7.43411420204978e-07, + "loss": 1.091, + "mean_token_accuracy": 0.6785544157028198, + "num_tokens": 52324906.0, + "step": 2032 + }, + { + "epoch": 0.22325938941357346, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.3159332275390625, + "learning_rate": 7.437774524158125e-07, + "loss": 0.9801, + "mean_token_accuracy": 0.7021956443786621, + "num_tokens": 52348354.0, + "step": 2033 + }, + { + "epoch": 0.22336920711618713, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.591506242752075, + "learning_rate": 7.441434846266471e-07, + "loss": 1.0297, + "mean_token_accuracy": 0.6885755062103271, + "num_tokens": 52369396.0, + "step": 2034 + }, + { + "epoch": 0.2234790248188008, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4507713317871094, + "learning_rate": 7.445095168374816e-07, + "loss": 0.8785, + "mean_token_accuracy": 0.7215151786804199, + "num_tokens": 52391080.0, + "step": 2035 + }, + { + "epoch": 0.22358884252141445, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.609581470489502, + "learning_rate": 7.448755490483162e-07, + "loss": 0.9622, + "mean_token_accuracy": 0.7081023454666138, + "num_tokens": 52410968.0, + "step": 2036 + }, + { + "epoch": 0.22369866022402812, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.016132354736328, + "learning_rate": 7.452415812591508e-07, + "loss": 1.0983, + "mean_token_accuracy": 0.6694475412368774, + "num_tokens": 52441502.0, + "step": 2037 + }, + { + "epoch": 0.22380847792664177, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.4043338298797607, + "learning_rate": 7.456076134699853e-07, + "loss": 1.0382, + "mean_token_accuracy": 0.6885977983474731, + "num_tokens": 52464356.0, + "step": 2038 + }, + { + "epoch": 0.22391829562925544, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.2190701961517334, + "learning_rate": 7.459736456808199e-07, + "loss": 1.0284, + "mean_token_accuracy": 0.6908168792724609, + "num_tokens": 52491196.0, + "step": 2039 + }, + { + "epoch": 0.2240281133318691, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.710627794265747, + "learning_rate": 7.463396778916544e-07, + "loss": 1.0245, + "mean_token_accuracy": 0.6922484636306763, + "num_tokens": 52510079.0, + "step": 2040 + }, + { + "epoch": 0.22413793103448276, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.2528018951416016, + "learning_rate": 7.467057101024889e-07, + "loss": 1.1062, + "mean_token_accuracy": 0.6777256727218628, + "num_tokens": 52536899.0, + "step": 2041 + }, + { + "epoch": 0.2242477487370964, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.304865598678589, + "learning_rate": 7.470717423133236e-07, + "loss": 1.113, + "mean_token_accuracy": 0.6734745502471924, + "num_tokens": 52565125.0, + "step": 2042 + }, + { + "epoch": 0.22435756643971008, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.2635443210601807, + "learning_rate": 7.474377745241581e-07, + "loss": 0.9966, + "mean_token_accuracy": 0.7022483348846436, + "num_tokens": 52589724.0, + "step": 2043 + }, + { + "epoch": 0.22446738414232376, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.1501646041870117, + "learning_rate": 7.478038067349926e-07, + "loss": 1.1277, + "mean_token_accuracy": 0.6673271059989929, + "num_tokens": 52620905.0, + "step": 2044 + }, + { + "epoch": 0.2245772018449374, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.466534376144409, + "learning_rate": 7.481698389458272e-07, + "loss": 0.99, + "mean_token_accuracy": 0.7043814659118652, + "num_tokens": 52647382.0, + "step": 2045 + }, + { + "epoch": 0.22468701954755108, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.4240305423736572, + "learning_rate": 7.485358711566617e-07, + "loss": 1.0329, + "mean_token_accuracy": 0.6864531636238098, + "num_tokens": 52670035.0, + "step": 2046 + }, + { + "epoch": 0.22479683725016472, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 1.958688735961914, + "learning_rate": 7.489019033674962e-07, + "loss": 1.0082, + "mean_token_accuracy": 0.692980945110321, + "num_tokens": 52702091.0, + "step": 2047 + }, + { + "epoch": 0.2249066549527784, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.377976894378662, + "learning_rate": 7.492679355783309e-07, + "loss": 0.9786, + "mean_token_accuracy": 0.7027292251586914, + "num_tokens": 52723476.0, + "step": 2048 + }, + { + "epoch": 0.22501647265539204, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.1523332595825195, + "learning_rate": 7.496339677891654e-07, + "loss": 0.929, + "mean_token_accuracy": 0.7151409387588501, + "num_tokens": 52750250.0, + "step": 2049 + }, + { + "epoch": 0.22512629035800572, + "ewc_loss": 1.2934207916259766e-05, + "grad_norm": 2.2851622104644775, + "learning_rate": 7.5e-07, + "loss": 1.0076, + "mean_token_accuracy": 0.69345623254776, + "num_tokens": 52774913.0, + "step": 2050 + }, + { + "epoch": 0.22523610806061936, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3131399154663086, + "learning_rate": 7.503660322108345e-07, + "loss": 1.0758, + "mean_token_accuracy": 0.6811903715133667, + "num_tokens": 52799061.0, + "step": 2051 + }, + { + "epoch": 0.22534592576323303, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1681737899780273, + "learning_rate": 7.50732064421669e-07, + "loss": 1.0318, + "mean_token_accuracy": 0.6885703206062317, + "num_tokens": 52827682.0, + "step": 2052 + }, + { + "epoch": 0.22545574346584668, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.232877731323242, + "learning_rate": 7.510980966325037e-07, + "loss": 1.0796, + "mean_token_accuracy": 0.6771611571311951, + "num_tokens": 52854004.0, + "step": 2053 + }, + { + "epoch": 0.22556556116846035, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5969316959381104, + "learning_rate": 7.514641288433382e-07, + "loss": 1.0489, + "mean_token_accuracy": 0.6965963840484619, + "num_tokens": 52875334.0, + "step": 2054 + }, + { + "epoch": 0.22567537887107403, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.258490562438965, + "learning_rate": 7.518301610541727e-07, + "loss": 1.0391, + "mean_token_accuracy": 0.6933624148368835, + "num_tokens": 52901072.0, + "step": 2055 + }, + { + "epoch": 0.22578519657368767, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.159205436706543, + "learning_rate": 7.521961932650073e-07, + "loss": 1.0393, + "mean_token_accuracy": 0.692175030708313, + "num_tokens": 52929140.0, + "step": 2056 + }, + { + "epoch": 0.22589501427630135, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.346015453338623, + "learning_rate": 7.525622254758418e-07, + "loss": 0.9782, + "mean_token_accuracy": 0.7028012275695801, + "num_tokens": 52953279.0, + "step": 2057 + }, + { + "epoch": 0.226004831978915, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.0365853309631348, + "learning_rate": 7.529282576866763e-07, + "loss": 1.0729, + "mean_token_accuracy": 0.6767573952674866, + "num_tokens": 52985489.0, + "step": 2058 + }, + { + "epoch": 0.22611464968152867, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.506448268890381, + "learning_rate": 7.53294289897511e-07, + "loss": 0.9952, + "mean_token_accuracy": 0.6968573331832886, + "num_tokens": 53008432.0, + "step": 2059 + }, + { + "epoch": 0.2262244673841423, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5677711963653564, + "learning_rate": 7.536603221083455e-07, + "loss": 1.006, + "mean_token_accuracy": 0.707015335559845, + "num_tokens": 53030196.0, + "step": 2060 + }, + { + "epoch": 0.226334285086756, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2503819465637207, + "learning_rate": 7.5402635431918e-07, + "loss": 1.004, + "mean_token_accuracy": 0.70432448387146, + "num_tokens": 53059442.0, + "step": 2061 + }, + { + "epoch": 0.22644410278936963, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.58333158493042, + "learning_rate": 7.543923865300146e-07, + "loss": 0.9893, + "mean_token_accuracy": 0.7027499079704285, + "num_tokens": 53081527.0, + "step": 2062 + }, + { + "epoch": 0.2265539204919833, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.155160665512085, + "learning_rate": 7.547584187408491e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.689726710319519, + "num_tokens": 53109281.0, + "step": 2063 + }, + { + "epoch": 0.22666373819459698, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.2427046298980713, + "learning_rate": 7.551244509516837e-07, + "loss": 1.0008, + "mean_token_accuracy": 0.70402991771698, + "num_tokens": 53134575.0, + "step": 2064 + }, + { + "epoch": 0.22677355589721063, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.183748960494995, + "learning_rate": 7.554904831625183e-07, + "loss": 1.0012, + "mean_token_accuracy": 0.6989303827285767, + "num_tokens": 53163685.0, + "step": 2065 + }, + { + "epoch": 0.2268833735998243, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.0837831497192383, + "learning_rate": 7.558565153733528e-07, + "loss": 1.0328, + "mean_token_accuracy": 0.6939947605133057, + "num_tokens": 53193254.0, + "step": 2066 + }, + { + "epoch": 0.22699319130243795, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.4565515518188477, + "learning_rate": 7.562225475841874e-07, + "loss": 1.0254, + "mean_token_accuracy": 0.6918078064918518, + "num_tokens": 53216268.0, + "step": 2067 + }, + { + "epoch": 0.22710300900505162, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3393197059631348, + "learning_rate": 7.565885797950219e-07, + "loss": 0.9277, + "mean_token_accuracy": 0.7151058912277222, + "num_tokens": 53239331.0, + "step": 2068 + }, + { + "epoch": 0.22721282670766527, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.150428056716919, + "learning_rate": 7.569546120058565e-07, + "loss": 1.0622, + "mean_token_accuracy": 0.691943883895874, + "num_tokens": 53269366.0, + "step": 2069 + }, + { + "epoch": 0.22732264441027894, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.4004170894622803, + "learning_rate": 7.573206442166911e-07, + "loss": 1.0194, + "mean_token_accuracy": 0.6938802599906921, + "num_tokens": 53293551.0, + "step": 2070 + }, + { + "epoch": 0.22743246211289259, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.3344240188598633, + "learning_rate": 7.576866764275256e-07, + "loss": 0.9683, + "mean_token_accuracy": 0.7011438608169556, + "num_tokens": 53317407.0, + "step": 2071 + }, + { + "epoch": 0.22754227981550626, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.0875985622406006, + "learning_rate": 7.580527086383601e-07, + "loss": 1.104, + "mean_token_accuracy": 0.6823805570602417, + "num_tokens": 53349129.0, + "step": 2072 + }, + { + "epoch": 0.22765209751811993, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.4290666580200195, + "learning_rate": 7.584187408491947e-07, + "loss": 0.994, + "mean_token_accuracy": 0.6976988315582275, + "num_tokens": 53373807.0, + "step": 2073 + }, + { + "epoch": 0.22776191522073358, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.408529758453369, + "learning_rate": 7.587847730600292e-07, + "loss": 1.1029, + "mean_token_accuracy": 0.6741502285003662, + "num_tokens": 53398993.0, + "step": 2074 + }, + { + "epoch": 0.22787173292334725, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.10267972946167, + "learning_rate": 7.591508052708638e-07, + "loss": 1.1144, + "mean_token_accuracy": 0.6738095879554749, + "num_tokens": 53430064.0, + "step": 2075 + }, + { + "epoch": 0.2279815506259609, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1505961418151855, + "learning_rate": 7.595168374816984e-07, + "loss": 0.9892, + "mean_token_accuracy": 0.6941893696784973, + "num_tokens": 53459080.0, + "step": 2076 + }, + { + "epoch": 0.22809136832857457, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1964380741119385, + "learning_rate": 7.598828696925329e-07, + "loss": 1.1091, + "mean_token_accuracy": 0.674208402633667, + "num_tokens": 53488249.0, + "step": 2077 + }, + { + "epoch": 0.22820118603118822, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.281898260116577, + "learning_rate": 7.602489019033674e-07, + "loss": 0.8927, + "mean_token_accuracy": 0.7259369492530823, + "num_tokens": 53512224.0, + "step": 2078 + }, + { + "epoch": 0.2283110037338019, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2551939487457275, + "learning_rate": 7.60614934114202e-07, + "loss": 1.0449, + "mean_token_accuracy": 0.6848801374435425, + "num_tokens": 53539216.0, + "step": 2079 + }, + { + "epoch": 0.22842082143641554, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6182801723480225, + "learning_rate": 7.609809663250366e-07, + "loss": 0.9561, + "mean_token_accuracy": 0.708185076713562, + "num_tokens": 53558834.0, + "step": 2080 + }, + { + "epoch": 0.2285306391390292, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.0776288509368896, + "learning_rate": 7.613469985358711e-07, + "loss": 1.0283, + "mean_token_accuracy": 0.691518247127533, + "num_tokens": 53587249.0, + "step": 2081 + }, + { + "epoch": 0.2286404568416429, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4779508113861084, + "learning_rate": 7.617130307467057e-07, + "loss": 0.9886, + "mean_token_accuracy": 0.701148509979248, + "num_tokens": 53610048.0, + "step": 2082 + }, + { + "epoch": 0.22875027454425653, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1304726600646973, + "learning_rate": 7.620790629575402e-07, + "loss": 1.0104, + "mean_token_accuracy": 0.6973831653594971, + "num_tokens": 53639322.0, + "step": 2083 + }, + { + "epoch": 0.2288600922468702, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4228806495666504, + "learning_rate": 7.624450951683748e-07, + "loss": 0.9595, + "mean_token_accuracy": 0.7099246978759766, + "num_tokens": 53661912.0, + "step": 2084 + }, + { + "epoch": 0.22896990994948385, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.192563772201538, + "learning_rate": 7.628111273792094e-07, + "loss": 1.0013, + "mean_token_accuracy": 0.7028775215148926, + "num_tokens": 53692100.0, + "step": 2085 + }, + { + "epoch": 0.22907972765209753, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.079585075378418, + "learning_rate": 7.631771595900439e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6831849217414856, + "num_tokens": 53721317.0, + "step": 2086 + }, + { + "epoch": 0.22918954535471117, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.489473581314087, + "learning_rate": 7.635431918008785e-07, + "loss": 1.0707, + "mean_token_accuracy": 0.6893606781959534, + "num_tokens": 53745152.0, + "step": 2087 + }, + { + "epoch": 0.22929936305732485, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6389734745025635, + "learning_rate": 7.63909224011713e-07, + "loss": 0.9706, + "mean_token_accuracy": 0.7115990519523621, + "num_tokens": 53766720.0, + "step": 2088 + }, + { + "epoch": 0.2294091807599385, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.151512861251831, + "learning_rate": 7.642752562225475e-07, + "loss": 0.9927, + "mean_token_accuracy": 0.6954758167266846, + "num_tokens": 53794504.0, + "step": 2089 + }, + { + "epoch": 0.22951899846255217, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.285928249359131, + "learning_rate": 7.646412884333821e-07, + "loss": 1.0239, + "mean_token_accuracy": 0.696064829826355, + "num_tokens": 53820018.0, + "step": 2090 + }, + { + "epoch": 0.2296288161651658, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0370969772338867, + "learning_rate": 7.650073206442167e-07, + "loss": 1.0518, + "mean_token_accuracy": 0.6879433393478394, + "num_tokens": 53851435.0, + "step": 2091 + }, + { + "epoch": 0.22973863386777948, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5445752143859863, + "learning_rate": 7.653733528550512e-07, + "loss": 0.9816, + "mean_token_accuracy": 0.7030807733535767, + "num_tokens": 53873241.0, + "step": 2092 + }, + { + "epoch": 0.22984845157039316, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1188066005706787, + "learning_rate": 7.657393850658858e-07, + "loss": 1.0538, + "mean_token_accuracy": 0.6880027055740356, + "num_tokens": 53903030.0, + "step": 2093 + }, + { + "epoch": 0.2299582692730068, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.531099319458008, + "learning_rate": 7.661054172767203e-07, + "loss": 1.0207, + "mean_token_accuracy": 0.6947214007377625, + "num_tokens": 53925218.0, + "step": 2094 + }, + { + "epoch": 0.23006808697562048, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4222891330718994, + "learning_rate": 7.664714494875548e-07, + "loss": 1.0779, + "mean_token_accuracy": 0.6829856634140015, + "num_tokens": 53948843.0, + "step": 2095 + }, + { + "epoch": 0.23017790467823412, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.058403253555298, + "learning_rate": 7.668374816983895e-07, + "loss": 0.9751, + "mean_token_accuracy": 0.7034921050071716, + "num_tokens": 53977112.0, + "step": 2096 + }, + { + "epoch": 0.2302877223808478, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3324074745178223, + "learning_rate": 7.67203513909224e-07, + "loss": 1.0059, + "mean_token_accuracy": 0.7031497955322266, + "num_tokens": 54002810.0, + "step": 2097 + }, + { + "epoch": 0.23039754008346144, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2722156047821045, + "learning_rate": 7.675695461200585e-07, + "loss": 1.0029, + "mean_token_accuracy": 0.7057522535324097, + "num_tokens": 54029780.0, + "step": 2098 + }, + { + "epoch": 0.23050735778607512, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2458224296569824, + "learning_rate": 7.679355783308931e-07, + "loss": 1.0774, + "mean_token_accuracy": 0.6863187551498413, + "num_tokens": 54054899.0, + "step": 2099 + }, + { + "epoch": 0.23061717548868876, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1584134101867676, + "learning_rate": 7.683016105417276e-07, + "loss": 1.101, + "mean_token_accuracy": 0.6822142601013184, + "num_tokens": 54084137.0, + "step": 2100 + }, + { + "epoch": 0.23072699319130244, + "ewc_loss": 1.3470649719238281e-05, + "grad_norm": 2.3189330101013184, + "learning_rate": 7.686676427525623e-07, + "loss": 0.9592, + "mean_token_accuracy": 0.7161327004432678, + "num_tokens": 54108823.0, + "step": 2101 + }, + { + "epoch": 0.2308368108939161, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.429964303970337, + "learning_rate": 7.690336749633968e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7024046778678894, + "num_tokens": 54131262.0, + "step": 2102 + }, + { + "epoch": 0.23094662859652976, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3584611415863037, + "learning_rate": 7.693997071742313e-07, + "loss": 1.0131, + "mean_token_accuracy": 0.6921427249908447, + "num_tokens": 54154640.0, + "step": 2103 + }, + { + "epoch": 0.23105644629914343, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.0221292972564697, + "learning_rate": 7.697657393850659e-07, + "loss": 0.9939, + "mean_token_accuracy": 0.7001744508743286, + "num_tokens": 54185330.0, + "step": 2104 + }, + { + "epoch": 0.23116626400175708, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5350799560546875, + "learning_rate": 7.701317715959004e-07, + "loss": 0.964, + "mean_token_accuracy": 0.7076954245567322, + "num_tokens": 54206266.0, + "step": 2105 + }, + { + "epoch": 0.23127608170437075, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.164363145828247, + "learning_rate": 7.704978038067349e-07, + "loss": 1.0748, + "mean_token_accuracy": 0.6758511066436768, + "num_tokens": 54234830.0, + "step": 2106 + }, + { + "epoch": 0.2313858994069844, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.329446315765381, + "learning_rate": 7.708638360175696e-07, + "loss": 0.8871, + "mean_token_accuracy": 0.72446608543396, + "num_tokens": 54258518.0, + "step": 2107 + }, + { + "epoch": 0.23149571710959807, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1804373264312744, + "learning_rate": 7.712298682284041e-07, + "loss": 1.0784, + "mean_token_accuracy": 0.6810320615768433, + "num_tokens": 54285139.0, + "step": 2108 + }, + { + "epoch": 0.23160553481221172, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3814897537231445, + "learning_rate": 7.715959004392386e-07, + "loss": 0.9844, + "mean_token_accuracy": 0.7017801403999329, + "num_tokens": 54308341.0, + "step": 2109 + }, + { + "epoch": 0.2317153525148254, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1850576400756836, + "learning_rate": 7.719619326500732e-07, + "loss": 1.1339, + "mean_token_accuracy": 0.6656463146209717, + "num_tokens": 54338009.0, + "step": 2110 + }, + { + "epoch": 0.23182517021743906, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4819529056549072, + "learning_rate": 7.723279648609077e-07, + "loss": 1.0834, + "mean_token_accuracy": 0.6793409585952759, + "num_tokens": 54361156.0, + "step": 2111 + }, + { + "epoch": 0.2319349879200527, + "ewc_loss": 1.3709068298339844e-05, + "grad_norm": 1.9382151365280151, + "learning_rate": 7.726939970717423e-07, + "loss": 0.8934, + "mean_token_accuracy": 0.7266330718994141, + "num_tokens": 54390645.0, + "step": 2112 + }, + { + "epoch": 0.23204480562266638, + "ewc_loss": 1.3709068298339844e-05, + "grad_norm": 2.4408481121063232, + "learning_rate": 7.730600292825769e-07, + "loss": 0.9948, + "mean_token_accuracy": 0.6940502524375916, + "num_tokens": 54412841.0, + "step": 2113 + }, + { + "epoch": 0.23215462332528003, + "ewc_loss": 1.3709068298339844e-05, + "grad_norm": 2.136518955230713, + "learning_rate": 7.734260614934114e-07, + "loss": 0.9704, + "mean_token_accuracy": 0.7039232850074768, + "num_tokens": 54440731.0, + "step": 2114 + }, + { + "epoch": 0.2322644410278937, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.713369607925415, + "learning_rate": 7.737920937042459e-07, + "loss": 1.0111, + "mean_token_accuracy": 0.6937790513038635, + "num_tokens": 54460803.0, + "step": 2115 + }, + { + "epoch": 0.23237425873050735, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4825358390808105, + "learning_rate": 7.741581259150805e-07, + "loss": 0.9866, + "mean_token_accuracy": 0.7075942754745483, + "num_tokens": 54483265.0, + "step": 2116 + }, + { + "epoch": 0.23248407643312102, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.1738626956939697, + "learning_rate": 7.745241581259151e-07, + "loss": 0.9873, + "mean_token_accuracy": 0.7020373940467834, + "num_tokens": 54509995.0, + "step": 2117 + }, + { + "epoch": 0.23259389413573467, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.406323194503784, + "learning_rate": 7.748901903367497e-07, + "loss": 0.9837, + "mean_token_accuracy": 0.7007649540901184, + "num_tokens": 54533911.0, + "step": 2118 + }, + { + "epoch": 0.23270371183834834, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5376381874084473, + "learning_rate": 7.752562225475842e-07, + "loss": 0.9901, + "mean_token_accuracy": 0.7003054022789001, + "num_tokens": 54555628.0, + "step": 2119 + }, + { + "epoch": 0.23281352954096202, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.126413345336914, + "learning_rate": 7.756222547584187e-07, + "loss": 1.0428, + "mean_token_accuracy": 0.6917967796325684, + "num_tokens": 54584663.0, + "step": 2120 + }, + { + "epoch": 0.23292334724357566, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.6321747303009033, + "learning_rate": 7.759882869692533e-07, + "loss": 1.0189, + "mean_token_accuracy": 0.7010713815689087, + "num_tokens": 54606332.0, + "step": 2121 + }, + { + "epoch": 0.23303316494618934, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.041486978530884, + "learning_rate": 7.763543191800878e-07, + "loss": 1.1008, + "mean_token_accuracy": 0.6774317026138306, + "num_tokens": 54639008.0, + "step": 2122 + }, + { + "epoch": 0.23314298264880298, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.088387966156006, + "learning_rate": 7.767203513909224e-07, + "loss": 1.042, + "mean_token_accuracy": 0.6859986186027527, + "num_tokens": 54668941.0, + "step": 2123 + }, + { + "epoch": 0.23325280035141666, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.4999423027038574, + "learning_rate": 7.77086383601757e-07, + "loss": 0.9911, + "mean_token_accuracy": 0.6982353925704956, + "num_tokens": 54691348.0, + "step": 2124 + }, + { + "epoch": 0.2333626180540303, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.257021188735962, + "learning_rate": 7.774524158125915e-07, + "loss": 0.9819, + "mean_token_accuracy": 0.705440104007721, + "num_tokens": 54715840.0, + "step": 2125 + }, + { + "epoch": 0.23347243575664398, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.27689790725708, + "learning_rate": 7.77818448023426e-07, + "loss": 1.109, + "mean_token_accuracy": 0.6669577956199646, + "num_tokens": 54742261.0, + "step": 2126 + }, + { + "epoch": 0.23358225345925762, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 1.9698808193206787, + "learning_rate": 7.781844802342606e-07, + "loss": 1.0923, + "mean_token_accuracy": 0.6724993586540222, + "num_tokens": 54775001.0, + "step": 2127 + }, + { + "epoch": 0.2336920711618713, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.307617425918579, + "learning_rate": 7.785505124450952e-07, + "loss": 1.0669, + "mean_token_accuracy": 0.6824973821640015, + "num_tokens": 54800287.0, + "step": 2128 + }, + { + "epoch": 0.23380188886448494, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.1862053871154785, + "learning_rate": 7.789165446559297e-07, + "loss": 1.0877, + "mean_token_accuracy": 0.6809177398681641, + "num_tokens": 54827212.0, + "step": 2129 + }, + { + "epoch": 0.23391170656709862, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.245046615600586, + "learning_rate": 7.792825768667643e-07, + "loss": 1.1603, + "mean_token_accuracy": 0.6589298248291016, + "num_tokens": 54855265.0, + "step": 2130 + }, + { + "epoch": 0.2340215242697123, + "ewc_loss": 1.3947486877441406e-05, + "grad_norm": 2.432196855545044, + "learning_rate": 7.796486090775988e-07, + "loss": 1.0719, + "mean_token_accuracy": 0.6780922412872314, + "num_tokens": 54880558.0, + "step": 2131 + }, + { + "epoch": 0.23413134197232593, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2060043811798096, + "learning_rate": 7.800146412884333e-07, + "loss": 1.104, + "mean_token_accuracy": 0.6757687330245972, + "num_tokens": 54908760.0, + "step": 2132 + }, + { + "epoch": 0.2342411596749396, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4249136447906494, + "learning_rate": 7.80380673499268e-07, + "loss": 0.9204, + "mean_token_accuracy": 0.714383602142334, + "num_tokens": 54927989.0, + "step": 2133 + }, + { + "epoch": 0.23435097737755325, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.167067050933838, + "learning_rate": 7.807467057101025e-07, + "loss": 1.0476, + "mean_token_accuracy": 0.6874265074729919, + "num_tokens": 54955289.0, + "step": 2134 + }, + { + "epoch": 0.23446079508016693, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.2288949489593506, + "learning_rate": 7.811127379209371e-07, + "loss": 0.9781, + "mean_token_accuracy": 0.6990408897399902, + "num_tokens": 54981124.0, + "step": 2135 + }, + { + "epoch": 0.23457061278278057, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.116457462310791, + "learning_rate": 7.814787701317716e-07, + "loss": 1.0698, + "mean_token_accuracy": 0.6828757524490356, + "num_tokens": 55009205.0, + "step": 2136 + }, + { + "epoch": 0.23468043048539425, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.241086959838867, + "learning_rate": 7.818448023426061e-07, + "loss": 1.1221, + "mean_token_accuracy": 0.6672942638397217, + "num_tokens": 55036118.0, + "step": 2137 + }, + { + "epoch": 0.2347902481880079, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.2519333362579346, + "learning_rate": 7.822108345534407e-07, + "loss": 0.9577, + "mean_token_accuracy": 0.7044615745544434, + "num_tokens": 55060467.0, + "step": 2138 + }, + { + "epoch": 0.23490006589062157, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.2479441165924072, + "learning_rate": 7.825768667642753e-07, + "loss": 1.0971, + "mean_token_accuracy": 0.6745800375938416, + "num_tokens": 55087375.0, + "step": 2139 + }, + { + "epoch": 0.23500988359323524, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.291926860809326, + "learning_rate": 7.829428989751098e-07, + "loss": 1.0301, + "mean_token_accuracy": 0.6972384452819824, + "num_tokens": 55112810.0, + "step": 2140 + }, + { + "epoch": 0.2351197012958489, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.522141456604004, + "learning_rate": 7.833089311859444e-07, + "loss": 1.0217, + "mean_token_accuracy": 0.6995327472686768, + "num_tokens": 55134221.0, + "step": 2141 + }, + { + "epoch": 0.23522951899846256, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.1874406337738037, + "learning_rate": 7.836749633967789e-07, + "loss": 1.0312, + "mean_token_accuracy": 0.6952881813049316, + "num_tokens": 55161282.0, + "step": 2142 + }, + { + "epoch": 0.2353393367010762, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.387099266052246, + "learning_rate": 7.840409956076133e-07, + "loss": 1.0516, + "mean_token_accuracy": 0.6889549493789673, + "num_tokens": 55185586.0, + "step": 2143 + }, + { + "epoch": 0.23544915440368988, + "ewc_loss": 1.4126300811767578e-05, + "grad_norm": 2.0927951335906982, + "learning_rate": 7.844070278184481e-07, + "loss": 1.155, + "mean_token_accuracy": 0.663084864616394, + "num_tokens": 55214146.0, + "step": 2144 + }, + { + "epoch": 0.23555897210630353, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.504086494445801, + "learning_rate": 7.847730600292826e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.6942468881607056, + "num_tokens": 55234676.0, + "step": 2145 + }, + { + "epoch": 0.2356687898089172, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.330045223236084, + "learning_rate": 7.85139092240117e-07, + "loss": 0.9362, + "mean_token_accuracy": 0.7111728191375732, + "num_tokens": 55257852.0, + "step": 2146 + }, + { + "epoch": 0.23577860751153085, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4078056812286377, + "learning_rate": 7.855051244509517e-07, + "loss": 1.0176, + "mean_token_accuracy": 0.6940369606018066, + "num_tokens": 55280103.0, + "step": 2147 + }, + { + "epoch": 0.23588842521414452, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.230588674545288, + "learning_rate": 7.858711566617861e-07, + "loss": 1.0534, + "mean_token_accuracy": 0.6782476305961609, + "num_tokens": 55306689.0, + "step": 2148 + }, + { + "epoch": 0.2359982429167582, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1897835731506348, + "learning_rate": 7.862371888726209e-07, + "loss": 1.0889, + "mean_token_accuracy": 0.6763337850570679, + "num_tokens": 55335557.0, + "step": 2149 + }, + { + "epoch": 0.23610806061937184, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3988187313079834, + "learning_rate": 7.866032210834554e-07, + "loss": 1.0392, + "mean_token_accuracy": 0.6885308027267456, + "num_tokens": 55361347.0, + "step": 2150 + }, + { + "epoch": 0.2362178783219855, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3154544830322266, + "learning_rate": 7.869692532942898e-07, + "loss": 1.0305, + "mean_token_accuracy": 0.6887891888618469, + "num_tokens": 55386252.0, + "step": 2151 + }, + { + "epoch": 0.23632769602459916, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3962206840515137, + "learning_rate": 7.873352855051244e-07, + "loss": 1.0571, + "mean_token_accuracy": 0.6927590370178223, + "num_tokens": 55407899.0, + "step": 2152 + }, + { + "epoch": 0.23643751372721283, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3140664100646973, + "learning_rate": 7.877013177159589e-07, + "loss": 0.869, + "mean_token_accuracy": 0.7298116683959961, + "num_tokens": 55430617.0, + "step": 2153 + }, + { + "epoch": 0.23654733142982648, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.5968706607818604, + "learning_rate": 7.880673499267934e-07, + "loss": 0.8577, + "mean_token_accuracy": 0.7317409515380859, + "num_tokens": 55449753.0, + "step": 2154 + }, + { + "epoch": 0.23665714913244015, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.2023677825927734, + "learning_rate": 7.884333821376281e-07, + "loss": 1.0423, + "mean_token_accuracy": 0.6880885362625122, + "num_tokens": 55477552.0, + "step": 2155 + }, + { + "epoch": 0.2367669668350538, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3413960933685303, + "learning_rate": 7.887994143484626e-07, + "loss": 0.9918, + "mean_token_accuracy": 0.7018099427223206, + "num_tokens": 55500986.0, + "step": 2156 + }, + { + "epoch": 0.23687678453766747, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.069380283355713, + "learning_rate": 7.891654465592971e-07, + "loss": 1.0569, + "mean_token_accuracy": 0.6861818432807922, + "num_tokens": 55533557.0, + "step": 2157 + }, + { + "epoch": 0.23698660224028115, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.1706173419952393, + "learning_rate": 7.895314787701317e-07, + "loss": 1.0515, + "mean_token_accuracy": 0.6813341379165649, + "num_tokens": 55562766.0, + "step": 2158 + }, + { + "epoch": 0.2370964199428948, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.2550113201141357, + "learning_rate": 7.898975109809662e-07, + "loss": 1.1054, + "mean_token_accuracy": 0.6692755818367004, + "num_tokens": 55590781.0, + "step": 2159 + }, + { + "epoch": 0.23720623764550847, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3870816230773926, + "learning_rate": 7.902635431918008e-07, + "loss": 1.0019, + "mean_token_accuracy": 0.6989096403121948, + "num_tokens": 55615755.0, + "step": 2160 + }, + { + "epoch": 0.2373160553481221, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.3283843994140625, + "learning_rate": 7.906295754026354e-07, + "loss": 0.9906, + "mean_token_accuracy": 0.7016564011573792, + "num_tokens": 55639958.0, + "step": 2161 + }, + { + "epoch": 0.2374258730507358, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.526747941970825, + "learning_rate": 7.909956076134699e-07, + "loss": 1.0654, + "mean_token_accuracy": 0.6814451813697815, + "num_tokens": 55661257.0, + "step": 2162 + }, + { + "epoch": 0.23753569075334943, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.188570261001587, + "learning_rate": 7.913616398243044e-07, + "loss": 0.9424, + "mean_token_accuracy": 0.7209254503250122, + "num_tokens": 55687147.0, + "step": 2163 + }, + { + "epoch": 0.2376455084559631, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.021409749984741, + "learning_rate": 7.91727672035139e-07, + "loss": 1.0547, + "mean_token_accuracy": 0.6851293444633484, + "num_tokens": 55718756.0, + "step": 2164 + }, + { + "epoch": 0.23775532615857675, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.2958250045776367, + "learning_rate": 7.920937042459736e-07, + "loss": 1.1015, + "mean_token_accuracy": 0.6750126481056213, + "num_tokens": 55745692.0, + "step": 2165 + }, + { + "epoch": 0.23786514386119043, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.200766086578369, + "learning_rate": 7.924597364568082e-07, + "loss": 0.9958, + "mean_token_accuracy": 0.7061862945556641, + "num_tokens": 55771595.0, + "step": 2166 + }, + { + "epoch": 0.23797496156380407, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 2.2824134826660156, + "learning_rate": 7.928257686676427e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.6890903115272522, + "num_tokens": 55795574.0, + "step": 2167 + }, + { + "epoch": 0.23808477926641775, + "ewc_loss": 1.436471939086914e-05, + "grad_norm": 2.203207492828369, + "learning_rate": 7.931918008784772e-07, + "loss": 1.0634, + "mean_token_accuracy": 0.6841180324554443, + "num_tokens": 55823116.0, + "step": 2168 + }, + { + "epoch": 0.23819459696903142, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.1882708072662354, + "learning_rate": 7.935578330893118e-07, + "loss": 1.0332, + "mean_token_accuracy": 0.6872217655181885, + "num_tokens": 55849908.0, + "step": 2169 + }, + { + "epoch": 0.23830441467164507, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 2.086331605911255, + "learning_rate": 7.939238653001463e-07, + "loss": 1.0596, + "mean_token_accuracy": 0.6833139657974243, + "num_tokens": 55879307.0, + "step": 2170 + }, + { + "epoch": 0.23841423237425874, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 2.3405721187591553, + "learning_rate": 7.942898975109809e-07, + "loss": 1.0987, + "mean_token_accuracy": 0.675906240940094, + "num_tokens": 55907879.0, + "step": 2171 + }, + { + "epoch": 0.23852405007687238, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 2.524338722229004, + "learning_rate": 7.946559297218155e-07, + "loss": 1.1104, + "mean_token_accuracy": 0.6766753196716309, + "num_tokens": 55930065.0, + "step": 2172 + }, + { + "epoch": 0.23863386777948606, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 5.098357677459717, + "learning_rate": 7.9502196193265e-07, + "loss": 1.0095, + "mean_token_accuracy": 0.6945990324020386, + "num_tokens": 55954500.0, + "step": 2173 + }, + { + "epoch": 0.2387436854820997, + "ewc_loss": 1.4483928680419922e-05, + "grad_norm": 2.2222959995269775, + "learning_rate": 7.953879941434845e-07, + "loss": 0.9368, + "mean_token_accuracy": 0.7138302326202393, + "num_tokens": 55980294.0, + "step": 2174 + }, + { + "epoch": 0.23885350318471338, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 1.9518098831176758, + "learning_rate": 7.957540263543191e-07, + "loss": 1.0228, + "mean_token_accuracy": 0.6963818669319153, + "num_tokens": 56012046.0, + "step": 2175 + }, + { + "epoch": 0.23896332088732702, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.242469072341919, + "learning_rate": 7.961200585651537e-07, + "loss": 1.0091, + "mean_token_accuracy": 0.6981155872344971, + "num_tokens": 56039172.0, + "step": 2176 + }, + { + "epoch": 0.2390731385899407, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4715182781219482, + "learning_rate": 7.964860907759882e-07, + "loss": 1.0017, + "mean_token_accuracy": 0.702181875705719, + "num_tokens": 56061746.0, + "step": 2177 + }, + { + "epoch": 0.23918295629255437, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3312768936157227, + "learning_rate": 7.968521229868228e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6813493967056274, + "num_tokens": 56086553.0, + "step": 2178 + }, + { + "epoch": 0.23929277399516802, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.091001272201538, + "learning_rate": 7.972181551976573e-07, + "loss": 1.1013, + "mean_token_accuracy": 0.6739495396614075, + "num_tokens": 56118829.0, + "step": 2179 + }, + { + "epoch": 0.2394025916977817, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.166900873184204, + "learning_rate": 7.975841874084918e-07, + "loss": 1.0475, + "mean_token_accuracy": 0.6862794160842896, + "num_tokens": 56145714.0, + "step": 2180 + }, + { + "epoch": 0.23951240940039534, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.397873640060425, + "learning_rate": 7.979502196193265e-07, + "loss": 1.0474, + "mean_token_accuracy": 0.683583676815033, + "num_tokens": 56170868.0, + "step": 2181 + }, + { + "epoch": 0.239622227103009, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3437037467956543, + "learning_rate": 7.98316251830161e-07, + "loss": 1.0215, + "mean_token_accuracy": 0.6944456100463867, + "num_tokens": 56195851.0, + "step": 2182 + }, + { + "epoch": 0.23973204480562266, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0375332832336426, + "learning_rate": 7.986822840409956e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.6867258548736572, + "num_tokens": 56227161.0, + "step": 2183 + }, + { + "epoch": 0.23984186250823633, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4498844146728516, + "learning_rate": 7.990483162518301e-07, + "loss": 1.088, + "mean_token_accuracy": 0.6777199506759644, + "num_tokens": 56249842.0, + "step": 2184 + }, + { + "epoch": 0.23995168021084998, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2363760471343994, + "learning_rate": 7.994143484626646e-07, + "loss": 1.0432, + "mean_token_accuracy": 0.6889955401420593, + "num_tokens": 56276810.0, + "step": 2185 + }, + { + "epoch": 0.24006149791346365, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.255317449569702, + "learning_rate": 7.997803806734992e-07, + "loss": 0.9934, + "mean_token_accuracy": 0.697111189365387, + "num_tokens": 56301375.0, + "step": 2186 + }, + { + "epoch": 0.24017131561607732, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2061033248901367, + "learning_rate": 8.001464128843338e-07, + "loss": 1.0517, + "mean_token_accuracy": 0.6835136413574219, + "num_tokens": 56327061.0, + "step": 2187 + }, + { + "epoch": 0.24028113331869097, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1104226112365723, + "learning_rate": 8.005124450951683e-07, + "loss": 1.0191, + "mean_token_accuracy": 0.6945523023605347, + "num_tokens": 56357185.0, + "step": 2188 + }, + { + "epoch": 0.24039095102130464, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3863437175750732, + "learning_rate": 8.008784773060029e-07, + "loss": 1.0455, + "mean_token_accuracy": 0.6856873035430908, + "num_tokens": 56380411.0, + "step": 2189 + }, + { + "epoch": 0.2405007687239183, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5422613620758057, + "learning_rate": 8.012445095168374e-07, + "loss": 1.0025, + "mean_token_accuracy": 0.7040855884552002, + "num_tokens": 56403539.0, + "step": 2190 + }, + { + "epoch": 0.24061058642653196, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2172322273254395, + "learning_rate": 8.016105417276719e-07, + "loss": 1.0095, + "mean_token_accuracy": 0.7006431818008423, + "num_tokens": 56430207.0, + "step": 2191 + }, + { + "epoch": 0.2407204041291456, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.05426287651062, + "learning_rate": 8.019765739385066e-07, + "loss": 1.0842, + "mean_token_accuracy": 0.6744686365127563, + "num_tokens": 56460386.0, + "step": 2192 + }, + { + "epoch": 0.24083022183175928, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.19185471534729, + "learning_rate": 8.023426061493411e-07, + "loss": 0.9333, + "mean_token_accuracy": 0.717103123664856, + "num_tokens": 56487140.0, + "step": 2193 + }, + { + "epoch": 0.24094003953437293, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5576469898223877, + "learning_rate": 8.027086383601756e-07, + "loss": 1.106, + "mean_token_accuracy": 0.6697673797607422, + "num_tokens": 56508781.0, + "step": 2194 + }, + { + "epoch": 0.2410498572369866, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.381678581237793, + "learning_rate": 8.030746705710102e-07, + "loss": 1.0411, + "mean_token_accuracy": 0.686728835105896, + "num_tokens": 56532472.0, + "step": 2195 + }, + { + "epoch": 0.24115967493960028, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 1.9514834880828857, + "learning_rate": 8.034407027818447e-07, + "loss": 1.0758, + "mean_token_accuracy": 0.6832079291343689, + "num_tokens": 56567185.0, + "step": 2196 + }, + { + "epoch": 0.24126949264221392, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0298874378204346, + "learning_rate": 8.038067349926793e-07, + "loss": 0.9843, + "mean_token_accuracy": 0.7065828442573547, + "num_tokens": 56596186.0, + "step": 2197 + }, + { + "epoch": 0.2413793103448276, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.233410596847534, + "learning_rate": 8.041727672035139e-07, + "loss": 0.968, + "mean_token_accuracy": 0.7019515037536621, + "num_tokens": 56621026.0, + "step": 2198 + }, + { + "epoch": 0.24148912804744124, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.513376235961914, + "learning_rate": 8.045387994143484e-07, + "loss": 0.9943, + "mean_token_accuracy": 0.7006656527519226, + "num_tokens": 56641149.0, + "step": 2199 + }, + { + "epoch": 0.24159894575005492, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.382636070251465, + "learning_rate": 8.04904831625183e-07, + "loss": 0.982, + "mean_token_accuracy": 0.7014245390892029, + "num_tokens": 56663743.0, + "step": 2200 + }, + { + "epoch": 0.24170876345266856, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.101655960083008, + "learning_rate": 8.052708638360175e-07, + "loss": 0.9664, + "mean_token_accuracy": 0.7156884074211121, + "num_tokens": 56690487.0, + "step": 2201 + }, + { + "epoch": 0.24181858115528224, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3679380416870117, + "learning_rate": 8.05636896046852e-07, + "loss": 0.986, + "mean_token_accuracy": 0.6957719922065735, + "num_tokens": 56714174.0, + "step": 2202 + }, + { + "epoch": 0.24192839885789588, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.22412109375, + "learning_rate": 8.060029282576867e-07, + "loss": 1.0909, + "mean_token_accuracy": 0.6775211095809937, + "num_tokens": 56743225.0, + "step": 2203 + }, + { + "epoch": 0.24203821656050956, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3351664543151855, + "learning_rate": 8.063689604685212e-07, + "loss": 1.0529, + "mean_token_accuracy": 0.6811546087265015, + "num_tokens": 56769588.0, + "step": 2204 + }, + { + "epoch": 0.2421480342631232, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3351521492004395, + "learning_rate": 8.067349926793557e-07, + "loss": 1.0169, + "mean_token_accuracy": 0.6938531398773193, + "num_tokens": 56793483.0, + "step": 2205 + }, + { + "epoch": 0.24225785196573688, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.078148126602173, + "learning_rate": 8.071010248901903e-07, + "loss": 1.0651, + "mean_token_accuracy": 0.6895890235900879, + "num_tokens": 56821789.0, + "step": 2206 + }, + { + "epoch": 0.24236766966835055, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.234175682067871, + "learning_rate": 8.074670571010248e-07, + "loss": 0.958, + "mean_token_accuracy": 0.7069905996322632, + "num_tokens": 56847788.0, + "step": 2207 + }, + { + "epoch": 0.2424774873709642, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.255460262298584, + "learning_rate": 8.078330893118594e-07, + "loss": 0.9685, + "mean_token_accuracy": 0.7085709571838379, + "num_tokens": 56874038.0, + "step": 2208 + }, + { + "epoch": 0.24258730507357787, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.488128900527954, + "learning_rate": 8.08199121522694e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6893446445465088, + "num_tokens": 56895943.0, + "step": 2209 + }, + { + "epoch": 0.24269712277619152, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.313723564147949, + "learning_rate": 8.085651537335285e-07, + "loss": 0.9848, + "mean_token_accuracy": 0.7003978490829468, + "num_tokens": 56920459.0, + "step": 2210 + }, + { + "epoch": 0.2428069404788052, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.261777639389038, + "learning_rate": 8.08931185944363e-07, + "loss": 1.1126, + "mean_token_accuracy": 0.6723490953445435, + "num_tokens": 56947307.0, + "step": 2211 + }, + { + "epoch": 0.24291675818141883, + "ewc_loss": 1.4841556549072266e-05, + "grad_norm": 2.1958255767822266, + "learning_rate": 8.092972181551976e-07, + "loss": 1.0391, + "mean_token_accuracy": 0.6891791820526123, + "num_tokens": 56974344.0, + "step": 2212 + }, + { + "epoch": 0.2430265758840325, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.0786142349243164, + "learning_rate": 8.096632503660322e-07, + "loss": 1.1612, + "mean_token_accuracy": 0.6680199503898621, + "num_tokens": 57005194.0, + "step": 2213 + }, + { + "epoch": 0.24313639358664615, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4547982215881348, + "learning_rate": 8.100292825768668e-07, + "loss": 0.9583, + "mean_token_accuracy": 0.7032166719436646, + "num_tokens": 57027359.0, + "step": 2214 + }, + { + "epoch": 0.24324621128925983, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4381649494171143, + "learning_rate": 8.103953147877013e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.700812041759491, + "num_tokens": 57050898.0, + "step": 2215 + }, + { + "epoch": 0.2433560289918735, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.347524642944336, + "learning_rate": 8.107613469985358e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.7016425132751465, + "num_tokens": 57075058.0, + "step": 2216 + }, + { + "epoch": 0.24346584669448715, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.269667625427246, + "learning_rate": 8.111273792093704e-07, + "loss": 0.9597, + "mean_token_accuracy": 0.7090820074081421, + "num_tokens": 57100386.0, + "step": 2217 + }, + { + "epoch": 0.24357566439710082, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.186284303665161, + "learning_rate": 8.114934114202049e-07, + "loss": 0.9977, + "mean_token_accuracy": 0.6971607804298401, + "num_tokens": 57126772.0, + "step": 2218 + }, + { + "epoch": 0.24368548209971447, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2769010066986084, + "learning_rate": 8.118594436310395e-07, + "loss": 1.0448, + "mean_token_accuracy": 0.6884826421737671, + "num_tokens": 57152044.0, + "step": 2219 + }, + { + "epoch": 0.24379529980232814, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2815632820129395, + "learning_rate": 8.122254758418741e-07, + "loss": 1.0458, + "mean_token_accuracy": 0.6916124224662781, + "num_tokens": 57176202.0, + "step": 2220 + }, + { + "epoch": 0.2439051175049418, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4223766326904297, + "learning_rate": 8.125915080527086e-07, + "loss": 0.9753, + "mean_token_accuracy": 0.703555703163147, + "num_tokens": 57198532.0, + "step": 2221 + }, + { + "epoch": 0.24401493520755546, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.160231828689575, + "learning_rate": 8.129575402635431e-07, + "loss": 1.03, + "mean_token_accuracy": 0.6899449825286865, + "num_tokens": 57228410.0, + "step": 2222 + }, + { + "epoch": 0.2441247529101691, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.0337319374084473, + "learning_rate": 8.133235724743777e-07, + "loss": 0.9792, + "mean_token_accuracy": 0.7040663957595825, + "num_tokens": 57257091.0, + "step": 2223 + }, + { + "epoch": 0.24423457061278278, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.272447347640991, + "learning_rate": 8.136896046852123e-07, + "loss": 0.9913, + "mean_token_accuracy": 0.6979944705963135, + "num_tokens": 57282667.0, + "step": 2224 + }, + { + "epoch": 0.24434438831539645, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4934914112091064, + "learning_rate": 8.140556368960468e-07, + "loss": 0.9398, + "mean_token_accuracy": 0.711999773979187, + "num_tokens": 57301885.0, + "step": 2225 + }, + { + "epoch": 0.2444542060180101, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5648880004882812, + "learning_rate": 8.144216691068814e-07, + "loss": 1.0279, + "mean_token_accuracy": 0.7019949555397034, + "num_tokens": 57322637.0, + "step": 2226 + }, + { + "epoch": 0.24456402372062377, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4817066192626953, + "learning_rate": 8.147877013177159e-07, + "loss": 1.0666, + "mean_token_accuracy": 0.6870372295379639, + "num_tokens": 57346366.0, + "step": 2227 + }, + { + "epoch": 0.24467384142323742, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2280709743499756, + "learning_rate": 8.151537335285504e-07, + "loss": 1.0494, + "mean_token_accuracy": 0.6956967115402222, + "num_tokens": 57373168.0, + "step": 2228 + }, + { + "epoch": 0.2447836591258511, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.1366593837738037, + "learning_rate": 8.155197657393851e-07, + "loss": 0.9953, + "mean_token_accuracy": 0.6962680816650391, + "num_tokens": 57402621.0, + "step": 2229 + }, + { + "epoch": 0.24489347682846474, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4204823970794678, + "learning_rate": 8.158857979502196e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6962364315986633, + "num_tokens": 57426230.0, + "step": 2230 + }, + { + "epoch": 0.2450032945310784, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2265889644622803, + "learning_rate": 8.162518301610542e-07, + "loss": 0.9924, + "mean_token_accuracy": 0.7030369639396667, + "num_tokens": 57453725.0, + "step": 2231 + }, + { + "epoch": 0.24511311223369206, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.072274684906006, + "learning_rate": 8.166178623718887e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6849203109741211, + "num_tokens": 57484306.0, + "step": 2232 + }, + { + "epoch": 0.24522292993630573, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2951626777648926, + "learning_rate": 8.169838945827232e-07, + "loss": 1.0406, + "mean_token_accuracy": 0.6912462115287781, + "num_tokens": 57510966.0, + "step": 2233 + }, + { + "epoch": 0.2453327476389194, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1867589950561523, + "learning_rate": 8.173499267935578e-07, + "loss": 1.0237, + "mean_token_accuracy": 0.6911948323249817, + "num_tokens": 57538533.0, + "step": 2234 + }, + { + "epoch": 0.24544256534153305, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.293710470199585, + "learning_rate": 8.177159590043924e-07, + "loss": 1.0126, + "mean_token_accuracy": 0.6921864151954651, + "num_tokens": 57564275.0, + "step": 2235 + }, + { + "epoch": 0.24555238304414673, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.020789384841919, + "learning_rate": 8.180819912152269e-07, + "loss": 1.0027, + "mean_token_accuracy": 0.7003322839736938, + "num_tokens": 57594837.0, + "step": 2236 + }, + { + "epoch": 0.24566220074676037, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.1666510105133057, + "learning_rate": 8.184480234260615e-07, + "loss": 0.9786, + "mean_token_accuracy": 0.713295578956604, + "num_tokens": 57620478.0, + "step": 2237 + }, + { + "epoch": 0.24577201844937405, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2165780067443848, + "learning_rate": 8.18814055636896e-07, + "loss": 1.0763, + "mean_token_accuracy": 0.683479368686676, + "num_tokens": 57647752.0, + "step": 2238 + }, + { + "epoch": 0.2458818361519877, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2960782051086426, + "learning_rate": 8.191800878477305e-07, + "loss": 0.9964, + "mean_token_accuracy": 0.6989092826843262, + "num_tokens": 57672591.0, + "step": 2239 + }, + { + "epoch": 0.24599165385460137, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.3039417266845703, + "learning_rate": 8.195461200585652e-07, + "loss": 0.9842, + "mean_token_accuracy": 0.7016793489456177, + "num_tokens": 57696420.0, + "step": 2240 + }, + { + "epoch": 0.246101471557215, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.3775410652160645, + "learning_rate": 8.199121522693997e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.6908932328224182, + "num_tokens": 57720386.0, + "step": 2241 + }, + { + "epoch": 0.2462112892598287, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.505080223083496, + "learning_rate": 8.202781844802342e-07, + "loss": 1.033, + "mean_token_accuracy": 0.6946554780006409, + "num_tokens": 57746173.0, + "step": 2242 + }, + { + "epoch": 0.24632110696244233, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3693203926086426, + "learning_rate": 8.206442166910688e-07, + "loss": 1.0286, + "mean_token_accuracy": 0.6913198232650757, + "num_tokens": 57770085.0, + "step": 2243 + }, + { + "epoch": 0.246430924665056, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.0373663902282715, + "learning_rate": 8.210102489019033e-07, + "loss": 0.9414, + "mean_token_accuracy": 0.7103884220123291, + "num_tokens": 57799026.0, + "step": 2244 + }, + { + "epoch": 0.24654074236766968, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2458834648132324, + "learning_rate": 8.213762811127379e-07, + "loss": 1.0175, + "mean_token_accuracy": 0.6929317116737366, + "num_tokens": 57824626.0, + "step": 2245 + }, + { + "epoch": 0.24665056007028333, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4997549057006836, + "learning_rate": 8.217423133235725e-07, + "loss": 1.0376, + "mean_token_accuracy": 0.6822088360786438, + "num_tokens": 57847648.0, + "step": 2246 + }, + { + "epoch": 0.246760377772897, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3721048831939697, + "learning_rate": 8.22108345534407e-07, + "loss": 1.0334, + "mean_token_accuracy": 0.6875414848327637, + "num_tokens": 57870239.0, + "step": 2247 + }, + { + "epoch": 0.24687019547551065, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.102384567260742, + "learning_rate": 8.224743777452416e-07, + "loss": 1.008, + "mean_token_accuracy": 0.6934018135070801, + "num_tokens": 57899702.0, + "step": 2248 + }, + { + "epoch": 0.24698001317812432, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.7758634090423584, + "learning_rate": 8.228404099560761e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.704525351524353, + "num_tokens": 57918400.0, + "step": 2249 + }, + { + "epoch": 0.24708983088073797, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.419046640396118, + "learning_rate": 8.232064421669106e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.693164587020874, + "num_tokens": 57940924.0, + "step": 2250 + }, + { + "epoch": 0.24719964858335164, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.683478593826294, + "learning_rate": 8.235724743777453e-07, + "loss": 0.9385, + "mean_token_accuracy": 0.7210991382598877, + "num_tokens": 57959917.0, + "step": 2251 + }, + { + "epoch": 0.24730946628596528, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.073145866394043, + "learning_rate": 8.239385065885798e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.689879298210144, + "num_tokens": 57989935.0, + "step": 2252 + }, + { + "epoch": 0.24741928398857896, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.612642526626587, + "learning_rate": 8.243045387994143e-07, + "loss": 0.9706, + "mean_token_accuracy": 0.7121334671974182, + "num_tokens": 58009541.0, + "step": 2253 + }, + { + "epoch": 0.24752910169119263, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 1.9915684461593628, + "learning_rate": 8.246705710102489e-07, + "loss": 1.0109, + "mean_token_accuracy": 0.6913092732429504, + "num_tokens": 58042087.0, + "step": 2254 + }, + { + "epoch": 0.24763891939380628, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.156081438064575, + "learning_rate": 8.250366032210834e-07, + "loss": 0.9614, + "mean_token_accuracy": 0.7149232029914856, + "num_tokens": 58070764.0, + "step": 2255 + }, + { + "epoch": 0.24774873709641995, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6932618618011475, + "learning_rate": 8.25402635431918e-07, + "loss": 0.9781, + "mean_token_accuracy": 0.7046912908554077, + "num_tokens": 58090387.0, + "step": 2256 + }, + { + "epoch": 0.2478585547990336, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3822593688964844, + "learning_rate": 8.257686676427526e-07, + "loss": 1.0654, + "mean_token_accuracy": 0.6870325207710266, + "num_tokens": 58113755.0, + "step": 2257 + }, + { + "epoch": 0.24796837250164727, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4876351356506348, + "learning_rate": 8.261346998535871e-07, + "loss": 0.9437, + "mean_token_accuracy": 0.7165980935096741, + "num_tokens": 58135562.0, + "step": 2258 + }, + { + "epoch": 0.24807819020426092, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.293689012527466, + "learning_rate": 8.265007320644216e-07, + "loss": 1.0531, + "mean_token_accuracy": 0.6837098598480225, + "num_tokens": 58161079.0, + "step": 2259 + }, + { + "epoch": 0.2481880079068746, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2790610790252686, + "learning_rate": 8.268667642752562e-07, + "loss": 0.9477, + "mean_token_accuracy": 0.7189596891403198, + "num_tokens": 58186104.0, + "step": 2260 + }, + { + "epoch": 0.24829782560948824, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 1.9712612628936768, + "learning_rate": 8.272327964860908e-07, + "loss": 1.0841, + "mean_token_accuracy": 0.6785820722579956, + "num_tokens": 58221477.0, + "step": 2261 + }, + { + "epoch": 0.2484076433121019, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.268777847290039, + "learning_rate": 8.275988286969253e-07, + "loss": 1.069, + "mean_token_accuracy": 0.6832952499389648, + "num_tokens": 58247976.0, + "step": 2262 + }, + { + "epoch": 0.24851746101471558, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5847086906433105, + "learning_rate": 8.279648609077599e-07, + "loss": 0.9546, + "mean_token_accuracy": 0.7123207449913025, + "num_tokens": 58267331.0, + "step": 2263 + }, + { + "epoch": 0.24862727871732923, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3171002864837646, + "learning_rate": 8.283308931185944e-07, + "loss": 1.0451, + "mean_token_accuracy": 0.6826906204223633, + "num_tokens": 58292582.0, + "step": 2264 + }, + { + "epoch": 0.2487370964199429, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2850182056427, + "learning_rate": 8.28696925329429e-07, + "loss": 1.1327, + "mean_token_accuracy": 0.6622159481048584, + "num_tokens": 58318884.0, + "step": 2265 + }, + { + "epoch": 0.24884691412255655, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.310375690460205, + "learning_rate": 8.290629575402635e-07, + "loss": 1.0185, + "mean_token_accuracy": 0.6911827325820923, + "num_tokens": 58344826.0, + "step": 2266 + }, + { + "epoch": 0.24895673182517022, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.313321828842163, + "learning_rate": 8.294289897510981e-07, + "loss": 1.0248, + "mean_token_accuracy": 0.6911145448684692, + "num_tokens": 58368859.0, + "step": 2267 + }, + { + "epoch": 0.24906654952778387, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2329680919647217, + "learning_rate": 8.297950219619327e-07, + "loss": 0.9952, + "mean_token_accuracy": 0.7039022445678711, + "num_tokens": 58397377.0, + "step": 2268 + }, + { + "epoch": 0.24917636723039754, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3445377349853516, + "learning_rate": 8.301610541727672e-07, + "loss": 0.8887, + "mean_token_accuracy": 0.7307580709457397, + "num_tokens": 58419411.0, + "step": 2269 + }, + { + "epoch": 0.2492861849330112, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1125943660736084, + "learning_rate": 8.305270863836017e-07, + "loss": 0.9753, + "mean_token_accuracy": 0.7013641595840454, + "num_tokens": 58447935.0, + "step": 2270 + }, + { + "epoch": 0.24939600263562486, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1322736740112305, + "learning_rate": 8.308931185944363e-07, + "loss": 0.9963, + "mean_token_accuracy": 0.6997091770172119, + "num_tokens": 58474769.0, + "step": 2271 + }, + { + "epoch": 0.24950582033823854, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.26194167137146, + "learning_rate": 8.312591508052709e-07, + "loss": 1.1377, + "mean_token_accuracy": 0.6623278856277466, + "num_tokens": 58504789.0, + "step": 2272 + }, + { + "epoch": 0.24961563804085218, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.602902889251709, + "learning_rate": 8.316251830161054e-07, + "loss": 0.9792, + "mean_token_accuracy": 0.7012861371040344, + "num_tokens": 58526700.0, + "step": 2273 + }, + { + "epoch": 0.24972545574346586, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.075974225997925, + "learning_rate": 8.3199121522694e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6858968734741211, + "num_tokens": 58555555.0, + "step": 2274 + }, + { + "epoch": 0.2498352734460795, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.341660261154175, + "learning_rate": 8.323572474377745e-07, + "loss": 0.8659, + "mean_token_accuracy": 0.7347883582115173, + "num_tokens": 58577667.0, + "step": 2275 + }, + { + "epoch": 0.24994509114869318, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.095801591873169, + "learning_rate": 8.32723279648609e-07, + "loss": 1.0379, + "mean_token_accuracy": 0.6920666098594666, + "num_tokens": 58606031.0, + "step": 2276 + }, + { + "epoch": 0.2500549088513068, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.352566719055176, + "learning_rate": 8.330893118594437e-07, + "loss": 0.9765, + "mean_token_accuracy": 0.7040821313858032, + "num_tokens": 58631967.0, + "step": 2277 + }, + { + "epoch": 0.25016472655392047, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.104326009750366, + "learning_rate": 8.334553440702782e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6862949132919312, + "num_tokens": 58662429.0, + "step": 2278 + }, + { + "epoch": 0.25027454425653417, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.250770330429077, + "learning_rate": 8.338213762811127e-07, + "loss": 1.0728, + "mean_token_accuracy": 0.6736264228820801, + "num_tokens": 58689786.0, + "step": 2279 + }, + { + "epoch": 0.2503843619591478, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3977227210998535, + "learning_rate": 8.341874084919473e-07, + "loss": 0.9772, + "mean_token_accuracy": 0.703896164894104, + "num_tokens": 58711364.0, + "step": 2280 + }, + { + "epoch": 0.25049417966176146, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.173560619354248, + "learning_rate": 8.345534407027818e-07, + "loss": 1.0433, + "mean_token_accuracy": 0.6880775094032288, + "num_tokens": 58737840.0, + "step": 2281 + }, + { + "epoch": 0.25060399736437516, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1383867263793945, + "learning_rate": 8.349194729136164e-07, + "loss": 1.0957, + "mean_token_accuracy": 0.6792306900024414, + "num_tokens": 58766062.0, + "step": 2282 + }, + { + "epoch": 0.2507138150669888, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.274538040161133, + "learning_rate": 8.35285505124451e-07, + "loss": 1.0352, + "mean_token_accuracy": 0.6847199201583862, + "num_tokens": 58792625.0, + "step": 2283 + }, + { + "epoch": 0.25082363276960246, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.397373914718628, + "learning_rate": 8.356515373352855e-07, + "loss": 1.1569, + "mean_token_accuracy": 0.6623319983482361, + "num_tokens": 58817547.0, + "step": 2284 + }, + { + "epoch": 0.2509334504722161, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.0882866382598877, + "learning_rate": 8.360175695461201e-07, + "loss": 1.0347, + "mean_token_accuracy": 0.6884901523590088, + "num_tokens": 58845967.0, + "step": 2285 + }, + { + "epoch": 0.2510432681748298, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0593252182006836, + "learning_rate": 8.363836017569546e-07, + "loss": 1.0223, + "mean_token_accuracy": 0.6960828900337219, + "num_tokens": 58876861.0, + "step": 2286 + }, + { + "epoch": 0.25115308587744345, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3531837463378906, + "learning_rate": 8.36749633967789e-07, + "loss": 1.0427, + "mean_token_accuracy": 0.6877244710922241, + "num_tokens": 58900848.0, + "step": 2287 + }, + { + "epoch": 0.2512629035800571, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.320322275161743, + "learning_rate": 8.371156661786238e-07, + "loss": 0.9791, + "mean_token_accuracy": 0.7039983868598938, + "num_tokens": 58922927.0, + "step": 2288 + }, + { + "epoch": 0.25137272128267074, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1399664878845215, + "learning_rate": 8.374816983894583e-07, + "loss": 1.0202, + "mean_token_accuracy": 0.6877932548522949, + "num_tokens": 58951123.0, + "step": 2289 + }, + { + "epoch": 0.25148253898528444, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5704562664031982, + "learning_rate": 8.378477306002928e-07, + "loss": 1.0499, + "mean_token_accuracy": 0.685348629951477, + "num_tokens": 58970496.0, + "step": 2290 + }, + { + "epoch": 0.2515923566878981, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2423887252807617, + "learning_rate": 8.382137628111274e-07, + "loss": 1.0432, + "mean_token_accuracy": 0.6919306516647339, + "num_tokens": 58994989.0, + "step": 2291 + }, + { + "epoch": 0.25170217439051173, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5615501403808594, + "learning_rate": 8.385797950219619e-07, + "loss": 0.9675, + "mean_token_accuracy": 0.7060415744781494, + "num_tokens": 59015235.0, + "step": 2292 + }, + { + "epoch": 0.25181199209312544, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3243491649627686, + "learning_rate": 8.389458272327965e-07, + "loss": 0.9756, + "mean_token_accuracy": 0.6994485259056091, + "num_tokens": 59038421.0, + "step": 2293 + }, + { + "epoch": 0.2519218097957391, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2715556621551514, + "learning_rate": 8.393118594436311e-07, + "loss": 1.0663, + "mean_token_accuracy": 0.6748558878898621, + "num_tokens": 59064410.0, + "step": 2294 + }, + { + "epoch": 0.25203162749835273, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.403348207473755, + "learning_rate": 8.396778916544656e-07, + "loss": 1.0382, + "mean_token_accuracy": 0.6898648738861084, + "num_tokens": 59087245.0, + "step": 2295 + }, + { + "epoch": 0.2521414452009664, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3120229244232178, + "learning_rate": 8.400439238653e-07, + "loss": 1.0265, + "mean_token_accuracy": 0.6896387338638306, + "num_tokens": 59111988.0, + "step": 2296 + }, + { + "epoch": 0.2522512629035801, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4680306911468506, + "learning_rate": 8.404099560761346e-07, + "loss": 1.0254, + "mean_token_accuracy": 0.6983803510665894, + "num_tokens": 59137371.0, + "step": 2297 + }, + { + "epoch": 0.2523610806061937, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.300724983215332, + "learning_rate": 8.407759882869691e-07, + "loss": 1.0202, + "mean_token_accuracy": 0.6965954899787903, + "num_tokens": 59163108.0, + "step": 2298 + }, + { + "epoch": 0.25247089830880737, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.203341484069824, + "learning_rate": 8.411420204978039e-07, + "loss": 1.0762, + "mean_token_accuracy": 0.676612377166748, + "num_tokens": 59191057.0, + "step": 2299 + }, + { + "epoch": 0.252580716011421, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.11928129196167, + "learning_rate": 8.415080527086383e-07, + "loss": 1.084, + "mean_token_accuracy": 0.6783764362335205, + "num_tokens": 59220690.0, + "step": 2300 + }, + { + "epoch": 0.2526905337140347, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1551830768585205, + "learning_rate": 8.418740849194728e-07, + "loss": 0.9867, + "mean_token_accuracy": 0.7056483626365662, + "num_tokens": 59247310.0, + "step": 2301 + }, + { + "epoch": 0.25280035141664836, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3232860565185547, + "learning_rate": 8.422401171303074e-07, + "loss": 1.0043, + "mean_token_accuracy": 0.7015058994293213, + "num_tokens": 59271797.0, + "step": 2302 + }, + { + "epoch": 0.252910169119262, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2403693199157715, + "learning_rate": 8.426061493411419e-07, + "loss": 0.9946, + "mean_token_accuracy": 0.7005903720855713, + "num_tokens": 59299996.0, + "step": 2303 + }, + { + "epoch": 0.2530199868218757, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.217536449432373, + "learning_rate": 8.429721815519765e-07, + "loss": 1.0747, + "mean_token_accuracy": 0.6815263032913208, + "num_tokens": 59326639.0, + "step": 2304 + }, + { + "epoch": 0.25312980452448935, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2935221195220947, + "learning_rate": 8.433382137628111e-07, + "loss": 1.0082, + "mean_token_accuracy": 0.7034334540367126, + "num_tokens": 59353176.0, + "step": 2305 + }, + { + "epoch": 0.253239622227103, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0921311378479004, + "learning_rate": 8.437042459736456e-07, + "loss": 1.0146, + "mean_token_accuracy": 0.6984748840332031, + "num_tokens": 59381590.0, + "step": 2306 + }, + { + "epoch": 0.25334943992971665, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.966269850730896, + "learning_rate": 8.440702781844801e-07, + "loss": 0.9777, + "mean_token_accuracy": 0.7049823999404907, + "num_tokens": 59411696.0, + "step": 2307 + }, + { + "epoch": 0.25345925763233035, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2085282802581787, + "learning_rate": 8.444363103953147e-07, + "loss": 1.0503, + "mean_token_accuracy": 0.6809872984886169, + "num_tokens": 59438691.0, + "step": 2308 + }, + { + "epoch": 0.253569075334944, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0960869789123535, + "learning_rate": 8.448023426061493e-07, + "loss": 1.0062, + "mean_token_accuracy": 0.6965578198432922, + "num_tokens": 59467636.0, + "step": 2309 + }, + { + "epoch": 0.25367889303755764, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0763611793518066, + "learning_rate": 8.451683748169838e-07, + "loss": 1.0422, + "mean_token_accuracy": 0.6900894045829773, + "num_tokens": 59496548.0, + "step": 2310 + }, + { + "epoch": 0.25378871074017134, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.453455686569214, + "learning_rate": 8.455344070278184e-07, + "loss": 1.0396, + "mean_token_accuracy": 0.6897817850112915, + "num_tokens": 59518194.0, + "step": 2311 + }, + { + "epoch": 0.253898528442785, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.463409423828125, + "learning_rate": 8.459004392386529e-07, + "loss": 0.9092, + "mean_token_accuracy": 0.7207636833190918, + "num_tokens": 59540317.0, + "step": 2312 + }, + { + "epoch": 0.25400834614539863, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.23038387298584, + "learning_rate": 8.462664714494874e-07, + "loss": 1.0753, + "mean_token_accuracy": 0.6748436689376831, + "num_tokens": 59569101.0, + "step": 2313 + }, + { + "epoch": 0.2541181638480123, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2378931045532227, + "learning_rate": 8.46632503660322e-07, + "loss": 1.0653, + "mean_token_accuracy": 0.6790201663970947, + "num_tokens": 59597709.0, + "step": 2314 + }, + { + "epoch": 0.254227981550626, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0956602096557617, + "learning_rate": 8.469985358711566e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6937670707702637, + "num_tokens": 59626571.0, + "step": 2315 + }, + { + "epoch": 0.2543377992532396, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2664763927459717, + "learning_rate": 8.473645680819912e-07, + "loss": 1.02, + "mean_token_accuracy": 0.6852641105651855, + "num_tokens": 59653507.0, + "step": 2316 + }, + { + "epoch": 0.2544476169558533, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2877440452575684, + "learning_rate": 8.477306002928257e-07, + "loss": 0.9829, + "mean_token_accuracy": 0.7102741003036499, + "num_tokens": 59682791.0, + "step": 2317 + }, + { + "epoch": 0.2545574346584669, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3861770629882812, + "learning_rate": 8.480966325036602e-07, + "loss": 1.1161, + "mean_token_accuracy": 0.6687558889389038, + "num_tokens": 59708535.0, + "step": 2318 + }, + { + "epoch": 0.2546672523610806, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.7495429515838623, + "learning_rate": 8.484626647144948e-07, + "loss": 0.9833, + "mean_token_accuracy": 0.7042183876037598, + "num_tokens": 59728949.0, + "step": 2319 + }, + { + "epoch": 0.25477707006369427, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4151999950408936, + "learning_rate": 8.488286969253294e-07, + "loss": 0.9763, + "mean_token_accuracy": 0.7091882228851318, + "num_tokens": 59750730.0, + "step": 2320 + }, + { + "epoch": 0.2548868877663079, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3558883666992188, + "learning_rate": 8.491947291361639e-07, + "loss": 1.0076, + "mean_token_accuracy": 0.6988022327423096, + "num_tokens": 59773449.0, + "step": 2321 + }, + { + "epoch": 0.2549967054689216, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2962305545806885, + "learning_rate": 8.495607613469985e-07, + "loss": 1.0487, + "mean_token_accuracy": 0.6834295988082886, + "num_tokens": 59797987.0, + "step": 2322 + }, + { + "epoch": 0.25510652317153526, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4877429008483887, + "learning_rate": 8.49926793557833e-07, + "loss": 0.9685, + "mean_token_accuracy": 0.7005395889282227, + "num_tokens": 59819138.0, + "step": 2323 + }, + { + "epoch": 0.2552163408741489, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4211745262145996, + "learning_rate": 8.502928257686675e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6903796195983887, + "num_tokens": 59842604.0, + "step": 2324 + }, + { + "epoch": 0.25532615857676255, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2223060131073, + "learning_rate": 8.506588579795022e-07, + "loss": 1.0293, + "mean_token_accuracy": 0.6918303370475769, + "num_tokens": 59869413.0, + "step": 2325 + }, + { + "epoch": 0.25543597627937625, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2546005249023438, + "learning_rate": 8.510248901903367e-07, + "loss": 1.042, + "mean_token_accuracy": 0.6863613724708557, + "num_tokens": 59896911.0, + "step": 2326 + }, + { + "epoch": 0.2555457939819899, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3472776412963867, + "learning_rate": 8.513909224011712e-07, + "loss": 0.9707, + "mean_token_accuracy": 0.7049030065536499, + "num_tokens": 59920209.0, + "step": 2327 + }, + { + "epoch": 0.25565561168460355, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2931036949157715, + "learning_rate": 8.517569546120058e-07, + "loss": 1.0602, + "mean_token_accuracy": 0.6826980113983154, + "num_tokens": 59946626.0, + "step": 2328 + }, + { + "epoch": 0.25576542938721725, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.13246750831604, + "learning_rate": 8.521229868228403e-07, + "loss": 1.056, + "mean_token_accuracy": 0.6891007423400879, + "num_tokens": 59975887.0, + "step": 2329 + }, + { + "epoch": 0.2558752470898309, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2592601776123047, + "learning_rate": 8.524890190336748e-07, + "loss": 0.9426, + "mean_token_accuracy": 0.7136672735214233, + "num_tokens": 60000691.0, + "step": 2330 + }, + { + "epoch": 0.25598506479244454, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.299298048019409, + "learning_rate": 8.528550512445095e-07, + "loss": 1.0187, + "mean_token_accuracy": 0.6964182257652283, + "num_tokens": 60025315.0, + "step": 2331 + }, + { + "epoch": 0.2560948824950582, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1142923831939697, + "learning_rate": 8.53221083455344e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6867493391036987, + "num_tokens": 60057120.0, + "step": 2332 + }, + { + "epoch": 0.2562047001976719, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1865789890289307, + "learning_rate": 8.535871156661786e-07, + "loss": 1.0233, + "mean_token_accuracy": 0.7000399827957153, + "num_tokens": 60084156.0, + "step": 2333 + }, + { + "epoch": 0.25631451790028553, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3363711833953857, + "learning_rate": 8.539531478770131e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6803444623947144, + "num_tokens": 60109948.0, + "step": 2334 + }, + { + "epoch": 0.2564243356028992, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6249752044677734, + "learning_rate": 8.543191800878476e-07, + "loss": 0.9363, + "mean_token_accuracy": 0.7090480327606201, + "num_tokens": 60129571.0, + "step": 2335 + }, + { + "epoch": 0.2565341533055128, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2061877250671387, + "learning_rate": 8.546852122986823e-07, + "loss": 1.0563, + "mean_token_accuracy": 0.683559775352478, + "num_tokens": 60157568.0, + "step": 2336 + }, + { + "epoch": 0.2566439710081265, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2765591144561768, + "learning_rate": 8.550512445095168e-07, + "loss": 1.0002, + "mean_token_accuracy": 0.6968024969100952, + "num_tokens": 60185034.0, + "step": 2337 + }, + { + "epoch": 0.25675378871074017, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 1.9263648986816406, + "learning_rate": 8.554172767203513e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.6938941478729248, + "num_tokens": 60218705.0, + "step": 2338 + }, + { + "epoch": 0.2568636064133538, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.271101236343384, + "learning_rate": 8.557833089311859e-07, + "loss": 0.9403, + "mean_token_accuracy": 0.717025101184845, + "num_tokens": 60243522.0, + "step": 2339 + }, + { + "epoch": 0.2569734241159675, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2090377807617188, + "learning_rate": 8.561493411420204e-07, + "loss": 1.0095, + "mean_token_accuracy": 0.6999465823173523, + "num_tokens": 60271606.0, + "step": 2340 + }, + { + "epoch": 0.25708324181858117, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1069998741149902, + "learning_rate": 8.56515373352855e-07, + "loss": 1.0353, + "mean_token_accuracy": 0.6932970881462097, + "num_tokens": 60300332.0, + "step": 2341 + }, + { + "epoch": 0.2571930595211948, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.160475969314575, + "learning_rate": 8.568814055636896e-07, + "loss": 0.8976, + "mean_token_accuracy": 0.727809488773346, + "num_tokens": 60328449.0, + "step": 2342 + }, + { + "epoch": 0.25730287722380846, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1809897422790527, + "learning_rate": 8.572474377745241e-07, + "loss": 1.0119, + "mean_token_accuracy": 0.6994061470031738, + "num_tokens": 60354910.0, + "step": 2343 + }, + { + "epoch": 0.25741269492642216, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2036232948303223, + "learning_rate": 8.576134699853586e-07, + "loss": 0.9916, + "mean_token_accuracy": 0.7040831446647644, + "num_tokens": 60380063.0, + "step": 2344 + }, + { + "epoch": 0.2575225126290358, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.381114959716797, + "learning_rate": 8.579795021961932e-07, + "loss": 1.0481, + "mean_token_accuracy": 0.6880713701248169, + "num_tokens": 60404706.0, + "step": 2345 + }, + { + "epoch": 0.25763233033164945, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1552302837371826, + "learning_rate": 8.583455344070277e-07, + "loss": 0.9971, + "mean_token_accuracy": 0.7005056738853455, + "num_tokens": 60433576.0, + "step": 2346 + }, + { + "epoch": 0.2577421480342631, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1881847381591797, + "learning_rate": 8.587115666178624e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.6973608732223511, + "num_tokens": 60458975.0, + "step": 2347 + }, + { + "epoch": 0.2578519657368768, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1000847816467285, + "learning_rate": 8.590775988286969e-07, + "loss": 1.0176, + "mean_token_accuracy": 0.7042943239212036, + "num_tokens": 60490043.0, + "step": 2348 + }, + { + "epoch": 0.25796178343949044, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3486061096191406, + "learning_rate": 8.594436310395314e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.6879854798316956, + "num_tokens": 60512267.0, + "step": 2349 + }, + { + "epoch": 0.2580716011421041, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.252997398376465, + "learning_rate": 8.59809663250366e-07, + "loss": 1.0464, + "mean_token_accuracy": 0.6900942921638489, + "num_tokens": 60537869.0, + "step": 2350 + }, + { + "epoch": 0.2581814188447178, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2349698543548584, + "learning_rate": 8.601756954612005e-07, + "loss": 1.0075, + "mean_token_accuracy": 0.6988993287086487, + "num_tokens": 60563603.0, + "step": 2351 + }, + { + "epoch": 0.25829123654733144, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.410919666290283, + "learning_rate": 8.605417276720351e-07, + "loss": 1.014, + "mean_token_accuracy": 0.6929965615272522, + "num_tokens": 60587192.0, + "step": 2352 + }, + { + "epoch": 0.2584010542499451, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.361361503601074, + "learning_rate": 8.609077598828697e-07, + "loss": 0.9635, + "mean_token_accuracy": 0.7026478052139282, + "num_tokens": 60610638.0, + "step": 2353 + }, + { + "epoch": 0.25851087195255873, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3497488498687744, + "learning_rate": 8.612737920937042e-07, + "loss": 1.0701, + "mean_token_accuracy": 0.6881793737411499, + "num_tokens": 60636598.0, + "step": 2354 + }, + { + "epoch": 0.25862068965517243, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0739452838897705, + "learning_rate": 8.616398243045387e-07, + "loss": 1.061, + "mean_token_accuracy": 0.6862406134605408, + "num_tokens": 60666496.0, + "step": 2355 + }, + { + "epoch": 0.2587305073577861, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0232744216918945, + "learning_rate": 8.620058565153733e-07, + "loss": 1.067, + "mean_token_accuracy": 0.6826609373092651, + "num_tokens": 60696995.0, + "step": 2356 + }, + { + "epoch": 0.2588403250603997, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.556108236312866, + "learning_rate": 8.623718887262079e-07, + "loss": 0.9618, + "mean_token_accuracy": 0.7128099799156189, + "num_tokens": 60717345.0, + "step": 2357 + }, + { + "epoch": 0.2589501427630134, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.28423810005188, + "learning_rate": 8.627379209370424e-07, + "loss": 1.0561, + "mean_token_accuracy": 0.6926653385162354, + "num_tokens": 60744639.0, + "step": 2358 + }, + { + "epoch": 0.25905996046562707, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0898334980010986, + "learning_rate": 8.63103953147877e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6824951767921448, + "num_tokens": 60773675.0, + "step": 2359 + }, + { + "epoch": 0.2591697781682407, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.19120192527771, + "learning_rate": 8.634699853587115e-07, + "loss": 1.011, + "mean_token_accuracy": 0.6980423927307129, + "num_tokens": 60799408.0, + "step": 2360 + }, + { + "epoch": 0.25927959587085436, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.393033981323242, + "learning_rate": 8.63836017569546e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.6838929653167725, + "num_tokens": 60823031.0, + "step": 2361 + }, + { + "epoch": 0.25938941357346806, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1404857635498047, + "learning_rate": 8.642020497803806e-07, + "loss": 1.0385, + "mean_token_accuracy": 0.6932716369628906, + "num_tokens": 60851807.0, + "step": 2362 + }, + { + "epoch": 0.2594992312760817, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.8682894706726074, + "learning_rate": 8.645680819912152e-07, + "loss": 1.0643, + "mean_token_accuracy": 0.6868191361427307, + "num_tokens": 60871805.0, + "step": 2363 + }, + { + "epoch": 0.25960904897869536, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 1.9349806308746338, + "learning_rate": 8.649341142020498e-07, + "loss": 1.1608, + "mean_token_accuracy": 0.6566181182861328, + "num_tokens": 60909599.0, + "step": 2364 + }, + { + "epoch": 0.259718866681309, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2701117992401123, + "learning_rate": 8.653001464128843e-07, + "loss": 1.0784, + "mean_token_accuracy": 0.6838812828063965, + "num_tokens": 60936870.0, + "step": 2365 + }, + { + "epoch": 0.2598286843839227, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.387226104736328, + "learning_rate": 8.656661786237188e-07, + "loss": 1.0091, + "mean_token_accuracy": 0.6967824697494507, + "num_tokens": 60958742.0, + "step": 2366 + }, + { + "epoch": 0.25993850208653635, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0862393379211426, + "learning_rate": 8.660322108345534e-07, + "loss": 1.0303, + "mean_token_accuracy": 0.6968276500701904, + "num_tokens": 60987581.0, + "step": 2367 + }, + { + "epoch": 0.26004831978915, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.169813871383667, + "learning_rate": 8.66398243045388e-07, + "loss": 1.0701, + "mean_token_accuracy": 0.6816694140434265, + "num_tokens": 61015388.0, + "step": 2368 + }, + { + "epoch": 0.2601581374917637, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0286834239959717, + "learning_rate": 8.667642752562225e-07, + "loss": 1.059, + "mean_token_accuracy": 0.6849969625473022, + "num_tokens": 61046490.0, + "step": 2369 + }, + { + "epoch": 0.26026795519437734, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.140293836593628, + "learning_rate": 8.671303074670571e-07, + "loss": 1.1076, + "mean_token_accuracy": 0.672986626625061, + "num_tokens": 61076806.0, + "step": 2370 + }, + { + "epoch": 0.260377772896991, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 1.9913954734802246, + "learning_rate": 8.674963396778916e-07, + "loss": 1.1102, + "mean_token_accuracy": 0.6774970889091492, + "num_tokens": 61109820.0, + "step": 2371 + }, + { + "epoch": 0.26048759059960463, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5184779167175293, + "learning_rate": 8.678623718887261e-07, + "loss": 0.9922, + "mean_token_accuracy": 0.6980130672454834, + "num_tokens": 61131765.0, + "step": 2372 + }, + { + "epoch": 0.26059740830221834, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0986416339874268, + "learning_rate": 8.682284040995608e-07, + "loss": 1.0396, + "mean_token_accuracy": 0.699774444103241, + "num_tokens": 61160748.0, + "step": 2373 + }, + { + "epoch": 0.260707226004832, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.080643892288208, + "learning_rate": 8.685944363103953e-07, + "loss": 0.9983, + "mean_token_accuracy": 0.7032941579818726, + "num_tokens": 61189525.0, + "step": 2374 + }, + { + "epoch": 0.26081704370744563, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2444355487823486, + "learning_rate": 8.689604685212298e-07, + "loss": 1.0356, + "mean_token_accuracy": 0.6950685381889343, + "num_tokens": 61217544.0, + "step": 2375 + }, + { + "epoch": 0.2609268614100593, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2050883769989014, + "learning_rate": 8.693265007320644e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6942716836929321, + "num_tokens": 61244972.0, + "step": 2376 + }, + { + "epoch": 0.261036679112673, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.218918561935425, + "learning_rate": 8.696925329428989e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6725341081619263, + "num_tokens": 61272586.0, + "step": 2377 + }, + { + "epoch": 0.2611464968152866, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.434204339981079, + "learning_rate": 8.700585651537334e-07, + "loss": 0.892, + "mean_token_accuracy": 0.7243863344192505, + "num_tokens": 61294173.0, + "step": 2378 + }, + { + "epoch": 0.26125631451790027, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1458799839019775, + "learning_rate": 8.704245973645681e-07, + "loss": 1.0429, + "mean_token_accuracy": 0.6893701553344727, + "num_tokens": 61321813.0, + "step": 2379 + }, + { + "epoch": 0.26136613222051397, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1260483264923096, + "learning_rate": 8.707906295754026e-07, + "loss": 1.1588, + "mean_token_accuracy": 0.6584691405296326, + "num_tokens": 61351003.0, + "step": 2380 + }, + { + "epoch": 0.2614759499231276, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.276646852493286, + "learning_rate": 8.711566617862372e-07, + "loss": 1.0805, + "mean_token_accuracy": 0.6827384829521179, + "num_tokens": 61376842.0, + "step": 2381 + }, + { + "epoch": 0.26158576762574126, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 1.967249870300293, + "learning_rate": 8.715226939970717e-07, + "loss": 1.0196, + "mean_token_accuracy": 0.6954066753387451, + "num_tokens": 61409980.0, + "step": 2382 + }, + { + "epoch": 0.2616955853283549, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.010324239730835, + "learning_rate": 8.718887262079062e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6889176368713379, + "num_tokens": 61442720.0, + "step": 2383 + }, + { + "epoch": 0.2618054030309686, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.359248638153076, + "learning_rate": 8.722547584187409e-07, + "loss": 0.9965, + "mean_token_accuracy": 0.7030735611915588, + "num_tokens": 61468322.0, + "step": 2384 + }, + { + "epoch": 0.26191522073358225, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4787042140960693, + "learning_rate": 8.726207906295754e-07, + "loss": 1.0247, + "mean_token_accuracy": 0.6862711906433105, + "num_tokens": 61490210.0, + "step": 2385 + }, + { + "epoch": 0.2620250384361959, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3971283435821533, + "learning_rate": 8.729868228404099e-07, + "loss": 1.0242, + "mean_token_accuracy": 0.6893548965454102, + "num_tokens": 61513905.0, + "step": 2386 + }, + { + "epoch": 0.2621348561388096, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5305466651916504, + "learning_rate": 8.733528550512445e-07, + "loss": 1.0131, + "mean_token_accuracy": 0.6919686198234558, + "num_tokens": 61533984.0, + "step": 2387 + }, + { + "epoch": 0.26224467384142325, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.169994831085205, + "learning_rate": 8.73718887262079e-07, + "loss": 0.9785, + "mean_token_accuracy": 0.7073692679405212, + "num_tokens": 61559133.0, + "step": 2388 + }, + { + "epoch": 0.2623544915440369, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2741219997406006, + "learning_rate": 8.740849194729136e-07, + "loss": 0.9982, + "mean_token_accuracy": 0.7008668184280396, + "num_tokens": 61583791.0, + "step": 2389 + }, + { + "epoch": 0.26246430924665054, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.570323944091797, + "learning_rate": 8.744509516837482e-07, + "loss": 0.9776, + "mean_token_accuracy": 0.7068204879760742, + "num_tokens": 61604249.0, + "step": 2390 + }, + { + "epoch": 0.26257412694926424, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.460852861404419, + "learning_rate": 8.748169838945827e-07, + "loss": 0.9848, + "mean_token_accuracy": 0.7068883180618286, + "num_tokens": 61627052.0, + "step": 2391 + }, + { + "epoch": 0.2626839446518779, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.147792100906372, + "learning_rate": 8.751830161054172e-07, + "loss": 1.0858, + "mean_token_accuracy": 0.6838252544403076, + "num_tokens": 61655979.0, + "step": 2392 + }, + { + "epoch": 0.26279376235449153, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.212651252746582, + "learning_rate": 8.755490483162518e-07, + "loss": 1.1039, + "mean_token_accuracy": 0.669238805770874, + "num_tokens": 61684326.0, + "step": 2393 + }, + { + "epoch": 0.2629035800571052, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3253285884857178, + "learning_rate": 8.759150805270863e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6797685623168945, + "num_tokens": 61712239.0, + "step": 2394 + }, + { + "epoch": 0.2630133977597189, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.159050703048706, + "learning_rate": 8.76281112737921e-07, + "loss": 1.0425, + "mean_token_accuracy": 0.6888074278831482, + "num_tokens": 61740619.0, + "step": 2395 + }, + { + "epoch": 0.2631232154623325, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2964510917663574, + "learning_rate": 8.766471449487555e-07, + "loss": 1.0622, + "mean_token_accuracy": 0.6798510551452637, + "num_tokens": 61766457.0, + "step": 2396 + }, + { + "epoch": 0.2632330331649462, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2985033988952637, + "learning_rate": 8.7701317715959e-07, + "loss": 1.0304, + "mean_token_accuracy": 0.6957420110702515, + "num_tokens": 61792109.0, + "step": 2397 + }, + { + "epoch": 0.2633428508675599, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.058518409729004, + "learning_rate": 8.773792093704246e-07, + "loss": 1.0052, + "mean_token_accuracy": 0.6968750953674316, + "num_tokens": 61821867.0, + "step": 2398 + }, + { + "epoch": 0.2634526685701735, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0252387523651123, + "learning_rate": 8.777452415812591e-07, + "loss": 0.9816, + "mean_token_accuracy": 0.7040755748748779, + "num_tokens": 61851076.0, + "step": 2399 + }, + { + "epoch": 0.26356248627278717, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0790598392486572, + "learning_rate": 8.781112737920937e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7005114555358887, + "num_tokens": 61883955.0, + "step": 2400 + }, + { + "epoch": 0.2636723039754008, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1668965816497803, + "learning_rate": 8.784773060029283e-07, + "loss": 0.9247, + "mean_token_accuracy": 0.7160698175430298, + "num_tokens": 61909733.0, + "step": 2401 + }, + { + "epoch": 0.2637821216780145, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1827402114868164, + "learning_rate": 8.788433382137628e-07, + "loss": 1.0285, + "mean_token_accuracy": 0.6886930465698242, + "num_tokens": 61937771.0, + "step": 2402 + }, + { + "epoch": 0.26389193938062816, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6361629962921143, + "learning_rate": 8.792093704245973e-07, + "loss": 1.0006, + "mean_token_accuracy": 0.6991419792175293, + "num_tokens": 61958207.0, + "step": 2403 + }, + { + "epoch": 0.2640017570832418, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1511709690093994, + "learning_rate": 8.795754026354319e-07, + "loss": 1.0515, + "mean_token_accuracy": 0.68086838722229, + "num_tokens": 61987806.0, + "step": 2404 + }, + { + "epoch": 0.2641115747858555, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.106102466583252, + "learning_rate": 8.799414348462665e-07, + "loss": 1.114, + "mean_token_accuracy": 0.6697179675102234, + "num_tokens": 62019461.0, + "step": 2405 + }, + { + "epoch": 0.26422139248846915, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0113942623138428, + "learning_rate": 8.80307467057101e-07, + "loss": 1.11, + "mean_token_accuracy": 0.6733551621437073, + "num_tokens": 62052648.0, + "step": 2406 + }, + { + "epoch": 0.2643312101910828, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3043324947357178, + "learning_rate": 8.806734992679356e-07, + "loss": 1.0202, + "mean_token_accuracy": 0.7019165754318237, + "num_tokens": 62077698.0, + "step": 2407 + }, + { + "epoch": 0.26444102789369645, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2150652408599854, + "learning_rate": 8.810395314787701e-07, + "loss": 1.0478, + "mean_token_accuracy": 0.6960216760635376, + "num_tokens": 62105774.0, + "step": 2408 + }, + { + "epoch": 0.26455084559631015, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4286413192749023, + "learning_rate": 8.814055636896046e-07, + "loss": 1.0268, + "mean_token_accuracy": 0.691206693649292, + "num_tokens": 62128899.0, + "step": 2409 + }, + { + "epoch": 0.2646606632989238, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.362412691116333, + "learning_rate": 8.817715959004392e-07, + "loss": 1.0683, + "mean_token_accuracy": 0.677950918674469, + "num_tokens": 62152471.0, + "step": 2410 + }, + { + "epoch": 0.26477048100153744, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.574641466140747, + "learning_rate": 8.821376281112738e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6913608312606812, + "num_tokens": 62174310.0, + "step": 2411 + }, + { + "epoch": 0.2648802987041511, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6608059406280518, + "learning_rate": 8.825036603221084e-07, + "loss": 1.0216, + "mean_token_accuracy": 0.7070050835609436, + "num_tokens": 62193956.0, + "step": 2412 + }, + { + "epoch": 0.2649901164067648, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1771600246429443, + "learning_rate": 8.828696925329429e-07, + "loss": 1.0529, + "mean_token_accuracy": 0.680404543876648, + "num_tokens": 62220723.0, + "step": 2413 + }, + { + "epoch": 0.26509993410937843, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.258086919784546, + "learning_rate": 8.832357247437774e-07, + "loss": 1.0454, + "mean_token_accuracy": 0.6892086267471313, + "num_tokens": 62246768.0, + "step": 2414 + }, + { + "epoch": 0.2652097518119921, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.125129222869873, + "learning_rate": 8.83601756954612e-07, + "loss": 0.9791, + "mean_token_accuracy": 0.7079653739929199, + "num_tokens": 62274476.0, + "step": 2415 + }, + { + "epoch": 0.2653195695146058, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.170442819595337, + "learning_rate": 8.839677891654466e-07, + "loss": 0.9734, + "mean_token_accuracy": 0.7077993154525757, + "num_tokens": 62300250.0, + "step": 2416 + }, + { + "epoch": 0.2654293872172194, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.162140130996704, + "learning_rate": 8.843338213762811e-07, + "loss": 0.9757, + "mean_token_accuracy": 0.71065753698349, + "num_tokens": 62325482.0, + "step": 2417 + }, + { + "epoch": 0.26553920491983307, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.425046920776367, + "learning_rate": 8.846998535871157e-07, + "loss": 1.0342, + "mean_token_accuracy": 0.6952356696128845, + "num_tokens": 62348385.0, + "step": 2418 + }, + { + "epoch": 0.2656490226224467, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3428757190704346, + "learning_rate": 8.850658857979502e-07, + "loss": 0.9949, + "mean_token_accuracy": 0.7007473707199097, + "num_tokens": 62371942.0, + "step": 2419 + }, + { + "epoch": 0.2657588403250604, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2890536785125732, + "learning_rate": 8.854319180087847e-07, + "loss": 0.9878, + "mean_token_accuracy": 0.7058552503585815, + "num_tokens": 62398163.0, + "step": 2420 + }, + { + "epoch": 0.26586865802767407, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5898563861846924, + "learning_rate": 8.857979502196194e-07, + "loss": 1.0296, + "mean_token_accuracy": 0.6856094598770142, + "num_tokens": 62419047.0, + "step": 2421 + }, + { + "epoch": 0.2659784757302877, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1867735385894775, + "learning_rate": 8.861639824304539e-07, + "loss": 1.114, + "mean_token_accuracy": 0.6816797256469727, + "num_tokens": 62446956.0, + "step": 2422 + }, + { + "epoch": 0.26608829343290136, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5145862102508545, + "learning_rate": 8.865300146412884e-07, + "loss": 1.0475, + "mean_token_accuracy": 0.6902070045471191, + "num_tokens": 62468806.0, + "step": 2423 + }, + { + "epoch": 0.26619811113551506, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.342477321624756, + "learning_rate": 8.86896046852123e-07, + "loss": 1.0557, + "mean_token_accuracy": 0.6867266893386841, + "num_tokens": 62493696.0, + "step": 2424 + }, + { + "epoch": 0.2663079288381287, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8490381240844727, + "learning_rate": 8.872620790629575e-07, + "loss": 0.9255, + "mean_token_accuracy": 0.7140993475914001, + "num_tokens": 62510545.0, + "step": 2425 + }, + { + "epoch": 0.26641774654074235, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.0759236812591553, + "learning_rate": 8.87628111273792e-07, + "loss": 1.0116, + "mean_token_accuracy": 0.7027046084403992, + "num_tokens": 62538602.0, + "step": 2426 + }, + { + "epoch": 0.26652756424335605, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8218181133270264, + "learning_rate": 8.879941434846267e-07, + "loss": 0.9247, + "mean_token_accuracy": 0.7154614329338074, + "num_tokens": 62556858.0, + "step": 2427 + }, + { + "epoch": 0.2666373819459697, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.322655200958252, + "learning_rate": 8.883601756954612e-07, + "loss": 1.0042, + "mean_token_accuracy": 0.7013260126113892, + "num_tokens": 62580368.0, + "step": 2428 + }, + { + "epoch": 0.26674719964858334, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.225116491317749, + "learning_rate": 8.887262079062958e-07, + "loss": 1.015, + "mean_token_accuracy": 0.6961312294006348, + "num_tokens": 62605952.0, + "step": 2429 + }, + { + "epoch": 0.266857017351197, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.097647190093994, + "learning_rate": 8.890922401171303e-07, + "loss": 1.0357, + "mean_token_accuracy": 0.687424898147583, + "num_tokens": 62635903.0, + "step": 2430 + }, + { + "epoch": 0.2669668350538107, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 1.8963756561279297, + "learning_rate": 8.894582723279648e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.6682145595550537, + "num_tokens": 62675543.0, + "step": 2431 + }, + { + "epoch": 0.26707665275642434, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.088190793991089, + "learning_rate": 8.898243045387995e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6753888130187988, + "num_tokens": 62705144.0, + "step": 2432 + }, + { + "epoch": 0.267186470459038, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3753724098205566, + "learning_rate": 8.90190336749634e-07, + "loss": 1.0079, + "mean_token_accuracy": 0.6939904689788818, + "num_tokens": 62730740.0, + "step": 2433 + }, + { + "epoch": 0.2672962881616517, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2709743976593018, + "learning_rate": 8.905563689604685e-07, + "loss": 1.0624, + "mean_token_accuracy": 0.6838912963867188, + "num_tokens": 62755879.0, + "step": 2434 + }, + { + "epoch": 0.26740610586426533, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.038201332092285, + "learning_rate": 8.909224011713031e-07, + "loss": 0.9952, + "mean_token_accuracy": 0.7008781433105469, + "num_tokens": 62786905.0, + "step": 2435 + }, + { + "epoch": 0.267515923566879, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.076305389404297, + "learning_rate": 8.912884333821376e-07, + "loss": 1.0417, + "mean_token_accuracy": 0.6920933127403259, + "num_tokens": 62815251.0, + "step": 2436 + }, + { + "epoch": 0.2676257412694926, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2370240688323975, + "learning_rate": 8.916544655929722e-07, + "loss": 0.9796, + "mean_token_accuracy": 0.7126747369766235, + "num_tokens": 62840071.0, + "step": 2437 + }, + { + "epoch": 0.2677355589721063, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.352142095565796, + "learning_rate": 8.920204978038068e-07, + "loss": 0.9435, + "mean_token_accuracy": 0.7113732099533081, + "num_tokens": 62863599.0, + "step": 2438 + }, + { + "epoch": 0.26784537667471997, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.383957624435425, + "learning_rate": 8.923865300146413e-07, + "loss": 0.9884, + "mean_token_accuracy": 0.7008799314498901, + "num_tokens": 62886063.0, + "step": 2439 + }, + { + "epoch": 0.2679551943773336, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2977354526519775, + "learning_rate": 8.927525622254757e-07, + "loss": 1.0141, + "mean_token_accuracy": 0.7042567729949951, + "num_tokens": 62911701.0, + "step": 2440 + }, + { + "epoch": 0.26806501207994726, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.677778959274292, + "learning_rate": 8.931185944363104e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.7072210907936096, + "num_tokens": 62931389.0, + "step": 2441 + }, + { + "epoch": 0.26817482978256096, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.1755354404449463, + "learning_rate": 8.934846266471448e-07, + "loss": 1.0927, + "mean_token_accuracy": 0.6739816665649414, + "num_tokens": 62959298.0, + "step": 2442 + }, + { + "epoch": 0.2682846474851746, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.1225149631500244, + "learning_rate": 8.938506588579794e-07, + "loss": 0.9932, + "mean_token_accuracy": 0.707341194152832, + "num_tokens": 62987228.0, + "step": 2443 + }, + { + "epoch": 0.26839446518778826, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.04866099357605, + "learning_rate": 8.94216691068814e-07, + "loss": 1.0003, + "mean_token_accuracy": 0.6960958242416382, + "num_tokens": 63016848.0, + "step": 2444 + }, + { + "epoch": 0.26850428289040196, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.2008721828460693, + "learning_rate": 8.945827232796485e-07, + "loss": 1.1033, + "mean_token_accuracy": 0.6745714545249939, + "num_tokens": 63047314.0, + "step": 2445 + }, + { + "epoch": 0.2686141005930156, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 1.998002290725708, + "learning_rate": 8.949487554904832e-07, + "loss": 1.1065, + "mean_token_accuracy": 0.6719143986701965, + "num_tokens": 63081054.0, + "step": 2446 + }, + { + "epoch": 0.26872391829562925, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.5306267738342285, + "learning_rate": 8.953147877013176e-07, + "loss": 1.0083, + "mean_token_accuracy": 0.6970951557159424, + "num_tokens": 63102763.0, + "step": 2447 + }, + { + "epoch": 0.2688337359982429, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.5672152042388916, + "learning_rate": 8.956808199121522e-07, + "loss": 1.0148, + "mean_token_accuracy": 0.6934731006622314, + "num_tokens": 63124068.0, + "step": 2448 + }, + { + "epoch": 0.2689435537008566, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3495612144470215, + "learning_rate": 8.960468521229869e-07, + "loss": 0.9646, + "mean_token_accuracy": 0.7074301242828369, + "num_tokens": 63145050.0, + "step": 2449 + }, + { + "epoch": 0.26905337140347024, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.396287441253662, + "learning_rate": 8.964128843338213e-07, + "loss": 0.9749, + "mean_token_accuracy": 0.7120481729507446, + "num_tokens": 63167810.0, + "step": 2450 + }, + { + "epoch": 0.2691631891060839, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1160168647766113, + "learning_rate": 8.967789165446558e-07, + "loss": 0.9853, + "mean_token_accuracy": 0.7049530744552612, + "num_tokens": 63197274.0, + "step": 2451 + }, + { + "epoch": 0.26927300680869753, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.237673044204712, + "learning_rate": 8.971449487554904e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6869872808456421, + "num_tokens": 63221700.0, + "step": 2452 + }, + { + "epoch": 0.26938282451131124, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.301271677017212, + "learning_rate": 8.97510980966325e-07, + "loss": 1.0611, + "mean_token_accuracy": 0.6806680560112, + "num_tokens": 63248117.0, + "step": 2453 + }, + { + "epoch": 0.2694926422139249, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.3704328536987305, + "learning_rate": 8.978770131771595e-07, + "loss": 0.926, + "mean_token_accuracy": 0.7169276475906372, + "num_tokens": 63271257.0, + "step": 2454 + }, + { + "epoch": 0.26960245991653853, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.4725120067596436, + "learning_rate": 8.982430453879941e-07, + "loss": 1.0219, + "mean_token_accuracy": 0.6892048716545105, + "num_tokens": 63293129.0, + "step": 2455 + }, + { + "epoch": 0.26971227761915223, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.5762064456939697, + "learning_rate": 8.986090775988286e-07, + "loss": 1.0063, + "mean_token_accuracy": 0.6989848017692566, + "num_tokens": 63315767.0, + "step": 2456 + }, + { + "epoch": 0.2698220953217659, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.1298980712890625, + "learning_rate": 8.989751098096631e-07, + "loss": 1.0284, + "mean_token_accuracy": 0.6962185502052307, + "num_tokens": 63345045.0, + "step": 2457 + }, + { + "epoch": 0.2699319130243795, + "ewc_loss": 1.704692840576172e-05, + "grad_norm": 2.2378995418548584, + "learning_rate": 8.993411420204977e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.6976054906845093, + "num_tokens": 63370935.0, + "step": 2458 + }, + { + "epoch": 0.27004173072699317, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.353264570236206, + "learning_rate": 8.997071742313323e-07, + "loss": 1.0208, + "mean_token_accuracy": 0.6883023977279663, + "num_tokens": 63395023.0, + "step": 2459 + }, + { + "epoch": 0.27015154842960687, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 1.97087824344635, + "learning_rate": 9.000732064421668e-07, + "loss": 1.1006, + "mean_token_accuracy": 0.6739176511764526, + "num_tokens": 63429774.0, + "step": 2460 + }, + { + "epoch": 0.2702613661322205, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.207401752471924, + "learning_rate": 9.004392386530014e-07, + "loss": 0.9573, + "mean_token_accuracy": 0.7069278955459595, + "num_tokens": 63454711.0, + "step": 2461 + }, + { + "epoch": 0.27037118383483416, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.2866382598876953, + "learning_rate": 9.008052708638359e-07, + "loss": 0.9869, + "mean_token_accuracy": 0.698864221572876, + "num_tokens": 63481256.0, + "step": 2462 + }, + { + "epoch": 0.27048100153744786, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.423686981201172, + "learning_rate": 9.011713030746705e-07, + "loss": 1.0565, + "mean_token_accuracy": 0.684366762638092, + "num_tokens": 63507253.0, + "step": 2463 + }, + { + "epoch": 0.2705908192400615, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.308889389038086, + "learning_rate": 9.015373352855051e-07, + "loss": 1.0029, + "mean_token_accuracy": 0.6888301372528076, + "num_tokens": 63532003.0, + "step": 2464 + }, + { + "epoch": 0.27070063694267515, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.465718984603882, + "learning_rate": 9.019033674963396e-07, + "loss": 0.9921, + "mean_token_accuracy": 0.698287844657898, + "num_tokens": 63554003.0, + "step": 2465 + }, + { + "epoch": 0.2708104546452888, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.5193419456481934, + "learning_rate": 9.022693997071742e-07, + "loss": 0.8862, + "mean_token_accuracy": 0.7264423966407776, + "num_tokens": 63574058.0, + "step": 2466 + }, + { + "epoch": 0.2709202723479025, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.4770848751068115, + "learning_rate": 9.026354319180087e-07, + "loss": 0.9634, + "mean_token_accuracy": 0.7045668363571167, + "num_tokens": 63596234.0, + "step": 2467 + }, + { + "epoch": 0.27103009005051615, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.0094850063323975, + "learning_rate": 9.030014641288432e-07, + "loss": 1.0574, + "mean_token_accuracy": 0.685294508934021, + "num_tokens": 63627837.0, + "step": 2468 + }, + { + "epoch": 0.2711399077531298, + "ewc_loss": 1.728534698486328e-05, + "grad_norm": 2.0731680393218994, + "learning_rate": 9.033674963396779e-07, + "loss": 1.1241, + "mean_token_accuracy": 0.6677904725074768, + "num_tokens": 63661751.0, + "step": 2469 + }, + { + "epoch": 0.27124972545574344, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.340151309967041, + "learning_rate": 9.037335285505124e-07, + "loss": 1.0442, + "mean_token_accuracy": 0.689642608165741, + "num_tokens": 63686230.0, + "step": 2470 + }, + { + "epoch": 0.27135954315835714, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.4079487323760986, + "learning_rate": 9.040995607613469e-07, + "loss": 1.0944, + "mean_token_accuracy": 0.6721394062042236, + "num_tokens": 63710974.0, + "step": 2471 + }, + { + "epoch": 0.2714693608609708, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.0752079486846924, + "learning_rate": 9.044655929721815e-07, + "loss": 1.0852, + "mean_token_accuracy": 0.6750141382217407, + "num_tokens": 63741057.0, + "step": 2472 + }, + { + "epoch": 0.27157917856358443, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.319857597351074, + "learning_rate": 9.04831625183016e-07, + "loss": 1.0871, + "mean_token_accuracy": 0.6777888536453247, + "num_tokens": 63766625.0, + "step": 2473 + }, + { + "epoch": 0.27168899626619814, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.152252435684204, + "learning_rate": 9.051976573938505e-07, + "loss": 1.0764, + "mean_token_accuracy": 0.6821164488792419, + "num_tokens": 63792679.0, + "step": 2474 + }, + { + "epoch": 0.2717988139688118, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.3464739322662354, + "learning_rate": 9.055636896046852e-07, + "loss": 1.0363, + "mean_token_accuracy": 0.6902163028717041, + "num_tokens": 63818259.0, + "step": 2475 + }, + { + "epoch": 0.2719086316714254, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.4062671661376953, + "learning_rate": 9.059297218155197e-07, + "loss": 0.9721, + "mean_token_accuracy": 0.7018895745277405, + "num_tokens": 63839630.0, + "step": 2476 + }, + { + "epoch": 0.2720184493740391, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.069650173187256, + "learning_rate": 9.062957540263542e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7032792568206787, + "num_tokens": 63868019.0, + "step": 2477 + }, + { + "epoch": 0.2721282670766528, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.0476233959198, + "learning_rate": 9.066617862371888e-07, + "loss": 1.0171, + "mean_token_accuracy": 0.7013158798217773, + "num_tokens": 63896304.0, + "step": 2478 + }, + { + "epoch": 0.2722380847792664, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.0596444606781006, + "learning_rate": 9.070278184480233e-07, + "loss": 0.9499, + "mean_token_accuracy": 0.7084035873413086, + "num_tokens": 63925436.0, + "step": 2479 + }, + { + "epoch": 0.27234790248188007, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.2916412353515625, + "learning_rate": 9.07393850658858e-07, + "loss": 0.9991, + "mean_token_accuracy": 0.6992746591567993, + "num_tokens": 63949487.0, + "step": 2480 + }, + { + "epoch": 0.27245772018449377, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.283710479736328, + "learning_rate": 9.077598828696925e-07, + "loss": 0.9279, + "mean_token_accuracy": 0.7229316830635071, + "num_tokens": 63972779.0, + "step": 2481 + }, + { + "epoch": 0.2725675378871074, + "ewc_loss": 1.7523765563964844e-05, + "grad_norm": 2.9230899810791016, + "learning_rate": 9.08125915080527e-07, + "loss": 0.9926, + "mean_token_accuracy": 0.6985440254211426, + "num_tokens": 63990229.0, + "step": 2482 + }, + { + "epoch": 0.27267735558972106, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.4125008583068848, + "learning_rate": 9.084919472913616e-07, + "loss": 1.0866, + "mean_token_accuracy": 0.6736463308334351, + "num_tokens": 64014964.0, + "step": 2483 + }, + { + "epoch": 0.2727871732923347, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.189427375793457, + "learning_rate": 9.088579795021961e-07, + "loss": 0.9946, + "mean_token_accuracy": 0.6983093023300171, + "num_tokens": 64041383.0, + "step": 2484 + }, + { + "epoch": 0.2728969909949484, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.2228667736053467, + "learning_rate": 9.092240117130307e-07, + "loss": 1.0594, + "mean_token_accuracy": 0.6801972985267639, + "num_tokens": 64070705.0, + "step": 2485 + }, + { + "epoch": 0.27300680869756205, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.258211612701416, + "learning_rate": 9.095900439238653e-07, + "loss": 0.9215, + "mean_token_accuracy": 0.7151790857315063, + "num_tokens": 64098527.0, + "step": 2486 + }, + { + "epoch": 0.2731166264001757, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 1.976714015007019, + "learning_rate": 9.099560761346998e-07, + "loss": 0.9749, + "mean_token_accuracy": 0.7015371918678284, + "num_tokens": 64129915.0, + "step": 2487 + }, + { + "epoch": 0.27322644410278935, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.3969621658325195, + "learning_rate": 9.103221083455343e-07, + "loss": 1.0309, + "mean_token_accuracy": 0.6918870806694031, + "num_tokens": 64154983.0, + "step": 2488 + }, + { + "epoch": 0.27333626180540305, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.0540006160736084, + "learning_rate": 9.106881405563689e-07, + "loss": 1.1412, + "mean_token_accuracy": 0.6728572249412537, + "num_tokens": 64186278.0, + "step": 2489 + }, + { + "epoch": 0.2734460795080167, + "ewc_loss": 1.7642974853515625e-05, + "grad_norm": 2.350020408630371, + "learning_rate": 9.110541727672034e-07, + "loss": 1.041, + "mean_token_accuracy": 0.6889253854751587, + "num_tokens": 64211406.0, + "step": 2490 + }, + { + "epoch": 0.27355589721063034, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.416313648223877, + "learning_rate": 9.11420204978038e-07, + "loss": 0.9526, + "mean_token_accuracy": 0.7108633518218994, + "num_tokens": 64236208.0, + "step": 2491 + }, + { + "epoch": 0.27366571491324404, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.013836622238159, + "learning_rate": 9.117862371888726e-07, + "loss": 1.1201, + "mean_token_accuracy": 0.6685180068016052, + "num_tokens": 64269029.0, + "step": 2492 + }, + { + "epoch": 0.2737755326158577, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.271632432937622, + "learning_rate": 9.121522693997071e-07, + "loss": 1.0837, + "mean_token_accuracy": 0.6704608201980591, + "num_tokens": 64295333.0, + "step": 2493 + }, + { + "epoch": 0.27388535031847133, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.147552490234375, + "learning_rate": 9.125183016105416e-07, + "loss": 1.0429, + "mean_token_accuracy": 0.6856265068054199, + "num_tokens": 64325395.0, + "step": 2494 + }, + { + "epoch": 0.273995168021085, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.205838918685913, + "learning_rate": 9.128843338213762e-07, + "loss": 1.0317, + "mean_token_accuracy": 0.6865711212158203, + "num_tokens": 64353858.0, + "step": 2495 + }, + { + "epoch": 0.2741049857236987, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.371464967727661, + "learning_rate": 9.132503660322108e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6822028160095215, + "num_tokens": 64378320.0, + "step": 2496 + }, + { + "epoch": 0.2742148034263123, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.3728885650634766, + "learning_rate": 9.136163982430454e-07, + "loss": 0.9575, + "mean_token_accuracy": 0.7097494006156921, + "num_tokens": 64402520.0, + "step": 2497 + }, + { + "epoch": 0.27432462112892597, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.283782482147217, + "learning_rate": 9.139824304538799e-07, + "loss": 0.9087, + "mean_token_accuracy": 0.7214457988739014, + "num_tokens": 64424227.0, + "step": 2498 + }, + { + "epoch": 0.2744344388315396, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.305814504623413, + "learning_rate": 9.143484626647144e-07, + "loss": 1.1144, + "mean_token_accuracy": 0.6749046444892883, + "num_tokens": 64449016.0, + "step": 2499 + }, + { + "epoch": 0.2745442565341533, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.306809902191162, + "learning_rate": 9.14714494875549e-07, + "loss": 0.8588, + "mean_token_accuracy": 0.732116162776947, + "num_tokens": 64471600.0, + "step": 2500 + }, + { + "epoch": 0.27465407423676697, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.121244430541992, + "learning_rate": 9.150805270863836e-07, + "loss": 1.0938, + "mean_token_accuracy": 0.6795761585235596, + "num_tokens": 64501456.0, + "step": 2501 + }, + { + "epoch": 0.2747638919393806, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.369868040084839, + "learning_rate": 9.154465592972181e-07, + "loss": 1.009, + "mean_token_accuracy": 0.699991762638092, + "num_tokens": 64525896.0, + "step": 2502 + }, + { + "epoch": 0.2748737096419943, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.119718551635742, + "learning_rate": 9.158125915080527e-07, + "loss": 1.0647, + "mean_token_accuracy": 0.6864942908287048, + "num_tokens": 64556903.0, + "step": 2503 + }, + { + "epoch": 0.27498352734460796, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.067664384841919, + "learning_rate": 9.161786237188872e-07, + "loss": 1.0579, + "mean_token_accuracy": 0.680995762348175, + "num_tokens": 64586040.0, + "step": 2504 + }, + { + "epoch": 0.2750933450472216, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.226526975631714, + "learning_rate": 9.165446559297217e-07, + "loss": 1.0767, + "mean_token_accuracy": 0.6833375692367554, + "num_tokens": 64611675.0, + "step": 2505 + }, + { + "epoch": 0.27520316274983525, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.1428301334381104, + "learning_rate": 9.169106881405563e-07, + "loss": 0.945, + "mean_token_accuracy": 0.716011106967926, + "num_tokens": 64637574.0, + "step": 2506 + }, + { + "epoch": 0.27531298045244895, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.457021951675415, + "learning_rate": 9.172767203513909e-07, + "loss": 0.8868, + "mean_token_accuracy": 0.7262017726898193, + "num_tokens": 64659812.0, + "step": 2507 + }, + { + "epoch": 0.2754227981550626, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.484377384185791, + "learning_rate": 9.176427525622254e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6782441735267639, + "num_tokens": 64683019.0, + "step": 2508 + }, + { + "epoch": 0.27553261585767624, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.3494582176208496, + "learning_rate": 9.1800878477306e-07, + "loss": 1.0306, + "mean_token_accuracy": 0.6965199112892151, + "num_tokens": 64708468.0, + "step": 2509 + }, + { + "epoch": 0.27564243356028995, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 1.9996347427368164, + "learning_rate": 9.183748169838945e-07, + "loss": 1.0502, + "mean_token_accuracy": 0.6808769106864929, + "num_tokens": 64740161.0, + "step": 2510 + }, + { + "epoch": 0.2757522512629036, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.5667800903320312, + "learning_rate": 9.18740849194729e-07, + "loss": 0.9665, + "mean_token_accuracy": 0.7069566249847412, + "num_tokens": 64759273.0, + "step": 2511 + }, + { + "epoch": 0.27586206896551724, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.024855613708496, + "learning_rate": 9.191068814055637e-07, + "loss": 0.993, + "mean_token_accuracy": 0.6998385190963745, + "num_tokens": 64788926.0, + "step": 2512 + }, + { + "epoch": 0.2759718866681309, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.2128946781158447, + "learning_rate": 9.194729136163982e-07, + "loss": 1.05, + "mean_token_accuracy": 0.6885518431663513, + "num_tokens": 64817940.0, + "step": 2513 + }, + { + "epoch": 0.2760817043707446, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 1.9070652723312378, + "learning_rate": 9.198389458272328e-07, + "loss": 1.1329, + "mean_token_accuracy": 0.6799931526184082, + "num_tokens": 64855373.0, + "step": 2514 + }, + { + "epoch": 0.27619152207335823, + "ewc_loss": 1.7762184143066406e-05, + "grad_norm": 2.141409158706665, + "learning_rate": 9.202049780380673e-07, + "loss": 1.0043, + "mean_token_accuracy": 0.6938308477401733, + "num_tokens": 64881957.0, + "step": 2515 + }, + { + "epoch": 0.2763013397759719, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.1616358757019043, + "learning_rate": 9.205710102489018e-07, + "loss": 0.9539, + "mean_token_accuracy": 0.7087489366531372, + "num_tokens": 64908585.0, + "step": 2516 + }, + { + "epoch": 0.2764111574785855, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.5357632637023926, + "learning_rate": 9.209370424597365e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.7066372036933899, + "num_tokens": 64930242.0, + "step": 2517 + }, + { + "epoch": 0.2765209751811992, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 1.9903985261917114, + "learning_rate": 9.21303074670571e-07, + "loss": 0.9804, + "mean_token_accuracy": 0.7009989023208618, + "num_tokens": 64959196.0, + "step": 2518 + }, + { + "epoch": 0.27663079288381287, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.157444477081299, + "learning_rate": 9.216691068814055e-07, + "loss": 1.0122, + "mean_token_accuracy": 0.6927249431610107, + "num_tokens": 64987422.0, + "step": 2519 + }, + { + "epoch": 0.2767406105864265, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.1716670989990234, + "learning_rate": 9.220351390922401e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6843642592430115, + "num_tokens": 65015218.0, + "step": 2520 + }, + { + "epoch": 0.2768504282890402, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.2141218185424805, + "learning_rate": 9.224011713030746e-07, + "loss": 1.0159, + "mean_token_accuracy": 0.6974296569824219, + "num_tokens": 65042379.0, + "step": 2521 + }, + { + "epoch": 0.27696024599165386, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.2108075618743896, + "learning_rate": 9.227672035139091e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6857573986053467, + "num_tokens": 65069290.0, + "step": 2522 + }, + { + "epoch": 0.2770700636942675, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.2157883644104004, + "learning_rate": 9.231332357247438e-07, + "loss": 0.9796, + "mean_token_accuracy": 0.6998990774154663, + "num_tokens": 65094108.0, + "step": 2523 + }, + { + "epoch": 0.27717988139688116, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.0994186401367188, + "learning_rate": 9.234992679355783e-07, + "loss": 0.9721, + "mean_token_accuracy": 0.7074329853057861, + "num_tokens": 65122212.0, + "step": 2524 + }, + { + "epoch": 0.27728969909949486, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.0802018642425537, + "learning_rate": 9.238653001464128e-07, + "loss": 1.0688, + "mean_token_accuracy": 0.6846351623535156, + "num_tokens": 65153331.0, + "step": 2525 + }, + { + "epoch": 0.2773995168021085, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.3578903675079346, + "learning_rate": 9.242313323572474e-07, + "loss": 0.9845, + "mean_token_accuracy": 0.7057970762252808, + "num_tokens": 65178441.0, + "step": 2526 + }, + { + "epoch": 0.27750933450472215, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.0566601753234863, + "learning_rate": 9.245973645680819e-07, + "loss": 0.9491, + "mean_token_accuracy": 0.7079523801803589, + "num_tokens": 65207757.0, + "step": 2527 + }, + { + "epoch": 0.2776191522073358, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.2029762268066406, + "learning_rate": 9.249633967789166e-07, + "loss": 1.064, + "mean_token_accuracy": 0.6873165369033813, + "num_tokens": 65236719.0, + "step": 2528 + }, + { + "epoch": 0.2777289699099495, + "ewc_loss": 1.7881393432617188e-05, + "grad_norm": 2.1979353427886963, + "learning_rate": 9.253294289897511e-07, + "loss": 0.9309, + "mean_token_accuracy": 0.7156223654747009, + "num_tokens": 65260911.0, + "step": 2529 + }, + { + "epoch": 0.27783878761256314, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.6312901973724365, + "learning_rate": 9.256954612005856e-07, + "loss": 0.9506, + "mean_token_accuracy": 0.7091953158378601, + "num_tokens": 65281523.0, + "step": 2530 + }, + { + "epoch": 0.2779486053151768, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.3907880783081055, + "learning_rate": 9.260614934114202e-07, + "loss": 1.0196, + "mean_token_accuracy": 0.6904376745223999, + "num_tokens": 65304936.0, + "step": 2531 + }, + { + "epoch": 0.2780584230177905, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.500900983810425, + "learning_rate": 9.264275256222547e-07, + "loss": 0.8914, + "mean_token_accuracy": 0.7235584259033203, + "num_tokens": 65329771.0, + "step": 2532 + }, + { + "epoch": 0.27816824072040414, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.447760581970215, + "learning_rate": 9.267935578330893e-07, + "loss": 1.1041, + "mean_token_accuracy": 0.6751816272735596, + "num_tokens": 65352739.0, + "step": 2533 + }, + { + "epoch": 0.2782780584230178, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.0985958576202393, + "learning_rate": 9.271595900439239e-07, + "loss": 1.0494, + "mean_token_accuracy": 0.6977158784866333, + "num_tokens": 65381469.0, + "step": 2534 + }, + { + "epoch": 0.27838787612563143, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.487440824508667, + "learning_rate": 9.275256222547584e-07, + "loss": 0.9507, + "mean_token_accuracy": 0.7105255126953125, + "num_tokens": 65402894.0, + "step": 2535 + }, + { + "epoch": 0.27849769382824513, + "ewc_loss": 1.800060272216797e-05, + "grad_norm": 2.1157374382019043, + "learning_rate": 9.278916544655929e-07, + "loss": 1.0897, + "mean_token_accuracy": 0.6839044094085693, + "num_tokens": 65433634.0, + "step": 2536 + }, + { + "epoch": 0.2786075115308588, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.7527575492858887, + "learning_rate": 9.282576866764275e-07, + "loss": 0.89, + "mean_token_accuracy": 0.7235995531082153, + "num_tokens": 65452164.0, + "step": 2537 + }, + { + "epoch": 0.2787173292334724, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.202519178390503, + "learning_rate": 9.28623718887262e-07, + "loss": 1.0742, + "mean_token_accuracy": 0.6873412132263184, + "num_tokens": 65478967.0, + "step": 2538 + }, + { + "epoch": 0.2788271469360861, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.6356213092803955, + "learning_rate": 9.289897510980966e-07, + "loss": 0.9472, + "mean_token_accuracy": 0.7109076976776123, + "num_tokens": 65499590.0, + "step": 2539 + }, + { + "epoch": 0.27893696463869977, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.354731321334839, + "learning_rate": 9.293557833089312e-07, + "loss": 0.9609, + "mean_token_accuracy": 0.7058085203170776, + "num_tokens": 65525619.0, + "step": 2540 + }, + { + "epoch": 0.2790467823413134, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.4519004821777344, + "learning_rate": 9.297218155197657e-07, + "loss": 0.9715, + "mean_token_accuracy": 0.7052979469299316, + "num_tokens": 65547328.0, + "step": 2541 + }, + { + "epoch": 0.27915660004392706, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.241285562515259, + "learning_rate": 9.300878477306002e-07, + "loss": 1.0794, + "mean_token_accuracy": 0.6783289909362793, + "num_tokens": 65574613.0, + "step": 2542 + }, + { + "epoch": 0.27926641774654076, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.123708486557007, + "learning_rate": 9.304538799414348e-07, + "loss": 0.9617, + "mean_token_accuracy": 0.7032659649848938, + "num_tokens": 65603136.0, + "step": 2543 + }, + { + "epoch": 0.2793762354491544, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.2749083042144775, + "learning_rate": 9.308199121522694e-07, + "loss": 1.0005, + "mean_token_accuracy": 0.6970584392547607, + "num_tokens": 65628228.0, + "step": 2544 + }, + { + "epoch": 0.27948605315176805, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.281656265258789, + "learning_rate": 9.31185944363104e-07, + "loss": 1.0306, + "mean_token_accuracy": 0.6884562969207764, + "num_tokens": 65655181.0, + "step": 2545 + }, + { + "epoch": 0.2795958708543817, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.043626308441162, + "learning_rate": 9.315519765739385e-07, + "loss": 1.1167, + "mean_token_accuracy": 0.6733680963516235, + "num_tokens": 65684935.0, + "step": 2546 + }, + { + "epoch": 0.2797056885569954, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.3740882873535156, + "learning_rate": 9.31918008784773e-07, + "loss": 1.0007, + "mean_token_accuracy": 0.6947690844535828, + "num_tokens": 65708943.0, + "step": 2547 + }, + { + "epoch": 0.27981550625960905, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.133467435836792, + "learning_rate": 9.322840409956076e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.7087428569793701, + "num_tokens": 65737123.0, + "step": 2548 + }, + { + "epoch": 0.2799253239622227, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.391235828399658, + "learning_rate": 9.326500732064422e-07, + "loss": 1.0351, + "mean_token_accuracy": 0.696284294128418, + "num_tokens": 65760927.0, + "step": 2549 + }, + { + "epoch": 0.2800351416648364, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.183957099914551, + "learning_rate": 9.330161054172767e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.6900432109832764, + "num_tokens": 65788313.0, + "step": 2550 + }, + { + "epoch": 0.28014495936745004, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.2252252101898193, + "learning_rate": 9.333821376281113e-07, + "loss": 1.0177, + "mean_token_accuracy": 0.6929830312728882, + "num_tokens": 65812857.0, + "step": 2551 + }, + { + "epoch": 0.2802547770700637, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.3005013465881348, + "learning_rate": 9.337481698389458e-07, + "loss": 0.9776, + "mean_token_accuracy": 0.7080503702163696, + "num_tokens": 65836955.0, + "step": 2552 + }, + { + "epoch": 0.28036459477267733, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.1749260425567627, + "learning_rate": 9.341142020497803e-07, + "loss": 1.0074, + "mean_token_accuracy": 0.6975719928741455, + "num_tokens": 65863603.0, + "step": 2553 + }, + { + "epoch": 0.28047441247529104, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.3279905319213867, + "learning_rate": 9.344802342606149e-07, + "loss": 1.0383, + "mean_token_accuracy": 0.6928204298019409, + "num_tokens": 65890120.0, + "step": 2554 + }, + { + "epoch": 0.2805842301779047, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 1.9620623588562012, + "learning_rate": 9.348462664714495e-07, + "loss": 1.0335, + "mean_token_accuracy": 0.6961924433708191, + "num_tokens": 65926236.0, + "step": 2555 + }, + { + "epoch": 0.2806940478805183, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.4638760089874268, + "learning_rate": 9.35212298682284e-07, + "loss": 0.9508, + "mean_token_accuracy": 0.7110896706581116, + "num_tokens": 65947643.0, + "step": 2556 + }, + { + "epoch": 0.28080386558313203, + "ewc_loss": 1.823902130126953e-05, + "grad_norm": 2.466909408569336, + "learning_rate": 9.355783308931186e-07, + "loss": 0.9765, + "mean_token_accuracy": 0.7111606597900391, + "num_tokens": 65970667.0, + "step": 2557 + }, + { + "epoch": 0.2809136832857457, + "ewc_loss": 1.811981201171875e-05, + "grad_norm": 2.265977621078491, + "learning_rate": 9.359443631039531e-07, + "loss": 1.0531, + "mean_token_accuracy": 0.6859272122383118, + "num_tokens": 65996991.0, + "step": 2558 + }, + { + "epoch": 0.2810235009883593, + "ewc_loss": 1.823902130126953e-05, + "grad_norm": 2.073958396911621, + "learning_rate": 9.363103953147876e-07, + "loss": 1.0181, + "mean_token_accuracy": 0.6928166747093201, + "num_tokens": 66027766.0, + "step": 2559 + }, + { + "epoch": 0.28113331869097297, + "ewc_loss": 1.823902130126953e-05, + "grad_norm": 2.6276276111602783, + "learning_rate": 9.366764275256223e-07, + "loss": 0.9205, + "mean_token_accuracy": 0.7146857380867004, + "num_tokens": 66046417.0, + "step": 2560 + }, + { + "epoch": 0.28124313639358667, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.1505510807037354, + "learning_rate": 9.370424597364568e-07, + "loss": 0.9655, + "mean_token_accuracy": 0.7072881460189819, + "num_tokens": 66073856.0, + "step": 2561 + }, + { + "epoch": 0.2813529540962003, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.236764669418335, + "learning_rate": 9.374084919472914e-07, + "loss": 1.0058, + "mean_token_accuracy": 0.6989960670471191, + "num_tokens": 66101429.0, + "step": 2562 + }, + { + "epoch": 0.28146277179881396, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.2094526290893555, + "learning_rate": 9.377745241581259e-07, + "loss": 1.0993, + "mean_token_accuracy": 0.6724527478218079, + "num_tokens": 66129613.0, + "step": 2563 + }, + { + "epoch": 0.2815725895014276, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.1264712810516357, + "learning_rate": 9.381405563689604e-07, + "loss": 1.0238, + "mean_token_accuracy": 0.6904342174530029, + "num_tokens": 66159666.0, + "step": 2564 + }, + { + "epoch": 0.2816824072040413, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.6826179027557373, + "learning_rate": 9.385065885797951e-07, + "loss": 0.9969, + "mean_token_accuracy": 0.7011045217514038, + "num_tokens": 66181171.0, + "step": 2565 + }, + { + "epoch": 0.28179222490665495, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.24678111076355, + "learning_rate": 9.388726207906296e-07, + "loss": 0.9319, + "mean_token_accuracy": 0.7209454774856567, + "num_tokens": 66205013.0, + "step": 2566 + }, + { + "epoch": 0.2819020426092686, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.2567014694213867, + "learning_rate": 9.392386530014641e-07, + "loss": 1.0014, + "mean_token_accuracy": 0.7023178339004517, + "num_tokens": 66230664.0, + "step": 2567 + }, + { + "epoch": 0.2820118603118823, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.0378060340881348, + "learning_rate": 9.396046852122987e-07, + "loss": 1.0005, + "mean_token_accuracy": 0.6995880007743835, + "num_tokens": 66260812.0, + "step": 2568 + }, + { + "epoch": 0.28212167801449595, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 2.436976909637451, + "learning_rate": 9.399707174231332e-07, + "loss": 1.0289, + "mean_token_accuracy": 0.685316264629364, + "num_tokens": 66282638.0, + "step": 2569 + }, + { + "epoch": 0.2822314957171096, + "ewc_loss": 1.8358230590820312e-05, + "grad_norm": 1.946360468864441, + "learning_rate": 9.403367496339677e-07, + "loss": 1.0562, + "mean_token_accuracy": 0.6869319677352905, + "num_tokens": 66315689.0, + "step": 2570 + }, + { + "epoch": 0.28234131341972324, + "ewc_loss": 1.8477439880371094e-05, + "grad_norm": 2.0502984523773193, + "learning_rate": 9.407027818448024e-07, + "loss": 0.9683, + "mean_token_accuracy": 0.7062983512878418, + "num_tokens": 66343772.0, + "step": 2571 + }, + { + "epoch": 0.28245113112233694, + "ewc_loss": 1.8477439880371094e-05, + "grad_norm": 2.250918388366699, + "learning_rate": 9.410688140556369e-07, + "loss": 0.9003, + "mean_token_accuracy": 0.7200033664703369, + "num_tokens": 66366792.0, + "step": 2572 + }, + { + "epoch": 0.2825609488249506, + "ewc_loss": 1.8477439880371094e-05, + "grad_norm": 2.375436305999756, + "learning_rate": 9.414348462664714e-07, + "loss": 0.9683, + "mean_token_accuracy": 0.7042126655578613, + "num_tokens": 66389908.0, + "step": 2573 + }, + { + "epoch": 0.28267076652756423, + "ewc_loss": 1.8477439880371094e-05, + "grad_norm": 1.9907368421554565, + "learning_rate": 9.41800878477306e-07, + "loss": 1.0532, + "mean_token_accuracy": 0.6881853342056274, + "num_tokens": 66423245.0, + "step": 2574 + }, + { + "epoch": 0.2827805842301779, + "ewc_loss": 1.8477439880371094e-05, + "grad_norm": 2.271298885345459, + "learning_rate": 9.421669106881405e-07, + "loss": 1.0821, + "mean_token_accuracy": 0.6758167147636414, + "num_tokens": 66453024.0, + "step": 2575 + }, + { + "epoch": 0.2828904019327916, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.198423385620117, + "learning_rate": 9.425329428989752e-07, + "loss": 1.0167, + "mean_token_accuracy": 0.6960006952285767, + "num_tokens": 66479736.0, + "step": 2576 + }, + { + "epoch": 0.2830002196354052, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.161104679107666, + "learning_rate": 9.428989751098097e-07, + "loss": 1.0594, + "mean_token_accuracy": 0.6787290573120117, + "num_tokens": 66509498.0, + "step": 2577 + }, + { + "epoch": 0.28311003733801887, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.3434057235717773, + "learning_rate": 9.432650073206442e-07, + "loss": 1.0134, + "mean_token_accuracy": 0.6969316601753235, + "num_tokens": 66534943.0, + "step": 2578 + }, + { + "epoch": 0.2832198550406326, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.154332160949707, + "learning_rate": 9.436310395314788e-07, + "loss": 0.9975, + "mean_token_accuracy": 0.6980386972427368, + "num_tokens": 66561088.0, + "step": 2579 + }, + { + "epoch": 0.2833296727432462, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.3017091751098633, + "learning_rate": 9.439970717423133e-07, + "loss": 0.8806, + "mean_token_accuracy": 0.7309657335281372, + "num_tokens": 66584705.0, + "step": 2580 + }, + { + "epoch": 0.28343949044585987, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.0084805488586426, + "learning_rate": 9.443631039531479e-07, + "loss": 1.0615, + "mean_token_accuracy": 0.681548535823822, + "num_tokens": 66617156.0, + "step": 2581 + }, + { + "epoch": 0.2835493081484735, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.3151628971099854, + "learning_rate": 9.447291361639825e-07, + "loss": 1.081, + "mean_token_accuracy": 0.6728923320770264, + "num_tokens": 66644407.0, + "step": 2582 + }, + { + "epoch": 0.2836591258510872, + "ewc_loss": 1.8596649169921875e-05, + "grad_norm": 2.139843702316284, + "learning_rate": 9.45095168374817e-07, + "loss": 1.1102, + "mean_token_accuracy": 0.6769801378250122, + "num_tokens": 66680089.0, + "step": 2583 + }, + { + "epoch": 0.28376894355370086, + "ewc_loss": 1.8715858459472656e-05, + "grad_norm": 1.8751611709594727, + "learning_rate": 9.454612005856515e-07, + "loss": 1.0226, + "mean_token_accuracy": 0.6925543546676636, + "num_tokens": 66716109.0, + "step": 2584 + }, + { + "epoch": 0.2838787612563145, + "ewc_loss": 1.8715858459472656e-05, + "grad_norm": 2.427595615386963, + "learning_rate": 9.458272327964861e-07, + "loss": 1.0469, + "mean_token_accuracy": 0.6917533874511719, + "num_tokens": 66740383.0, + "step": 2585 + }, + { + "epoch": 0.2839885789589282, + "ewc_loss": 1.8715858459472656e-05, + "grad_norm": 1.9530154466629028, + "learning_rate": 9.461932650073206e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6939975023269653, + "num_tokens": 66774475.0, + "step": 2586 + }, + { + "epoch": 0.28409839666154185, + "ewc_loss": 1.8715858459472656e-05, + "grad_norm": 2.0584311485290527, + "learning_rate": 9.465592972181552e-07, + "loss": 0.9344, + "mean_token_accuracy": 0.7177002429962158, + "num_tokens": 66802774.0, + "step": 2587 + }, + { + "epoch": 0.2842082143641555, + "ewc_loss": 1.8715858459472656e-05, + "grad_norm": 2.638458490371704, + "learning_rate": 9.469253294289898e-07, + "loss": 0.9785, + "mean_token_accuracy": 0.6973217129707336, + "num_tokens": 66822965.0, + "step": 2588 + }, + { + "epoch": 0.28431803206676914, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.1387431621551514, + "learning_rate": 9.472913616398243e-07, + "loss": 1.0477, + "mean_token_accuracy": 0.6909966468811035, + "num_tokens": 66852734.0, + "step": 2589 + }, + { + "epoch": 0.28442784976938285, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.246891975402832, + "learning_rate": 9.476573938506587e-07, + "loss": 0.9339, + "mean_token_accuracy": 0.7191079258918762, + "num_tokens": 66877006.0, + "step": 2590 + }, + { + "epoch": 0.2845376674719965, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.343841314315796, + "learning_rate": 9.480234260614933e-07, + "loss": 1.0571, + "mean_token_accuracy": 0.687900960445404, + "num_tokens": 66902300.0, + "step": 2591 + }, + { + "epoch": 0.28464748517461014, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.132124423980713, + "learning_rate": 9.48389458272328e-07, + "loss": 0.9284, + "mean_token_accuracy": 0.7194933891296387, + "num_tokens": 66930991.0, + "step": 2592 + }, + { + "epoch": 0.2847573028772238, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.261802911758423, + "learning_rate": 9.487554904831626e-07, + "loss": 1.0209, + "mean_token_accuracy": 0.6960703134536743, + "num_tokens": 66957004.0, + "step": 2593 + }, + { + "epoch": 0.2848671205798375, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.5231544971466064, + "learning_rate": 9.49121522693997e-07, + "loss": 1.0268, + "mean_token_accuracy": 0.7018491625785828, + "num_tokens": 66980931.0, + "step": 2594 + }, + { + "epoch": 0.28497693828245113, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.359623908996582, + "learning_rate": 9.494875549048315e-07, + "loss": 0.9844, + "mean_token_accuracy": 0.7090133428573608, + "num_tokens": 67005148.0, + "step": 2595 + }, + { + "epoch": 0.2850867559850648, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.611755132675171, + "learning_rate": 9.498535871156661e-07, + "loss": 1.014, + "mean_token_accuracy": 0.6826661825180054, + "num_tokens": 67024780.0, + "step": 2596 + }, + { + "epoch": 0.2851965736876785, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.226757764816284, + "learning_rate": 9.502196193265007e-07, + "loss": 1.0356, + "mean_token_accuracy": 0.693846583366394, + "num_tokens": 67051088.0, + "step": 2597 + }, + { + "epoch": 0.2853063913902921, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.5627410411834717, + "learning_rate": 9.505856515373352e-07, + "loss": 1.004, + "mean_token_accuracy": 0.7004649043083191, + "num_tokens": 67073148.0, + "step": 2598 + }, + { + "epoch": 0.28541620909290577, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.1771938800811768, + "learning_rate": 9.509516837481698e-07, + "loss": 1.0768, + "mean_token_accuracy": 0.685815691947937, + "num_tokens": 67103289.0, + "step": 2599 + }, + { + "epoch": 0.2855260267955194, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.151297092437744, + "learning_rate": 9.513177159590043e-07, + "loss": 1.1299, + "mean_token_accuracy": 0.6645423173904419, + "num_tokens": 67134940.0, + "step": 2600 + }, + { + "epoch": 0.2856358444981331, + "ewc_loss": 1.8835067749023438e-05, + "grad_norm": 2.148181915283203, + "learning_rate": 9.516837481698388e-07, + "loss": 1.036, + "mean_token_accuracy": 0.6912847757339478, + "num_tokens": 67162861.0, + "step": 2601 + }, + { + "epoch": 0.28574566220074676, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.5051257610321045, + "learning_rate": 9.520497803806734e-07, + "loss": 0.9821, + "mean_token_accuracy": 0.7004819512367249, + "num_tokens": 67183291.0, + "step": 2602 + }, + { + "epoch": 0.2858554799033604, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.306668996810913, + "learning_rate": 9.52415812591508e-07, + "loss": 1.0175, + "mean_token_accuracy": 0.6946600675582886, + "num_tokens": 67208977.0, + "step": 2603 + }, + { + "epoch": 0.28596529760597406, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.271650791168213, + "learning_rate": 9.527818448023425e-07, + "loss": 0.9897, + "mean_token_accuracy": 0.7026593089103699, + "num_tokens": 67233401.0, + "step": 2604 + }, + { + "epoch": 0.28607511530858776, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.232133388519287, + "learning_rate": 9.531478770131771e-07, + "loss": 1.0367, + "mean_token_accuracy": 0.6990716457366943, + "num_tokens": 67258898.0, + "step": 2605 + }, + { + "epoch": 0.2861849330112014, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.0260918140411377, + "learning_rate": 9.535139092240116e-07, + "loss": 0.9933, + "mean_token_accuracy": 0.702005922794342, + "num_tokens": 67288247.0, + "step": 2606 + }, + { + "epoch": 0.28629475071381505, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.260732412338257, + "learning_rate": 9.538799414348462e-07, + "loss": 1.0083, + "mean_token_accuracy": 0.696510910987854, + "num_tokens": 67312209.0, + "step": 2607 + }, + { + "epoch": 0.28640456841642875, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.2283833026885986, + "learning_rate": 9.542459736456808e-07, + "loss": 1.0072, + "mean_token_accuracy": 0.6996848583221436, + "num_tokens": 67336959.0, + "step": 2608 + }, + { + "epoch": 0.2865143861190424, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.070423126220703, + "learning_rate": 9.546120058565154e-07, + "loss": 1.0001, + "mean_token_accuracy": 0.6948407292366028, + "num_tokens": 67365192.0, + "step": 2609 + }, + { + "epoch": 0.28662420382165604, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.2505457401275635, + "learning_rate": 9.549780380673498e-07, + "loss": 1.0016, + "mean_token_accuracy": 0.6951435804367065, + "num_tokens": 67392262.0, + "step": 2610 + }, + { + "epoch": 0.2867340215242697, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.1317296028137207, + "learning_rate": 9.553440702781844e-07, + "loss": 1.005, + "mean_token_accuracy": 0.7016627192497253, + "num_tokens": 67419572.0, + "step": 2611 + }, + { + "epoch": 0.2868438392268834, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.4859752655029297, + "learning_rate": 9.55710102489019e-07, + "loss": 0.9472, + "mean_token_accuracy": 0.717566728591919, + "num_tokens": 67440259.0, + "step": 2612 + }, + { + "epoch": 0.28695365692949704, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 2.4285027980804443, + "learning_rate": 9.560761346998536e-07, + "loss": 0.947, + "mean_token_accuracy": 0.7068654894828796, + "num_tokens": 67464103.0, + "step": 2613 + }, + { + "epoch": 0.2870634746321107, + "ewc_loss": 1.895427703857422e-05, + "grad_norm": 1.9332072734832764, + "learning_rate": 9.56442166910688e-07, + "loss": 1.0964, + "mean_token_accuracy": 0.6734377145767212, + "num_tokens": 67498304.0, + "step": 2614 + }, + { + "epoch": 0.2871732923347244, + "ewc_loss": 1.9073486328125e-05, + "grad_norm": 2.5243756771087646, + "learning_rate": 9.568081991215226e-07, + "loss": 1.0779, + "mean_token_accuracy": 0.6778856515884399, + "num_tokens": 67521821.0, + "step": 2615 + }, + { + "epoch": 0.28728311003733803, + "ewc_loss": 1.9073486328125e-05, + "grad_norm": 2.640737771987915, + "learning_rate": 9.571742313323572e-07, + "loss": 0.9662, + "mean_token_accuracy": 0.7082554697990417, + "num_tokens": 67541624.0, + "step": 2616 + }, + { + "epoch": 0.2873929277399517, + "ewc_loss": 1.9073486328125e-05, + "grad_norm": 2.3941545486450195, + "learning_rate": 9.575402635431916e-07, + "loss": 0.9864, + "mean_token_accuracy": 0.7096352577209473, + "num_tokens": 67563863.0, + "step": 2617 + }, + { + "epoch": 0.2875027454425653, + "ewc_loss": 1.9073486328125e-05, + "grad_norm": 2.3544697761535645, + "learning_rate": 9.579062957540262e-07, + "loss": 1.075, + "mean_token_accuracy": 0.6761769652366638, + "num_tokens": 67588335.0, + "step": 2618 + }, + { + "epoch": 0.287612563145179, + "ewc_loss": 1.9073486328125e-05, + "grad_norm": 2.4538698196411133, + "learning_rate": 9.582723279648608e-07, + "loss": 0.9343, + "mean_token_accuracy": 0.7116667032241821, + "num_tokens": 67610513.0, + "step": 2619 + }, + { + "epoch": 0.28772238084779267, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.0384252071380615, + "learning_rate": 9.586383601756954e-07, + "loss": 1.1326, + "mean_token_accuracy": 0.665282130241394, + "num_tokens": 67643857.0, + "step": 2620 + }, + { + "epoch": 0.2878321985504063, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.423076629638672, + "learning_rate": 9.5900439238653e-07, + "loss": 0.9821, + "mean_token_accuracy": 0.6980658769607544, + "num_tokens": 67665589.0, + "step": 2621 + }, + { + "epoch": 0.28794201625301996, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.039531946182251, + "learning_rate": 9.593704245973644e-07, + "loss": 1.0185, + "mean_token_accuracy": 0.6964030265808105, + "num_tokens": 67695418.0, + "step": 2622 + }, + { + "epoch": 0.28805183395563366, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.4336516857147217, + "learning_rate": 9.59736456808199e-07, + "loss": 0.9857, + "mean_token_accuracy": 0.6981784701347351, + "num_tokens": 67716503.0, + "step": 2623 + }, + { + "epoch": 0.2881616516582473, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.1740894317626953, + "learning_rate": 9.601024890190336e-07, + "loss": 0.9946, + "mean_token_accuracy": 0.7035720348358154, + "num_tokens": 67743036.0, + "step": 2624 + }, + { + "epoch": 0.28827146936086095, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 1.9833958148956299, + "learning_rate": 9.604685212298682e-07, + "loss": 1.0735, + "mean_token_accuracy": 0.6796619892120361, + "num_tokens": 67775477.0, + "step": 2625 + }, + { + "epoch": 0.28838128706347466, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.6098580360412598, + "learning_rate": 9.608345534407028e-07, + "loss": 1.0265, + "mean_token_accuracy": 0.692672073841095, + "num_tokens": 67796246.0, + "step": 2626 + }, + { + "epoch": 0.2884911047660883, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.4428863525390625, + "learning_rate": 9.612005856515372e-07, + "loss": 0.984, + "mean_token_accuracy": 0.7086300849914551, + "num_tokens": 67820758.0, + "step": 2627 + }, + { + "epoch": 0.28860092246870195, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.321490526199341, + "learning_rate": 9.615666178623718e-07, + "loss": 1.0088, + "mean_token_accuracy": 0.697411060333252, + "num_tokens": 67844105.0, + "step": 2628 + }, + { + "epoch": 0.2887107401713156, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.0306482315063477, + "learning_rate": 9.619326500732064e-07, + "loss": 1.0834, + "mean_token_accuracy": 0.6726131439208984, + "num_tokens": 67876536.0, + "step": 2629 + }, + { + "epoch": 0.2888205578739293, + "ewc_loss": 1.919269561767578e-05, + "grad_norm": 2.102092981338501, + "learning_rate": 9.62298682284041e-07, + "loss": 0.9772, + "mean_token_accuracy": 0.7091368436813354, + "num_tokens": 67906726.0, + "step": 2630 + }, + { + "epoch": 0.28893037557654294, + "ewc_loss": 1.9311904907226562e-05, + "grad_norm": 2.1940996646881104, + "learning_rate": 9.626647144948754e-07, + "loss": 1.0279, + "mean_token_accuracy": 0.6817902326583862, + "num_tokens": 67934853.0, + "step": 2631 + }, + { + "epoch": 0.2890401932791566, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.227048635482788, + "learning_rate": 9.6303074670571e-07, + "loss": 1.0175, + "mean_token_accuracy": 0.7038172483444214, + "num_tokens": 67960487.0, + "step": 2632 + }, + { + "epoch": 0.2891500109817703, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.2460832595825195, + "learning_rate": 9.633967789165446e-07, + "loss": 0.9979, + "mean_token_accuracy": 0.6959781646728516, + "num_tokens": 67985099.0, + "step": 2633 + }, + { + "epoch": 0.28925982868438394, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.6327426433563232, + "learning_rate": 9.63762811127379e-07, + "loss": 1.1012, + "mean_token_accuracy": 0.6747446656227112, + "num_tokens": 68005907.0, + "step": 2634 + }, + { + "epoch": 0.2893696463869976, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.5375375747680664, + "learning_rate": 9.641288433382138e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.7119896411895752, + "num_tokens": 68025604.0, + "step": 2635 + }, + { + "epoch": 0.2894794640896112, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.1018576622009277, + "learning_rate": 9.644948755490482e-07, + "loss": 0.8938, + "mean_token_accuracy": 0.7254599332809448, + "num_tokens": 68053834.0, + "step": 2636 + }, + { + "epoch": 0.28958928179222493, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.4554569721221924, + "learning_rate": 9.648609077598828e-07, + "loss": 1.0192, + "mean_token_accuracy": 0.692143976688385, + "num_tokens": 68075654.0, + "step": 2637 + }, + { + "epoch": 0.2896990994948386, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.3091237545013428, + "learning_rate": 9.652269399707174e-07, + "loss": 0.9974, + "mean_token_accuracy": 0.7018048763275146, + "num_tokens": 68100679.0, + "step": 2638 + }, + { + "epoch": 0.2898089171974522, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.270434617996216, + "learning_rate": 9.655929721815518e-07, + "loss": 1.0442, + "mean_token_accuracy": 0.6934249997138977, + "num_tokens": 68126152.0, + "step": 2639 + }, + { + "epoch": 0.28991873490006587, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 1.865087866783142, + "learning_rate": 9.659590043923866e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.6943784952163696, + "num_tokens": 68160904.0, + "step": 2640 + }, + { + "epoch": 0.29002855260267957, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.0616626739501953, + "learning_rate": 9.66325036603221e-07, + "loss": 1.1084, + "mean_token_accuracy": 0.6693817377090454, + "num_tokens": 68194352.0, + "step": 2641 + }, + { + "epoch": 0.2901383703052932, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.285015344619751, + "learning_rate": 9.666910688140556e-07, + "loss": 1.0275, + "mean_token_accuracy": 0.6955013275146484, + "num_tokens": 68220095.0, + "step": 2642 + }, + { + "epoch": 0.29024818800790686, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.2814650535583496, + "learning_rate": 9.670571010248902e-07, + "loss": 0.9879, + "mean_token_accuracy": 0.7034315466880798, + "num_tokens": 68245461.0, + "step": 2643 + }, + { + "epoch": 0.29035800571052056, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.2763333320617676, + "learning_rate": 9.674231332357246e-07, + "loss": 0.9967, + "mean_token_accuracy": 0.6944820880889893, + "num_tokens": 68270879.0, + "step": 2644 + }, + { + "epoch": 0.2904678234131342, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.433607816696167, + "learning_rate": 9.677891654465592e-07, + "loss": 0.9292, + "mean_token_accuracy": 0.7115107774734497, + "num_tokens": 68292893.0, + "step": 2645 + }, + { + "epoch": 0.29057764111574785, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.1639621257781982, + "learning_rate": 9.681551976573938e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6838124990463257, + "num_tokens": 68319075.0, + "step": 2646 + }, + { + "epoch": 0.2906874588183615, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.5080881118774414, + "learning_rate": 9.685212298682284e-07, + "loss": 0.9464, + "mean_token_accuracy": 0.711647629737854, + "num_tokens": 68342098.0, + "step": 2647 + }, + { + "epoch": 0.2907972765209752, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.1916873455047607, + "learning_rate": 9.688872620790628e-07, + "loss": 1.0887, + "mean_token_accuracy": 0.6710832715034485, + "num_tokens": 68371225.0, + "step": 2648 + }, + { + "epoch": 0.29090709422358885, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.3046669960021973, + "learning_rate": 9.692532942898974e-07, + "loss": 1.0465, + "mean_token_accuracy": 0.6862226128578186, + "num_tokens": 68398188.0, + "step": 2649 + }, + { + "epoch": 0.2910169119262025, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 2.223391056060791, + "learning_rate": 9.69619326500732e-07, + "loss": 0.9629, + "mean_token_accuracy": 0.7089900970458984, + "num_tokens": 68425520.0, + "step": 2650 + }, + { + "epoch": 0.29112672962881614, + "ewc_loss": 1.9431114196777344e-05, + "grad_norm": 1.8257559537887573, + "learning_rate": 9.699853587115666e-07, + "loss": 1.076, + "mean_token_accuracy": 0.6780347228050232, + "num_tokens": 68462456.0, + "step": 2651 + }, + { + "epoch": 0.29123654733142984, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.249399423599243, + "learning_rate": 9.703513909224012e-07, + "loss": 1.0866, + "mean_token_accuracy": 0.6782732605934143, + "num_tokens": 68489331.0, + "step": 2652 + }, + { + "epoch": 0.2913463650340435, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.2290420532226562, + "learning_rate": 9.707174231332356e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6881572008132935, + "num_tokens": 68516051.0, + "step": 2653 + }, + { + "epoch": 0.29145618273665713, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.341676712036133, + "learning_rate": 9.710834553440702e-07, + "loss": 1.0362, + "mean_token_accuracy": 0.6950737237930298, + "num_tokens": 68539805.0, + "step": 2654 + }, + { + "epoch": 0.29156600043927083, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.293536424636841, + "learning_rate": 9.714494875549048e-07, + "loss": 1.0005, + "mean_token_accuracy": 0.6991139054298401, + "num_tokens": 68565221.0, + "step": 2655 + }, + { + "epoch": 0.2916758181418845, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.177633762359619, + "learning_rate": 9.718155197657394e-07, + "loss": 0.9898, + "mean_token_accuracy": 0.7030451893806458, + "num_tokens": 68593703.0, + "step": 2656 + }, + { + "epoch": 0.2917856358444981, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.351841688156128, + "learning_rate": 9.72181551976574e-07, + "loss": 0.9871, + "mean_token_accuracy": 0.7004258632659912, + "num_tokens": 68615507.0, + "step": 2657 + }, + { + "epoch": 0.29189545354711177, + "ewc_loss": 1.9550323486328125e-05, + "grad_norm": 2.126976251602173, + "learning_rate": 9.725475841874084e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.697677731513977, + "num_tokens": 68644423.0, + "step": 2658 + }, + { + "epoch": 0.2920052712497255, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.182356357574463, + "learning_rate": 9.72913616398243e-07, + "loss": 1.024, + "mean_token_accuracy": 0.6955576539039612, + "num_tokens": 68670329.0, + "step": 2659 + }, + { + "epoch": 0.2921150889523391, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.5879313945770264, + "learning_rate": 9.732796486090776e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.699839174747467, + "num_tokens": 68691542.0, + "step": 2660 + }, + { + "epoch": 0.29222490665495277, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.4663641452789307, + "learning_rate": 9.736456808199122e-07, + "loss": 0.9845, + "mean_token_accuracy": 0.701757550239563, + "num_tokens": 68713333.0, + "step": 2661 + }, + { + "epoch": 0.29233472435756647, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.3603453636169434, + "learning_rate": 9.740117130307466e-07, + "loss": 0.9873, + "mean_token_accuracy": 0.7069805860519409, + "num_tokens": 68735818.0, + "step": 2662 + }, + { + "epoch": 0.2924445420601801, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.1955878734588623, + "learning_rate": 9.743777452415812e-07, + "loss": 1.0165, + "mean_token_accuracy": 0.6950299143791199, + "num_tokens": 68762899.0, + "step": 2663 + }, + { + "epoch": 0.29255435976279376, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.3607919216156006, + "learning_rate": 9.747437774524158e-07, + "loss": 1.0315, + "mean_token_accuracy": 0.6980929374694824, + "num_tokens": 68786750.0, + "step": 2664 + }, + { + "epoch": 0.2926641774654074, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.392683267593384, + "learning_rate": 9.751098096632502e-07, + "loss": 0.9812, + "mean_token_accuracy": 0.7035221457481384, + "num_tokens": 68810131.0, + "step": 2665 + }, + { + "epoch": 0.2927739951680211, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.3043813705444336, + "learning_rate": 9.754758418740848e-07, + "loss": 1.0918, + "mean_token_accuracy": 0.67745041847229, + "num_tokens": 68837302.0, + "step": 2666 + }, + { + "epoch": 0.29288381287063475, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.108328342437744, + "learning_rate": 9.758418740849194e-07, + "loss": 1.0138, + "mean_token_accuracy": 0.7021608948707581, + "num_tokens": 68866848.0, + "step": 2667 + }, + { + "epoch": 0.2929936305732484, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.3781919479370117, + "learning_rate": 9.76207906295754e-07, + "loss": 1.0702, + "mean_token_accuracy": 0.6841300129890442, + "num_tokens": 68892221.0, + "step": 2668 + }, + { + "epoch": 0.29310344827586204, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.095762014389038, + "learning_rate": 9.765739385065886e-07, + "loss": 0.9788, + "mean_token_accuracy": 0.7028450965881348, + "num_tokens": 68920022.0, + "step": 2669 + }, + { + "epoch": 0.29321326597847575, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.507089853286743, + "learning_rate": 9.76939970717423e-07, + "loss": 0.9794, + "mean_token_accuracy": 0.7090009450912476, + "num_tokens": 68941852.0, + "step": 2670 + }, + { + "epoch": 0.2933230836810894, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.1001150608062744, + "learning_rate": 9.773060029282576e-07, + "loss": 1.0614, + "mean_token_accuracy": 0.6807420253753662, + "num_tokens": 68973292.0, + "step": 2671 + }, + { + "epoch": 0.29343290138370304, + "ewc_loss": 1.9669532775878906e-05, + "grad_norm": 2.539144515991211, + "learning_rate": 9.776720351390922e-07, + "loss": 1.0499, + "mean_token_accuracy": 0.6904965043067932, + "num_tokens": 68995369.0, + "step": 2672 + }, + { + "epoch": 0.29354271908631674, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.2733278274536133, + "learning_rate": 9.780380673499268e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.6854879260063171, + "num_tokens": 69018419.0, + "step": 2673 + }, + { + "epoch": 0.2936525367889304, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.6591832637786865, + "learning_rate": 9.784040995607614e-07, + "loss": 0.972, + "mean_token_accuracy": 0.7097082138061523, + "num_tokens": 69039570.0, + "step": 2674 + }, + { + "epoch": 0.29376235449154403, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.294966459274292, + "learning_rate": 9.787701317715958e-07, + "loss": 0.9697, + "mean_token_accuracy": 0.7091662287712097, + "num_tokens": 69065295.0, + "step": 2675 + }, + { + "epoch": 0.2938721721941577, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.398026943206787, + "learning_rate": 9.791361639824304e-07, + "loss": 0.9898, + "mean_token_accuracy": 0.7079254388809204, + "num_tokens": 69088292.0, + "step": 2676 + }, + { + "epoch": 0.2939819898967714, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.5131306648254395, + "learning_rate": 9.79502196193265e-07, + "loss": 0.9761, + "mean_token_accuracy": 0.6981006860733032, + "num_tokens": 69109009.0, + "step": 2677 + }, + { + "epoch": 0.294091807599385, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.1419661045074463, + "learning_rate": 9.798682284040996e-07, + "loss": 0.9697, + "mean_token_accuracy": 0.7135502099990845, + "num_tokens": 69135296.0, + "step": 2678 + }, + { + "epoch": 0.29420162530199867, + "ewc_loss": 1.9788742065429688e-05, + "grad_norm": 2.056159734725952, + "learning_rate": 9.80234260614934e-07, + "loss": 1.0648, + "mean_token_accuracy": 0.6781383752822876, + "num_tokens": 69167474.0, + "step": 2679 + }, + { + "epoch": 0.2943114430046123, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.0189144611358643, + "learning_rate": 9.806002928257686e-07, + "loss": 1.0456, + "mean_token_accuracy": 0.6915770173072815, + "num_tokens": 69199833.0, + "step": 2680 + }, + { + "epoch": 0.294421260707226, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.1529345512390137, + "learning_rate": 9.809663250366032e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6861186623573303, + "num_tokens": 69228030.0, + "step": 2681 + }, + { + "epoch": 0.29453107840983966, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.233980178833008, + "learning_rate": 9.813323572474376e-07, + "loss": 1.0403, + "mean_token_accuracy": 0.6806430816650391, + "num_tokens": 69253176.0, + "step": 2682 + }, + { + "epoch": 0.2946408961124533, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.1940982341766357, + "learning_rate": 9.816983894582724e-07, + "loss": 0.962, + "mean_token_accuracy": 0.7093433141708374, + "num_tokens": 69279049.0, + "step": 2683 + }, + { + "epoch": 0.294750713815067, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.0004448890686035, + "learning_rate": 9.820644216691068e-07, + "loss": 1.05, + "mean_token_accuracy": 0.6873661875724792, + "num_tokens": 69311963.0, + "step": 2684 + }, + { + "epoch": 0.29486053151768066, + "ewc_loss": 1.990795135498047e-05, + "grad_norm": 2.211252212524414, + "learning_rate": 9.824304538799414e-07, + "loss": 0.9506, + "mean_token_accuracy": 0.7130745649337769, + "num_tokens": 69338490.0, + "step": 2685 + }, + { + "epoch": 0.2949703492202943, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.4761736392974854, + "learning_rate": 9.82796486090776e-07, + "loss": 0.9806, + "mean_token_accuracy": 0.6979385614395142, + "num_tokens": 69360211.0, + "step": 2686 + }, + { + "epoch": 0.29508016692290795, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.3128957748413086, + "learning_rate": 9.831625183016104e-07, + "loss": 1.1463, + "mean_token_accuracy": 0.6730855703353882, + "num_tokens": 69389732.0, + "step": 2687 + }, + { + "epoch": 0.29518998462552165, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.4375860691070557, + "learning_rate": 9.835285505124452e-07, + "loss": 0.9158, + "mean_token_accuracy": 0.7166172862052917, + "num_tokens": 69412227.0, + "step": 2688 + }, + { + "epoch": 0.2952998023281353, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.332145929336548, + "learning_rate": 9.838945827232796e-07, + "loss": 1.0349, + "mean_token_accuracy": 0.6912117600440979, + "num_tokens": 69435342.0, + "step": 2689 + }, + { + "epoch": 0.29540962003074894, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 1.9740477800369263, + "learning_rate": 9.842606149341142e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6850454807281494, + "num_tokens": 69466875.0, + "step": 2690 + }, + { + "epoch": 0.29551943773336264, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.2372865676879883, + "learning_rate": 9.846266471449488e-07, + "loss": 1.0365, + "mean_token_accuracy": 0.6997523307800293, + "num_tokens": 69493073.0, + "step": 2691 + }, + { + "epoch": 0.2956292554359763, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.142469882965088, + "learning_rate": 9.849926793557832e-07, + "loss": 0.9656, + "mean_token_accuracy": 0.7051995992660522, + "num_tokens": 69520144.0, + "step": 2692 + }, + { + "epoch": 0.29573907313858994, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.238359212875366, + "learning_rate": 9.853587115666178e-07, + "loss": 0.9927, + "mean_token_accuracy": 0.6986492872238159, + "num_tokens": 69545286.0, + "step": 2693 + }, + { + "epoch": 0.2958488908412036, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.131575345993042, + "learning_rate": 9.857247437774524e-07, + "loss": 1.1255, + "mean_token_accuracy": 0.6728522777557373, + "num_tokens": 69575666.0, + "step": 2694 + }, + { + "epoch": 0.2959587085438173, + "ewc_loss": 2.002716064453125e-05, + "grad_norm": 2.3691153526306152, + "learning_rate": 9.86090775988287e-07, + "loss": 0.9876, + "mean_token_accuracy": 0.6988418102264404, + "num_tokens": 69597333.0, + "step": 2695 + }, + { + "epoch": 0.29606852624643093, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.7761175632476807, + "learning_rate": 9.864568081991214e-07, + "loss": 0.9586, + "mean_token_accuracy": 0.7062435746192932, + "num_tokens": 69615995.0, + "step": 2696 + }, + { + "epoch": 0.2961783439490446, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.332517147064209, + "learning_rate": 9.86822840409956e-07, + "loss": 1.1006, + "mean_token_accuracy": 0.6764038801193237, + "num_tokens": 69640263.0, + "step": 2697 + }, + { + "epoch": 0.2962881616516582, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.4303157329559326, + "learning_rate": 9.871888726207906e-07, + "loss": 1.0075, + "mean_token_accuracy": 0.7059226036071777, + "num_tokens": 69662277.0, + "step": 2698 + }, + { + "epoch": 0.2963979793542719, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 1.9776455163955688, + "learning_rate": 9.875549048316252e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6788171529769897, + "num_tokens": 69694745.0, + "step": 2699 + }, + { + "epoch": 0.29650779705688557, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.2772390842437744, + "learning_rate": 9.879209370424598e-07, + "loss": 1.015, + "mean_token_accuracy": 0.6950960159301758, + "num_tokens": 69720316.0, + "step": 2700 + }, + { + "epoch": 0.2966176147594992, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.1333980560302734, + "learning_rate": 9.882869692532942e-07, + "loss": 1.0358, + "mean_token_accuracy": 0.6885896921157837, + "num_tokens": 69750100.0, + "step": 2701 + }, + { + "epoch": 0.2967274324621129, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.198852300643921, + "learning_rate": 9.886530014641288e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6899479627609253, + "num_tokens": 69777792.0, + "step": 2702 + }, + { + "epoch": 0.29683725016472656, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.236795663833618, + "learning_rate": 9.890190336749634e-07, + "loss": 1.0682, + "mean_token_accuracy": 0.675616979598999, + "num_tokens": 69804220.0, + "step": 2703 + }, + { + "epoch": 0.2969470678673402, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.106940269470215, + "learning_rate": 9.89385065885798e-07, + "loss": 1.0683, + "mean_token_accuracy": 0.6754322648048401, + "num_tokens": 69837893.0, + "step": 2704 + }, + { + "epoch": 0.29705688556995385, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.275562286376953, + "learning_rate": 9.897510980966326e-07, + "loss": 0.9829, + "mean_token_accuracy": 0.7106798887252808, + "num_tokens": 69862269.0, + "step": 2705 + }, + { + "epoch": 0.29716670327256756, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.068889617919922, + "learning_rate": 9.90117130307467e-07, + "loss": 1.0822, + "mean_token_accuracy": 0.6742672920227051, + "num_tokens": 69893396.0, + "step": 2706 + }, + { + "epoch": 0.2972765209751812, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.197601318359375, + "learning_rate": 9.904831625183016e-07, + "loss": 0.9896, + "mean_token_accuracy": 0.697611391544342, + "num_tokens": 69920648.0, + "step": 2707 + }, + { + "epoch": 0.29738633867779485, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.158801555633545, + "learning_rate": 9.908491947291362e-07, + "loss": 0.9978, + "mean_token_accuracy": 0.698631763458252, + "num_tokens": 69947077.0, + "step": 2708 + }, + { + "epoch": 0.29749615638040855, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.1872994899749756, + "learning_rate": 9.912152269399708e-07, + "loss": 1.0768, + "mean_token_accuracy": 0.6834585070610046, + "num_tokens": 69977260.0, + "step": 2709 + }, + { + "epoch": 0.2976059740830222, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.04644775390625, + "learning_rate": 9.915812591508052e-07, + "loss": 0.998, + "mean_token_accuracy": 0.7030585408210754, + "num_tokens": 70006314.0, + "step": 2710 + }, + { + "epoch": 0.29771579178563584, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.168936252593994, + "learning_rate": 9.919472913616398e-07, + "loss": 0.9921, + "mean_token_accuracy": 0.699994683265686, + "num_tokens": 70033098.0, + "step": 2711 + }, + { + "epoch": 0.2978256094882495, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.1756536960601807, + "learning_rate": 9.923133235724744e-07, + "loss": 1.0088, + "mean_token_accuracy": 0.7043259143829346, + "num_tokens": 70062200.0, + "step": 2712 + }, + { + "epoch": 0.2979354271908632, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.164767026901245, + "learning_rate": 9.926793557833088e-07, + "loss": 1.0679, + "mean_token_accuracy": 0.6956838369369507, + "num_tokens": 70089711.0, + "step": 2713 + }, + { + "epoch": 0.29804524489347684, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.226576566696167, + "learning_rate": 9.930453879941434e-07, + "loss": 0.9242, + "mean_token_accuracy": 0.7205860614776611, + "num_tokens": 70113978.0, + "step": 2714 + }, + { + "epoch": 0.2981550625960905, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.2500133514404297, + "learning_rate": 9.93411420204978e-07, + "loss": 0.9851, + "mean_token_accuracy": 0.7008099555969238, + "num_tokens": 70138238.0, + "step": 2715 + }, + { + "epoch": 0.2982648802987041, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.3969364166259766, + "learning_rate": 9.937774524158126e-07, + "loss": 0.9644, + "mean_token_accuracy": 0.7100931406021118, + "num_tokens": 70161906.0, + "step": 2716 + }, + { + "epoch": 0.29837469800131783, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.0725443363189697, + "learning_rate": 9.941434846266472e-07, + "loss": 0.9612, + "mean_token_accuracy": 0.7069438695907593, + "num_tokens": 70192723.0, + "step": 2717 + }, + { + "epoch": 0.2984845157039315, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.1695945262908936, + "learning_rate": 9.945095168374816e-07, + "loss": 0.9901, + "mean_token_accuracy": 0.6999576091766357, + "num_tokens": 70221730.0, + "step": 2718 + }, + { + "epoch": 0.2985943334065451, + "ewc_loss": 2.0265579223632812e-05, + "grad_norm": 2.1894822120666504, + "learning_rate": 9.948755490483162e-07, + "loss": 0.9706, + "mean_token_accuracy": 0.7081934213638306, + "num_tokens": 70249072.0, + "step": 2719 + }, + { + "epoch": 0.2987041511091588, + "ewc_loss": 2.0503997802734375e-05, + "grad_norm": 2.5899887084960938, + "learning_rate": 9.952415812591508e-07, + "loss": 0.9686, + "mean_token_accuracy": 0.7035871744155884, + "num_tokens": 70270451.0, + "step": 2720 + }, + { + "epoch": 0.29881396881177247, + "ewc_loss": 2.0503997802734375e-05, + "grad_norm": 2.544355869293213, + "learning_rate": 9.956076134699854e-07, + "loss": 0.9152, + "mean_token_accuracy": 0.7239418029785156, + "num_tokens": 70290888.0, + "step": 2721 + }, + { + "epoch": 0.2989237865143861, + "ewc_loss": 2.0503997802734375e-05, + "grad_norm": 2.444922924041748, + "learning_rate": 9.9597364568082e-07, + "loss": 1.0875, + "mean_token_accuracy": 0.6754580736160278, + "num_tokens": 70314777.0, + "step": 2722 + }, + { + "epoch": 0.29903360421699976, + "ewc_loss": 2.0503997802734375e-05, + "grad_norm": 2.0845816135406494, + "learning_rate": 9.963396778916544e-07, + "loss": 0.9784, + "mean_token_accuracy": 0.7056581377983093, + "num_tokens": 70343692.0, + "step": 2723 + }, + { + "epoch": 0.29914342191961346, + "ewc_loss": 2.0503997802734375e-05, + "grad_norm": 2.216904878616333, + "learning_rate": 9.96705710102489e-07, + "loss": 0.9947, + "mean_token_accuracy": 0.6961025595664978, + "num_tokens": 70367561.0, + "step": 2724 + }, + { + "epoch": 0.2992532396222271, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.4136974811553955, + "learning_rate": 9.970717423133236e-07, + "loss": 0.9277, + "mean_token_accuracy": 0.7107765674591064, + "num_tokens": 70389900.0, + "step": 2725 + }, + { + "epoch": 0.29936305732484075, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.3952276706695557, + "learning_rate": 9.974377745241582e-07, + "loss": 0.9679, + "mean_token_accuracy": 0.7080049514770508, + "num_tokens": 70415989.0, + "step": 2726 + }, + { + "epoch": 0.2994728750274544, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 1.9679832458496094, + "learning_rate": 9.978038067349926e-07, + "loss": 1.0596, + "mean_token_accuracy": 0.6827673316001892, + "num_tokens": 70448872.0, + "step": 2727 + }, + { + "epoch": 0.2995826927300681, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.0748510360717773, + "learning_rate": 9.981698389458272e-07, + "loss": 1.0287, + "mean_token_accuracy": 0.6927861571311951, + "num_tokens": 70479507.0, + "step": 2728 + }, + { + "epoch": 0.29969251043268175, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.142306089401245, + "learning_rate": 9.985358711566618e-07, + "loss": 0.9747, + "mean_token_accuracy": 0.7007603645324707, + "num_tokens": 70510493.0, + "step": 2729 + }, + { + "epoch": 0.2998023281352954, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.5496978759765625, + "learning_rate": 9.989019033674961e-07, + "loss": 0.9351, + "mean_token_accuracy": 0.7108913064002991, + "num_tokens": 70532115.0, + "step": 2730 + }, + { + "epoch": 0.2999121458379091, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 1.938139796257019, + "learning_rate": 9.99267935578331e-07, + "loss": 0.9085, + "mean_token_accuracy": 0.7110803127288818, + "num_tokens": 70563634.0, + "step": 2731 + }, + { + "epoch": 0.30002196354052274, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.1920740604400635, + "learning_rate": 9.996339677891654e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.7064803838729858, + "num_tokens": 70591530.0, + "step": 2732 + }, + { + "epoch": 0.3001317812431364, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.6155338287353516, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6899624466896057, + "num_tokens": 70611898.0, + "step": 2733 + }, + { + "epoch": 0.30024159894575003, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.5739126205444336, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7191765308380127, + "num_tokens": 70631779.0, + "step": 2734 + }, + { + "epoch": 0.30035141664836373, + "ewc_loss": 2.0623207092285156e-05, + "grad_norm": 2.04547119140625, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6965054273605347, + "num_tokens": 70660246.0, + "step": 2735 + }, + { + "epoch": 0.3004612343509774, + "ewc_loss": 2.086162567138672e-05, + "grad_norm": 2.0516397953033447, + "learning_rate": 1e-06, + "loss": 1.0342, + "mean_token_accuracy": 0.6915371417999268, + "num_tokens": 70691353.0, + "step": 2736 + }, + { + "epoch": 0.300571052053591, + "ewc_loss": 2.086162567138672e-05, + "grad_norm": 2.5638580322265625, + "learning_rate": 1e-06, + "loss": 1.0901, + "mean_token_accuracy": 0.6728840470314026, + "num_tokens": 70714143.0, + "step": 2737 + }, + { + "epoch": 0.3006808697562047, + "ewc_loss": 2.086162567138672e-05, + "grad_norm": 2.2743258476257324, + "learning_rate": 1e-06, + "loss": 1.039, + "mean_token_accuracy": 0.6883540153503418, + "num_tokens": 70738976.0, + "step": 2738 + }, + { + "epoch": 0.3007906874588184, + "ewc_loss": 2.086162567138672e-05, + "grad_norm": 2.4358150959014893, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7150764465332031, + "num_tokens": 70760448.0, + "step": 2739 + }, + { + "epoch": 0.300900505161432, + "ewc_loss": 2.086162567138672e-05, + "grad_norm": 2.1432080268859863, + "learning_rate": 1e-06, + "loss": 1.0921, + "mean_token_accuracy": 0.6700248718261719, + "num_tokens": 70786483.0, + "step": 2740 + }, + { + "epoch": 0.30101032286404567, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.263608455657959, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7102153301239014, + "num_tokens": 70809208.0, + "step": 2741 + }, + { + "epoch": 0.30112014056665937, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.3455348014831543, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.708892822265625, + "num_tokens": 70834608.0, + "step": 2742 + }, + { + "epoch": 0.301229958269273, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.2489118576049805, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7125208377838135, + "num_tokens": 70861261.0, + "step": 2743 + }, + { + "epoch": 0.30133977597188666, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.237534284591675, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6875295042991638, + "num_tokens": 70889551.0, + "step": 2744 + }, + { + "epoch": 0.3014495936745003, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.2213432788848877, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6955165863037109, + "num_tokens": 70916162.0, + "step": 2745 + }, + { + "epoch": 0.301559411377114, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.1483829021453857, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7087819576263428, + "num_tokens": 70944163.0, + "step": 2746 + }, + { + "epoch": 0.30166922907972765, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.1506271362304688, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7062479853630066, + "num_tokens": 70970646.0, + "step": 2747 + }, + { + "epoch": 0.3017790467823413, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 1.948801875114441, + "learning_rate": 1e-06, + "loss": 1.0754, + "mean_token_accuracy": 0.6793326735496521, + "num_tokens": 71003040.0, + "step": 2748 + }, + { + "epoch": 0.301888864484955, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 1.9841593503952026, + "learning_rate": 1e-06, + "loss": 1.068, + "mean_token_accuracy": 0.6801609992980957, + "num_tokens": 71034419.0, + "step": 2749 + }, + { + "epoch": 0.30199868218756865, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.3021209239959717, + "learning_rate": 1e-06, + "loss": 1.0295, + "mean_token_accuracy": 0.6940499544143677, + "num_tokens": 71060266.0, + "step": 2750 + }, + { + "epoch": 0.3021084998901823, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.2842867374420166, + "learning_rate": 1e-06, + "loss": 1.1114, + "mean_token_accuracy": 0.6658810377120972, + "num_tokens": 71085939.0, + "step": 2751 + }, + { + "epoch": 0.30221831759279594, + "ewc_loss": 2.09808349609375e-05, + "grad_norm": 2.235799551010132, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6929861903190613, + "num_tokens": 71110518.0, + "step": 2752 + }, + { + "epoch": 0.30232813529540964, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.4039955139160156, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7075978517532349, + "num_tokens": 71132356.0, + "step": 2753 + }, + { + "epoch": 0.3024379529980233, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.3408401012420654, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6940345764160156, + "num_tokens": 71155990.0, + "step": 2754 + }, + { + "epoch": 0.30254777070063693, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.524664878845215, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7166804671287537, + "num_tokens": 71176187.0, + "step": 2755 + }, + { + "epoch": 0.3026575884032506, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.2457175254821777, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6994879841804504, + "num_tokens": 71200764.0, + "step": 2756 + }, + { + "epoch": 0.3027674061058643, + "ewc_loss": 2.110004425048828e-05, + "grad_norm": 2.3065829277038574, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.6946570873260498, + "num_tokens": 71228226.0, + "step": 2757 + }, + { + "epoch": 0.3028772238084779, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.248607873916626, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6916608810424805, + "num_tokens": 71252338.0, + "step": 2758 + }, + { + "epoch": 0.30298704151109157, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.4478957653045654, + "learning_rate": 1e-06, + "loss": 1.0487, + "mean_token_accuracy": 0.6823898553848267, + "num_tokens": 71276029.0, + "step": 2759 + }, + { + "epoch": 0.30309685921370527, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.3317208290100098, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7169308662414551, + "num_tokens": 71300798.0, + "step": 2760 + }, + { + "epoch": 0.3032066769163189, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 1.972666621208191, + "learning_rate": 1e-06, + "loss": 1.094, + "mean_token_accuracy": 0.6727476716041565, + "num_tokens": 71335051.0, + "step": 2761 + }, + { + "epoch": 0.30331649461893256, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.2813849449157715, + "learning_rate": 1e-06, + "loss": 1.0425, + "mean_token_accuracy": 0.6899818181991577, + "num_tokens": 71361661.0, + "step": 2762 + }, + { + "epoch": 0.3034263123215462, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.451730728149414, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7008770704269409, + "num_tokens": 71382572.0, + "step": 2763 + }, + { + "epoch": 0.3035361300241599, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.461573600769043, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7155926823616028, + "num_tokens": 71403425.0, + "step": 2764 + }, + { + "epoch": 0.30364594772677356, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.2390568256378174, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6950787901878357, + "num_tokens": 71428908.0, + "step": 2765 + }, + { + "epoch": 0.3037557654293872, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.177565336227417, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.7153798341751099, + "num_tokens": 71454772.0, + "step": 2766 + }, + { + "epoch": 0.3038655831320009, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.296173095703125, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6852425932884216, + "num_tokens": 71482693.0, + "step": 2767 + }, + { + "epoch": 0.30397540083461455, + "ewc_loss": 2.1219253540039062e-05, + "grad_norm": 2.4037656784057617, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7155054807662964, + "num_tokens": 71504932.0, + "step": 2768 + }, + { + "epoch": 0.3040852185372282, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.691744804382324, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7082381248474121, + "num_tokens": 71524792.0, + "step": 2769 + }, + { + "epoch": 0.30419503623984184, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.134195566177368, + "learning_rate": 1e-06, + "loss": 1.1291, + "mean_token_accuracy": 0.6634024381637573, + "num_tokens": 71555938.0, + "step": 2770 + }, + { + "epoch": 0.30430485394245554, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.344332218170166, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6853756308555603, + "num_tokens": 71582278.0, + "step": 2771 + }, + { + "epoch": 0.3044146716450692, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.8862907886505127, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6946631073951721, + "num_tokens": 71611249.0, + "step": 2772 + }, + { + "epoch": 0.30452448934768284, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.1221187114715576, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6959906220436096, + "num_tokens": 71640985.0, + "step": 2773 + }, + { + "epoch": 0.3046343070502965, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.449260950088501, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7055462598800659, + "num_tokens": 71663987.0, + "step": 2774 + }, + { + "epoch": 0.3047441247529102, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.42969012260437, + "learning_rate": 1e-06, + "loss": 1.0772, + "mean_token_accuracy": 0.6819278597831726, + "num_tokens": 71688019.0, + "step": 2775 + }, + { + "epoch": 0.30485394245552383, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.1028709411621094, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7149361968040466, + "num_tokens": 71715223.0, + "step": 2776 + }, + { + "epoch": 0.3049637601581375, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.2248237133026123, + "learning_rate": 1e-06, + "loss": 1.0529, + "mean_token_accuracy": 0.6782480478286743, + "num_tokens": 71742935.0, + "step": 2777 + }, + { + "epoch": 0.3050735778607512, + "ewc_loss": 2.1338462829589844e-05, + "grad_norm": 2.122953414916992, + "learning_rate": 1e-06, + "loss": 1.1274, + "mean_token_accuracy": 0.6691171526908875, + "num_tokens": 71774364.0, + "step": 2778 + }, + { + "epoch": 0.3051833955633648, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 1.9800841808319092, + "learning_rate": 1e-06, + "loss": 1.1288, + "mean_token_accuracy": 0.669556736946106, + "num_tokens": 71808027.0, + "step": 2779 + }, + { + "epoch": 0.30529321326597847, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.3658957481384277, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.709089994430542, + "num_tokens": 71830722.0, + "step": 2780 + }, + { + "epoch": 0.3054030309685921, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.3859174251556396, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.7010642886161804, + "num_tokens": 71852928.0, + "step": 2781 + }, + { + "epoch": 0.3055128486712058, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.198636293411255, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.6918073296546936, + "num_tokens": 71882433.0, + "step": 2782 + }, + { + "epoch": 0.30562266637381946, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 1.9622770547866821, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6868563890457153, + "num_tokens": 71914594.0, + "step": 2783 + }, + { + "epoch": 0.3057324840764331, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.3117620944976807, + "learning_rate": 1e-06, + "loss": 1.0619, + "mean_token_accuracy": 0.6822980642318726, + "num_tokens": 71938171.0, + "step": 2784 + }, + { + "epoch": 0.30584230177904675, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.2171919345855713, + "learning_rate": 1e-06, + "loss": 1.0865, + "mean_token_accuracy": 0.6899933815002441, + "num_tokens": 71966619.0, + "step": 2785 + }, + { + "epoch": 0.30595211948166046, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.1525466442108154, + "learning_rate": 1e-06, + "loss": 1.1103, + "mean_token_accuracy": 0.6676919460296631, + "num_tokens": 71996236.0, + "step": 2786 + }, + { + "epoch": 0.3060619371842741, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.430917978286743, + "learning_rate": 1e-06, + "loss": 1.0891, + "mean_token_accuracy": 0.6768337488174438, + "num_tokens": 72020186.0, + "step": 2787 + }, + { + "epoch": 0.30617175488688775, + "ewc_loss": 2.1457672119140625e-05, + "grad_norm": 2.1941323280334473, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7062413096427917, + "num_tokens": 72045722.0, + "step": 2788 + }, + { + "epoch": 0.30628157258950145, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.3373162746429443, + "learning_rate": 1e-06, + "loss": 1.1036, + "mean_token_accuracy": 0.6736330986022949, + "num_tokens": 72072531.0, + "step": 2789 + }, + { + "epoch": 0.3063913902921151, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.2169668674468994, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6974419355392456, + "num_tokens": 72098928.0, + "step": 2790 + }, + { + "epoch": 0.30650120799472874, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.3937926292419434, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6870871782302856, + "num_tokens": 72123697.0, + "step": 2791 + }, + { + "epoch": 0.3066110256973424, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.634281635284424, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7048567533493042, + "num_tokens": 72144223.0, + "step": 2792 + }, + { + "epoch": 0.3067208433999561, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.3452181816101074, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7075574398040771, + "num_tokens": 72167845.0, + "step": 2793 + }, + { + "epoch": 0.30683066110256974, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.315833330154419, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6994976997375488, + "num_tokens": 72190607.0, + "step": 2794 + }, + { + "epoch": 0.3069404788051834, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.238138437271118, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6914718747138977, + "num_tokens": 72216993.0, + "step": 2795 + }, + { + "epoch": 0.3070502965077971, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.3485233783721924, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7065662145614624, + "num_tokens": 72240619.0, + "step": 2796 + }, + { + "epoch": 0.30716011421041073, + "ewc_loss": 2.1576881408691406e-05, + "grad_norm": 2.579662322998047, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6885228157043457, + "num_tokens": 72261970.0, + "step": 2797 + }, + { + "epoch": 0.3072699319130244, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.536464214324951, + "learning_rate": 1e-06, + "loss": 1.0847, + "mean_token_accuracy": 0.6802189946174622, + "num_tokens": 72286977.0, + "step": 2798 + }, + { + "epoch": 0.307379749615638, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.1930830478668213, + "learning_rate": 1e-06, + "loss": 1.0203, + "mean_token_accuracy": 0.6906090378761292, + "num_tokens": 72312612.0, + "step": 2799 + }, + { + "epoch": 0.3074895673182517, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.0827836990356445, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6843084096908569, + "num_tokens": 72342397.0, + "step": 2800 + }, + { + "epoch": 0.30759938502086537, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.0277044773101807, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6855270266532898, + "num_tokens": 72373431.0, + "step": 2801 + }, + { + "epoch": 0.307709202723479, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.2426295280456543, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6953368782997131, + "num_tokens": 72399898.0, + "step": 2802 + }, + { + "epoch": 0.30781902042609266, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.323805093765259, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7537247538566589, + "num_tokens": 72421490.0, + "step": 2803 + }, + { + "epoch": 0.30792883812870636, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.493961811065674, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7170169353485107, + "num_tokens": 72441399.0, + "step": 2804 + }, + { + "epoch": 0.30803865583132, + "ewc_loss": 2.181529998779297e-05, + "grad_norm": 2.518810272216797, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7206031084060669, + "num_tokens": 72465359.0, + "step": 2805 + }, + { + "epoch": 0.30814847353393365, + "ewc_loss": 2.193450927734375e-05, + "grad_norm": 2.135838747024536, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7073369026184082, + "num_tokens": 72492832.0, + "step": 2806 + }, + { + "epoch": 0.30825829123654735, + "ewc_loss": 2.193450927734375e-05, + "grad_norm": 2.2705111503601074, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6924537420272827, + "num_tokens": 72517740.0, + "step": 2807 + }, + { + "epoch": 0.308368108939161, + "ewc_loss": 2.193450927734375e-05, + "grad_norm": 2.0550215244293213, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7181491851806641, + "num_tokens": 72546634.0, + "step": 2808 + }, + { + "epoch": 0.30847792664177465, + "ewc_loss": 2.193450927734375e-05, + "grad_norm": 2.215888023376465, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.697425901889801, + "num_tokens": 72574884.0, + "step": 2809 + }, + { + "epoch": 0.3085877443443883, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.2965216636657715, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7111800909042358, + "num_tokens": 72599568.0, + "step": 2810 + }, + { + "epoch": 0.308697562047002, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.3907058238983154, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7050924301147461, + "num_tokens": 72623881.0, + "step": 2811 + }, + { + "epoch": 0.30880737974961564, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.003056764602661, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6962067484855652, + "num_tokens": 72656182.0, + "step": 2812 + }, + { + "epoch": 0.3089171974522293, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.1688737869262695, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7148867845535278, + "num_tokens": 72681467.0, + "step": 2813 + }, + { + "epoch": 0.309027015154843, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.474168062210083, + "learning_rate": 1e-06, + "loss": 1.0863, + "mean_token_accuracy": 0.6843811869621277, + "num_tokens": 72704525.0, + "step": 2814 + }, + { + "epoch": 0.30913683285745663, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.277203321456909, + "learning_rate": 1e-06, + "loss": 1.0393, + "mean_token_accuracy": 0.6869522929191589, + "num_tokens": 72729940.0, + "step": 2815 + }, + { + "epoch": 0.3092466505600703, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.2861568927764893, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7093007564544678, + "num_tokens": 72754018.0, + "step": 2816 + }, + { + "epoch": 0.3093564682626839, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.2969882488250732, + "learning_rate": 1e-06, + "loss": 1.0756, + "mean_token_accuracy": 0.6841202974319458, + "num_tokens": 72780723.0, + "step": 2817 + }, + { + "epoch": 0.3094662859652976, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.1937129497528076, + "learning_rate": 1e-06, + "loss": 1.0611, + "mean_token_accuracy": 0.6806184649467468, + "num_tokens": 72809617.0, + "step": 2818 + }, + { + "epoch": 0.3095761036679113, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.5942540168762207, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7373570203781128, + "num_tokens": 72829097.0, + "step": 2819 + }, + { + "epoch": 0.3096859213705249, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.163053035736084, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.70569908618927, + "num_tokens": 72856357.0, + "step": 2820 + }, + { + "epoch": 0.30979573907313857, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.2513740062713623, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.707855224609375, + "num_tokens": 72882233.0, + "step": 2821 + }, + { + "epoch": 0.30990555677575227, + "ewc_loss": 2.205371856689453e-05, + "grad_norm": 2.3648030757904053, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7102674245834351, + "num_tokens": 72905153.0, + "step": 2822 + }, + { + "epoch": 0.3100153744783659, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.5530617237091064, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7177774906158447, + "num_tokens": 72926296.0, + "step": 2823 + }, + { + "epoch": 0.31012519218097956, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.108308792114258, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6910134553909302, + "num_tokens": 72955950.0, + "step": 2824 + }, + { + "epoch": 0.31023500988359326, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.355536460876465, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6820985078811646, + "num_tokens": 72982794.0, + "step": 2825 + }, + { + "epoch": 0.3103448275862069, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.3382487297058105, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7002490758895874, + "num_tokens": 73007331.0, + "step": 2826 + }, + { + "epoch": 0.31045464528882055, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.5629470348358154, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6959168910980225, + "num_tokens": 73027103.0, + "step": 2827 + }, + { + "epoch": 0.3105644629914342, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.4257943630218506, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6958218216896057, + "num_tokens": 73049063.0, + "step": 2828 + }, + { + "epoch": 0.3106742806940479, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.0173871517181396, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6957231760025024, + "num_tokens": 73078869.0, + "step": 2829 + }, + { + "epoch": 0.31078409839666155, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.292205333709717, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6984171867370605, + "num_tokens": 73105191.0, + "step": 2830 + }, + { + "epoch": 0.3108939160992752, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.1928117275238037, + "learning_rate": 1e-06, + "loss": 1.062, + "mean_token_accuracy": 0.6895335912704468, + "num_tokens": 73131768.0, + "step": 2831 + }, + { + "epoch": 0.31100373380188884, + "ewc_loss": 2.2172927856445312e-05, + "grad_norm": 2.4722394943237305, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6987114548683167, + "num_tokens": 73158912.0, + "step": 2832 + }, + { + "epoch": 0.31111355150450254, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.3606276512145996, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7008543610572815, + "num_tokens": 73183322.0, + "step": 2833 + }, + { + "epoch": 0.3112233692071162, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.055208206176758, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.699325680732727, + "num_tokens": 73212338.0, + "step": 2834 + }, + { + "epoch": 0.31133318690972983, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.230537176132202, + "learning_rate": 1e-06, + "loss": 1.0338, + "mean_token_accuracy": 0.6901431083679199, + "num_tokens": 73241951.0, + "step": 2835 + }, + { + "epoch": 0.31144300461234353, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.162827968597412, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7180370688438416, + "num_tokens": 73269042.0, + "step": 2836 + }, + { + "epoch": 0.3115528223149572, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.2455739974975586, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7015409469604492, + "num_tokens": 73295104.0, + "step": 2837 + }, + { + "epoch": 0.3116626400175708, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.1656525135040283, + "learning_rate": 1e-06, + "loss": 1.066, + "mean_token_accuracy": 0.6842523813247681, + "num_tokens": 73322207.0, + "step": 2838 + }, + { + "epoch": 0.31177245772018447, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.4358277320861816, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.6908100843429565, + "num_tokens": 73344542.0, + "step": 2839 + }, + { + "epoch": 0.31188227542279817, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.517568826675415, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.7044293880462646, + "num_tokens": 73366839.0, + "step": 2840 + }, + { + "epoch": 0.3119920931254118, + "ewc_loss": 2.2292137145996094e-05, + "grad_norm": 2.385911703109741, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6983014941215515, + "num_tokens": 73390568.0, + "step": 2841 + }, + { + "epoch": 0.31210191082802546, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.4572649002075195, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7102348804473877, + "num_tokens": 73411547.0, + "step": 2842 + }, + { + "epoch": 0.31221172853063917, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.4642906188964844, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6958029270172119, + "num_tokens": 73436034.0, + "step": 2843 + }, + { + "epoch": 0.3123215462332528, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.3163552284240723, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6933982372283936, + "num_tokens": 73461319.0, + "step": 2844 + }, + { + "epoch": 0.31243136393586646, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.4212281703948975, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6954742670059204, + "num_tokens": 73483607.0, + "step": 2845 + }, + { + "epoch": 0.3125411816384801, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.1995224952697754, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.681739330291748, + "num_tokens": 73510148.0, + "step": 2846 + }, + { + "epoch": 0.3126509993410938, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.7982189655303955, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7383860945701599, + "num_tokens": 73527606.0, + "step": 2847 + }, + { + "epoch": 0.31276081704370745, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.4700214862823486, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7015900015830994, + "num_tokens": 73548554.0, + "step": 2848 + }, + { + "epoch": 0.3128706347463211, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.074923038482666, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7255322933197021, + "num_tokens": 73575796.0, + "step": 2849 + }, + { + "epoch": 0.31298045244893474, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.133702278137207, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.699766993522644, + "num_tokens": 73605313.0, + "step": 2850 + }, + { + "epoch": 0.31309027015154844, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.337752342224121, + "learning_rate": 1e-06, + "loss": 1.0277, + "mean_token_accuracy": 0.6899622082710266, + "num_tokens": 73630136.0, + "step": 2851 + }, + { + "epoch": 0.3132000878541621, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.51725435256958, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.712795615196228, + "num_tokens": 73650701.0, + "step": 2852 + }, + { + "epoch": 0.31330990555677574, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.371307611465454, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.71943598985672, + "num_tokens": 73675451.0, + "step": 2853 + }, + { + "epoch": 0.31341972325938944, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.413846015930176, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6924409866333008, + "num_tokens": 73699631.0, + "step": 2854 + }, + { + "epoch": 0.3135295409620031, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.038384199142456, + "learning_rate": 1e-06, + "loss": 1.1448, + "mean_token_accuracy": 0.668519914150238, + "num_tokens": 73734841.0, + "step": 2855 + }, + { + "epoch": 0.31363935866461673, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.086101531982422, + "learning_rate": 1e-06, + "loss": 1.0631, + "mean_token_accuracy": 0.6778174638748169, + "num_tokens": 73766927.0, + "step": 2856 + }, + { + "epoch": 0.3137491763672304, + "ewc_loss": 2.2411346435546875e-05, + "grad_norm": 2.1857454776763916, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6972770690917969, + "num_tokens": 73794957.0, + "step": 2857 + }, + { + "epoch": 0.3138589940698441, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.607492685317993, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7011572122573853, + "num_tokens": 73814952.0, + "step": 2858 + }, + { + "epoch": 0.3139688117724577, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.1464180946350098, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6929745674133301, + "num_tokens": 73840972.0, + "step": 2859 + }, + { + "epoch": 0.31407862947507137, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.0399458408355713, + "learning_rate": 1e-06, + "loss": 1.0454, + "mean_token_accuracy": 0.6814696788787842, + "num_tokens": 73871707.0, + "step": 2860 + }, + { + "epoch": 0.314188447177685, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.4369418621063232, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7060168385505676, + "num_tokens": 73896564.0, + "step": 2861 + }, + { + "epoch": 0.3142982648802987, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.213887929916382, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7035818099975586, + "num_tokens": 73923668.0, + "step": 2862 + }, + { + "epoch": 0.31440808258291236, + "ewc_loss": 2.2530555725097656e-05, + "grad_norm": 2.1775567531585693, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6873382925987244, + "num_tokens": 73949813.0, + "step": 2863 + }, + { + "epoch": 0.314517900285526, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.4898183345794678, + "learning_rate": 1e-06, + "loss": 1.0637, + "mean_token_accuracy": 0.6888289451599121, + "num_tokens": 73972394.0, + "step": 2864 + }, + { + "epoch": 0.3146277179881397, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.261570692062378, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6959919929504395, + "num_tokens": 73996907.0, + "step": 2865 + }, + { + "epoch": 0.31473753569075336, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.119258403778076, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.724765956401825, + "num_tokens": 74023279.0, + "step": 2866 + }, + { + "epoch": 0.314847353393367, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.048323392868042, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6995440125465393, + "num_tokens": 74055139.0, + "step": 2867 + }, + { + "epoch": 0.31495717109598065, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.4996018409729004, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7031097412109375, + "num_tokens": 74077381.0, + "step": 2868 + }, + { + "epoch": 0.31506698879859435, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.3684473037719727, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6977965831756592, + "num_tokens": 74102507.0, + "step": 2869 + }, + { + "epoch": 0.315176806501208, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.1669063568115234, + "learning_rate": 1e-06, + "loss": 1.0489, + "mean_token_accuracy": 0.6836527585983276, + "num_tokens": 74131402.0, + "step": 2870 + }, + { + "epoch": 0.31528662420382164, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.666701316833496, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7142596244812012, + "num_tokens": 74150490.0, + "step": 2871 + }, + { + "epoch": 0.31539644190643534, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.2141175270080566, + "learning_rate": 1e-06, + "loss": 1.066, + "mean_token_accuracy": 0.679966151714325, + "num_tokens": 74177232.0, + "step": 2872 + }, + { + "epoch": 0.315506259609049, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.34653902053833, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7168023586273193, + "num_tokens": 74201663.0, + "step": 2873 + }, + { + "epoch": 0.31561607731166264, + "ewc_loss": 2.2649765014648438e-05, + "grad_norm": 2.3163559436798096, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7023289203643799, + "num_tokens": 74225721.0, + "step": 2874 + }, + { + "epoch": 0.3157258950142763, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.077644109725952, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.7017101645469666, + "num_tokens": 74254833.0, + "step": 2875 + }, + { + "epoch": 0.31583571271689, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.2585983276367188, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6974811553955078, + "num_tokens": 74280086.0, + "step": 2876 + }, + { + "epoch": 0.31594553041950363, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.19867205619812, + "learning_rate": 1e-06, + "loss": 1.1249, + "mean_token_accuracy": 0.6796655654907227, + "num_tokens": 74308202.0, + "step": 2877 + }, + { + "epoch": 0.3160553481221173, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 1.969391107559204, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6963748931884766, + "num_tokens": 74340124.0, + "step": 2878 + }, + { + "epoch": 0.3161651658247309, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.4564733505249023, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7028775215148926, + "num_tokens": 74363775.0, + "step": 2879 + }, + { + "epoch": 0.3162749835273446, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.3998823165893555, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7136480808258057, + "num_tokens": 74385059.0, + "step": 2880 + }, + { + "epoch": 0.31638480122995827, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 1.991507649421692, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6924533843994141, + "num_tokens": 74415274.0, + "step": 2881 + }, + { + "epoch": 0.3164946189325719, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 3.077152967453003, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6971930861473083, + "num_tokens": 74433509.0, + "step": 2882 + }, + { + "epoch": 0.3166044366351856, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.255600929260254, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.7093107104301453, + "num_tokens": 74456649.0, + "step": 2883 + }, + { + "epoch": 0.31671425433779926, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.5119340419769287, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6973146796226501, + "num_tokens": 74478535.0, + "step": 2884 + }, + { + "epoch": 0.3168240720404129, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.3537709712982178, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.710995078086853, + "num_tokens": 74501778.0, + "step": 2885 + }, + { + "epoch": 0.31693388974302655, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.1588661670684814, + "learning_rate": 1e-06, + "loss": 1.0903, + "mean_token_accuracy": 0.6750245690345764, + "num_tokens": 74530204.0, + "step": 2886 + }, + { + "epoch": 0.31704370744564025, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.3688747882843018, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.737634003162384, + "num_tokens": 74552726.0, + "step": 2887 + }, + { + "epoch": 0.3171535251482539, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.3933849334716797, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.7048961520195007, + "num_tokens": 74574343.0, + "step": 2888 + }, + { + "epoch": 0.31726334285086755, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.3424835205078125, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.706221342086792, + "num_tokens": 74597489.0, + "step": 2889 + }, + { + "epoch": 0.31737316055348125, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 1.9140368700027466, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6998311281204224, + "num_tokens": 74630000.0, + "step": 2890 + }, + { + "epoch": 0.3174829782560949, + "ewc_loss": 2.276897430419922e-05, + "grad_norm": 2.246511697769165, + "learning_rate": 1e-06, + "loss": 1.0963, + "mean_token_accuracy": 0.6785062551498413, + "num_tokens": 74656838.0, + "step": 2891 + }, + { + "epoch": 0.31759279595870854, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.2727482318878174, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6933825016021729, + "num_tokens": 74680459.0, + "step": 2892 + }, + { + "epoch": 0.3177026136613222, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.0889947414398193, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7074062824249268, + "num_tokens": 74709266.0, + "step": 2893 + }, + { + "epoch": 0.3178124313639359, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.0876879692077637, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7162353992462158, + "num_tokens": 74737699.0, + "step": 2894 + }, + { + "epoch": 0.31792224906654953, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.4427545070648193, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7074002623558044, + "num_tokens": 74761079.0, + "step": 2895 + }, + { + "epoch": 0.3180320667691632, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.0151243209838867, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.70183265209198, + "num_tokens": 74790938.0, + "step": 2896 + }, + { + "epoch": 0.3181418844717768, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.09033203125, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7078354358673096, + "num_tokens": 74819771.0, + "step": 2897 + }, + { + "epoch": 0.3182517021743905, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.3168461322784424, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6982403993606567, + "num_tokens": 74845767.0, + "step": 2898 + }, + { + "epoch": 0.3183615198770042, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.133606433868408, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7060837745666504, + "num_tokens": 74873917.0, + "step": 2899 + }, + { + "epoch": 0.3184713375796178, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.252804756164551, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7161735892295837, + "num_tokens": 74898683.0, + "step": 2900 + }, + { + "epoch": 0.3185811552822315, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.2986598014831543, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6929740309715271, + "num_tokens": 74923942.0, + "step": 2901 + }, + { + "epoch": 0.31869097298484517, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.1473121643066406, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7103639841079712, + "num_tokens": 74950771.0, + "step": 2902 + }, + { + "epoch": 0.3188007906874588, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.4685497283935547, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6956192255020142, + "num_tokens": 74973253.0, + "step": 2903 + }, + { + "epoch": 0.31891060839007246, + "ewc_loss": 2.300739288330078e-05, + "grad_norm": 2.269855260848999, + "learning_rate": 1e-06, + "loss": 1.0799, + "mean_token_accuracy": 0.6793640851974487, + "num_tokens": 74999470.0, + "step": 2904 + }, + { + "epoch": 0.31902042609268616, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.634479284286499, + "learning_rate": 1e-06, + "loss": 1.0445, + "mean_token_accuracy": 0.6882386207580566, + "num_tokens": 75019191.0, + "step": 2905 + }, + { + "epoch": 0.3191302437952998, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.0985169410705566, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6864224672317505, + "num_tokens": 75047634.0, + "step": 2906 + }, + { + "epoch": 0.31924006149791345, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.0844898223876953, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7053178548812866, + "num_tokens": 75077206.0, + "step": 2907 + }, + { + "epoch": 0.3193498792005271, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.330328941345215, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.7007828950881958, + "num_tokens": 75103531.0, + "step": 2908 + }, + { + "epoch": 0.3194596969031408, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 1.9899115562438965, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7135206460952759, + "num_tokens": 75135218.0, + "step": 2909 + }, + { + "epoch": 0.31956951460575445, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.275174856185913, + "learning_rate": 1e-06, + "loss": 1.0736, + "mean_token_accuracy": 0.6718542575836182, + "num_tokens": 75162189.0, + "step": 2910 + }, + { + "epoch": 0.3196793323083681, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.4220974445343018, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7264107465744019, + "num_tokens": 75183522.0, + "step": 2911 + }, + { + "epoch": 0.3197891500109818, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.0829689502716064, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7037561535835266, + "num_tokens": 75211860.0, + "step": 2912 + }, + { + "epoch": 0.31989896771359544, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 1.8045458793640137, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6851258277893066, + "num_tokens": 75250506.0, + "step": 2913 + }, + { + "epoch": 0.3200087854162091, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 1.9666383266448975, + "learning_rate": 1e-06, + "loss": 1.1104, + "mean_token_accuracy": 0.6740204095840454, + "num_tokens": 75285783.0, + "step": 2914 + }, + { + "epoch": 0.32011860311882273, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.006333112716675, + "learning_rate": 1e-06, + "loss": 1.1419, + "mean_token_accuracy": 0.6568835973739624, + "num_tokens": 75318059.0, + "step": 2915 + }, + { + "epoch": 0.32022842082143643, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.2903311252593994, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.695476770401001, + "num_tokens": 75345039.0, + "step": 2916 + }, + { + "epoch": 0.3203382385240501, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.046020030975342, + "learning_rate": 1e-06, + "loss": 1.0899, + "mean_token_accuracy": 0.6831006407737732, + "num_tokens": 75374543.0, + "step": 2917 + }, + { + "epoch": 0.3204480562266637, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.4625229835510254, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.7032634019851685, + "num_tokens": 75397039.0, + "step": 2918 + }, + { + "epoch": 0.3205578739292774, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.2562854290008545, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6902481317520142, + "num_tokens": 75422381.0, + "step": 2919 + }, + { + "epoch": 0.32066769163189107, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.3920929431915283, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.721002459526062, + "num_tokens": 75446255.0, + "step": 2920 + }, + { + "epoch": 0.3207775093345047, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.3976528644561768, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7206209301948547, + "num_tokens": 75468639.0, + "step": 2921 + }, + { + "epoch": 0.32088732703711836, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.207268714904785, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7211436629295349, + "num_tokens": 75492994.0, + "step": 2922 + }, + { + "epoch": 0.32099714473973207, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.135974884033203, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.6912968754768372, + "num_tokens": 75520922.0, + "step": 2923 + }, + { + "epoch": 0.3211069624423457, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.1170554161071777, + "learning_rate": 1e-06, + "loss": 1.0695, + "mean_token_accuracy": 0.688843846321106, + "num_tokens": 75548894.0, + "step": 2924 + }, + { + "epoch": 0.32121678014495936, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.4833834171295166, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.694881796836853, + "num_tokens": 75569737.0, + "step": 2925 + }, + { + "epoch": 0.321326597847573, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.1813583374023438, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7006450295448303, + "num_tokens": 75594651.0, + "step": 2926 + }, + { + "epoch": 0.3214364155501867, + "ewc_loss": 2.3126602172851562e-05, + "grad_norm": 2.1724953651428223, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.6989402770996094, + "num_tokens": 75620646.0, + "step": 2927 + }, + { + "epoch": 0.32154623325280035, + "ewc_loss": 2.3245811462402344e-05, + "grad_norm": 2.333340644836426, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7147439122200012, + "num_tokens": 75642082.0, + "step": 2928 + }, + { + "epoch": 0.321656050955414, + "ewc_loss": 2.3245811462402344e-05, + "grad_norm": 2.0747907161712646, + "learning_rate": 1e-06, + "loss": 1.0442, + "mean_token_accuracy": 0.6895505785942078, + "num_tokens": 75671471.0, + "step": 2929 + }, + { + "epoch": 0.3217658686580277, + "ewc_loss": 2.3245811462402344e-05, + "grad_norm": 2.224018096923828, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6862406730651855, + "num_tokens": 75698969.0, + "step": 2930 + }, + { + "epoch": 0.32187568636064134, + "ewc_loss": 2.3245811462402344e-05, + "grad_norm": 2.120607852935791, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7259605526924133, + "num_tokens": 75726030.0, + "step": 2931 + }, + { + "epoch": 0.321985504063255, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.350005626678467, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7118099927902222, + "num_tokens": 75749806.0, + "step": 2932 + }, + { + "epoch": 0.32209532176586864, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.4971389770507812, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7027842998504639, + "num_tokens": 75774356.0, + "step": 2933 + }, + { + "epoch": 0.32220513946848234, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.2098896503448486, + "learning_rate": 1e-06, + "loss": 1.0452, + "mean_token_accuracy": 0.6865798830986023, + "num_tokens": 75801477.0, + "step": 2934 + }, + { + "epoch": 0.322314957171096, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.247807025909424, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7121579647064209, + "num_tokens": 75826278.0, + "step": 2935 + }, + { + "epoch": 0.32242477487370963, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.2143213748931885, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7006477117538452, + "num_tokens": 75853384.0, + "step": 2936 + }, + { + "epoch": 0.3225345925763233, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.496936082839966, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7165152430534363, + "num_tokens": 75875598.0, + "step": 2937 + }, + { + "epoch": 0.322644410278937, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.2078726291656494, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.6992293000221252, + "num_tokens": 75901014.0, + "step": 2938 + }, + { + "epoch": 0.3227542279815506, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.2306134700775146, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.706505537033081, + "num_tokens": 75927081.0, + "step": 2939 + }, + { + "epoch": 0.32286404568416427, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.472461223602295, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.736635684967041, + "num_tokens": 75947874.0, + "step": 2940 + }, + { + "epoch": 0.32297386338677797, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.230259656906128, + "learning_rate": 1e-06, + "loss": 1.0878, + "mean_token_accuracy": 0.6773456931114197, + "num_tokens": 75974473.0, + "step": 2941 + }, + { + "epoch": 0.3230836810893916, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.4350638389587402, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7042466402053833, + "num_tokens": 75997171.0, + "step": 2942 + }, + { + "epoch": 0.32319349879200526, + "ewc_loss": 2.3365020751953125e-05, + "grad_norm": 2.4505093097686768, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.7056190967559814, + "num_tokens": 76020257.0, + "step": 2943 + }, + { + "epoch": 0.3233033164946189, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.351156711578369, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7068004608154297, + "num_tokens": 76042908.0, + "step": 2944 + }, + { + "epoch": 0.3234131341972326, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.115910530090332, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6989632844924927, + "num_tokens": 76071996.0, + "step": 2945 + }, + { + "epoch": 0.32352295189984626, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.264092445373535, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7230380773544312, + "num_tokens": 76095474.0, + "step": 2946 + }, + { + "epoch": 0.3236327696024599, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.373603105545044, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7106336355209351, + "num_tokens": 76119182.0, + "step": 2947 + }, + { + "epoch": 0.3237425873050736, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.323955535888672, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7151806354522705, + "num_tokens": 76141518.0, + "step": 2948 + }, + { + "epoch": 0.32385240500768725, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.139328956604004, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6985732913017273, + "num_tokens": 76166749.0, + "step": 2949 + }, + { + "epoch": 0.3239622227103009, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.129847526550293, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7211214900016785, + "num_tokens": 76194031.0, + "step": 2950 + }, + { + "epoch": 0.32407204041291454, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.381970167160034, + "learning_rate": 1e-06, + "loss": 1.1282, + "mean_token_accuracy": 0.6722413897514343, + "num_tokens": 76220946.0, + "step": 2951 + }, + { + "epoch": 0.32418185811552824, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.207028865814209, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.6996739506721497, + "num_tokens": 76247852.0, + "step": 2952 + }, + { + "epoch": 0.3242916758181419, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.6992123126983643, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7458739876747131, + "num_tokens": 76265952.0, + "step": 2953 + }, + { + "epoch": 0.32440149352075554, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.0799734592437744, + "learning_rate": 1e-06, + "loss": 1.0665, + "mean_token_accuracy": 0.6888275742530823, + "num_tokens": 76295479.0, + "step": 2954 + }, + { + "epoch": 0.3245113112233692, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.4067416191101074, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7134109139442444, + "num_tokens": 76319832.0, + "step": 2955 + }, + { + "epoch": 0.3246211289259829, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.3109171390533447, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.69899582862854, + "num_tokens": 76346144.0, + "step": 2956 + }, + { + "epoch": 0.32473094662859653, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.2071170806884766, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7142654657363892, + "num_tokens": 76373994.0, + "step": 2957 + }, + { + "epoch": 0.3248407643312102, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.1483640670776367, + "learning_rate": 1e-06, + "loss": 1.1161, + "mean_token_accuracy": 0.668064534664154, + "num_tokens": 76404158.0, + "step": 2958 + }, + { + "epoch": 0.3249505820338239, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 1.8843529224395752, + "learning_rate": 1e-06, + "loss": 1.0595, + "mean_token_accuracy": 0.6898930072784424, + "num_tokens": 76440125.0, + "step": 2959 + }, + { + "epoch": 0.3250603997364375, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.4090700149536133, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6841031312942505, + "num_tokens": 76464590.0, + "step": 2960 + }, + { + "epoch": 0.32517021743905117, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.3276379108428955, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6855335235595703, + "num_tokens": 76488566.0, + "step": 2961 + }, + { + "epoch": 0.3252800351416648, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.1148324012756348, + "learning_rate": 1e-06, + "loss": 1.1051, + "mean_token_accuracy": 0.6829125285148621, + "num_tokens": 76517572.0, + "step": 2962 + }, + { + "epoch": 0.3253898528442785, + "ewc_loss": 2.372264862060547e-05, + "grad_norm": 2.417336940765381, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6917012929916382, + "num_tokens": 76539776.0, + "step": 2963 + }, + { + "epoch": 0.32549967054689216, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.170992136001587, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7201811075210571, + "num_tokens": 76566605.0, + "step": 2964 + }, + { + "epoch": 0.3256094882495058, + "ewc_loss": 2.3603439331054688e-05, + "grad_norm": 2.2051403522491455, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7112257480621338, + "num_tokens": 76591486.0, + "step": 2965 + }, + { + "epoch": 0.3257193059521195, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.498622417449951, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7053579092025757, + "num_tokens": 76611866.0, + "step": 2966 + }, + { + "epoch": 0.32582912365473315, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.246934413909912, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6992113590240479, + "num_tokens": 76637126.0, + "step": 2967 + }, + { + "epoch": 0.3259389413573468, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.4549560546875, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6993564963340759, + "num_tokens": 76659032.0, + "step": 2968 + }, + { + "epoch": 0.32604875905996045, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.544269323348999, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7039417028427124, + "num_tokens": 76681382.0, + "step": 2969 + }, + { + "epoch": 0.32615857676257415, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.469078779220581, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.7029145956039429, + "num_tokens": 76703872.0, + "step": 2970 + }, + { + "epoch": 0.3262683944651878, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.1119134426116943, + "learning_rate": 1e-06, + "loss": 1.115, + "mean_token_accuracy": 0.6685454845428467, + "num_tokens": 76732208.0, + "step": 2971 + }, + { + "epoch": 0.32637821216780144, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.0852437019348145, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6884779930114746, + "num_tokens": 76761485.0, + "step": 2972 + }, + { + "epoch": 0.3264880298704151, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.110694646835327, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6991206407546997, + "num_tokens": 76790799.0, + "step": 2973 + }, + { + "epoch": 0.3265978475730288, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.3660695552825928, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6927222013473511, + "num_tokens": 76814140.0, + "step": 2974 + }, + { + "epoch": 0.32670766527564243, + "ewc_loss": 2.384185791015625e-05, + "grad_norm": 2.408742666244507, + "learning_rate": 1e-06, + "loss": 1.0506, + "mean_token_accuracy": 0.6919451355934143, + "num_tokens": 76836439.0, + "step": 2975 + }, + { + "epoch": 0.3268174829782561, + "ewc_loss": 2.396106719970703e-05, + "grad_norm": 2.4262821674346924, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7061296105384827, + "num_tokens": 76857722.0, + "step": 2976 + }, + { + "epoch": 0.3269273006808698, + "ewc_loss": 2.396106719970703e-05, + "grad_norm": 2.3522047996520996, + "learning_rate": 1e-06, + "loss": 1.1372, + "mean_token_accuracy": 0.6672409772872925, + "num_tokens": 76883370.0, + "step": 2977 + }, + { + "epoch": 0.3270371183834834, + "ewc_loss": 2.396106719970703e-05, + "grad_norm": 2.4194514751434326, + "learning_rate": 1e-06, + "loss": 1.0549, + "mean_token_accuracy": 0.6774511933326721, + "num_tokens": 76907397.0, + "step": 2978 + }, + { + "epoch": 0.3271469360860971, + "ewc_loss": 2.396106719970703e-05, + "grad_norm": 2.3078856468200684, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6974721550941467, + "num_tokens": 76931225.0, + "step": 2979 + }, + { + "epoch": 0.3272567537887107, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.3043885231018066, + "learning_rate": 1e-06, + "loss": 1.0886, + "mean_token_accuracy": 0.6708062887191772, + "num_tokens": 76956296.0, + "step": 2980 + }, + { + "epoch": 0.3273665714913244, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.243694305419922, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7155905365943909, + "num_tokens": 76979668.0, + "step": 2981 + }, + { + "epoch": 0.32747638919393807, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.867521047592163, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7008965015411377, + "num_tokens": 76999757.0, + "step": 2982 + }, + { + "epoch": 0.3275862068965517, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 1.8958624601364136, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.710732102394104, + "num_tokens": 77031664.0, + "step": 2983 + }, + { + "epoch": 0.32769602459916536, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.6594491004943848, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.6982294321060181, + "num_tokens": 77050012.0, + "step": 2984 + }, + { + "epoch": 0.32780584230177906, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.453721046447754, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.6974841356277466, + "num_tokens": 77071442.0, + "step": 2985 + }, + { + "epoch": 0.3279156600043927, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.284057855606079, + "learning_rate": 1e-06, + "loss": 1.0529, + "mean_token_accuracy": 0.6915501952171326, + "num_tokens": 77098835.0, + "step": 2986 + }, + { + "epoch": 0.32802547770700635, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.1624269485473633, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7126833200454712, + "num_tokens": 77125012.0, + "step": 2987 + }, + { + "epoch": 0.32813529540962005, + "ewc_loss": 2.4199485778808594e-05, + "grad_norm": 2.3064980506896973, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7265394330024719, + "num_tokens": 77148158.0, + "step": 2988 + }, + { + "epoch": 0.3282451131122337, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.2826623916625977, + "learning_rate": 1e-06, + "loss": 1.0641, + "mean_token_accuracy": 0.6819236874580383, + "num_tokens": 77175883.0, + "step": 2989 + }, + { + "epoch": 0.32835493081484735, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.442167043685913, + "learning_rate": 1e-06, + "loss": 1.109, + "mean_token_accuracy": 0.6711908578872681, + "num_tokens": 77200164.0, + "step": 2990 + }, + { + "epoch": 0.328464748517461, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.245408296585083, + "learning_rate": 1e-06, + "loss": 1.0925, + "mean_token_accuracy": 0.6786428093910217, + "num_tokens": 77226914.0, + "step": 2991 + }, + { + "epoch": 0.3285745662200747, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.1931045055389404, + "learning_rate": 1e-06, + "loss": 1.045, + "mean_token_accuracy": 0.6957470774650574, + "num_tokens": 77253262.0, + "step": 2992 + }, + { + "epoch": 0.32868438392268834, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.342247724533081, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7309280633926392, + "num_tokens": 77280602.0, + "step": 2993 + }, + { + "epoch": 0.328794201625302, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.1614952087402344, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7023254036903381, + "num_tokens": 77308632.0, + "step": 2994 + }, + { + "epoch": 0.3289040193279157, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.524226427078247, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6962102055549622, + "num_tokens": 77331211.0, + "step": 2995 + }, + { + "epoch": 0.32901383703052933, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.1230661869049072, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7168990969657898, + "num_tokens": 77358266.0, + "step": 2996 + }, + { + "epoch": 0.329123654733143, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.324199676513672, + "learning_rate": 1e-06, + "loss": 1.0885, + "mean_token_accuracy": 0.6889735460281372, + "num_tokens": 77382026.0, + "step": 2997 + }, + { + "epoch": 0.3292334724357566, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 1.9579432010650635, + "learning_rate": 1e-06, + "loss": 1.1239, + "mean_token_accuracy": 0.6726570725440979, + "num_tokens": 77420022.0, + "step": 2998 + }, + { + "epoch": 0.3293432901383703, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.020969867706299, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6910820603370667, + "num_tokens": 77451856.0, + "step": 2999 + }, + { + "epoch": 0.32945310784098397, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.0834295749664307, + "learning_rate": 1e-06, + "loss": 1.0514, + "mean_token_accuracy": 0.6826743483543396, + "num_tokens": 77481883.0, + "step": 3000 + }, + { + "epoch": 0.3295629255435976, + "ewc_loss": 2.4318695068359375e-05, + "grad_norm": 2.6242423057556152, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.718290388584137, + "num_tokens": 77501260.0, + "step": 3001 + }, + { + "epoch": 0.32967274324621126, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.452664852142334, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.706531822681427, + "num_tokens": 77522736.0, + "step": 3002 + }, + { + "epoch": 0.32978256094882497, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.1132090091705322, + "learning_rate": 1e-06, + "loss": 1.1272, + "mean_token_accuracy": 0.6603951454162598, + "num_tokens": 77552652.0, + "step": 3003 + }, + { + "epoch": 0.3298923786514386, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.1673905849456787, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7254571914672852, + "num_tokens": 77577360.0, + "step": 3004 + }, + { + "epoch": 0.33000219635405226, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.03252911567688, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7094206213951111, + "num_tokens": 77606752.0, + "step": 3005 + }, + { + "epoch": 0.33011201405666596, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.3723533153533936, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7096368074417114, + "num_tokens": 77628709.0, + "step": 3006 + }, + { + "epoch": 0.3302218317592796, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.141824960708618, + "learning_rate": 1e-06, + "loss": 1.0569, + "mean_token_accuracy": 0.680827260017395, + "num_tokens": 77656252.0, + "step": 3007 + }, + { + "epoch": 0.33033164946189325, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.250239133834839, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.7015699148178101, + "num_tokens": 77681961.0, + "step": 3008 + }, + { + "epoch": 0.3304414671645069, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.1584832668304443, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7013965845108032, + "num_tokens": 77709226.0, + "step": 3009 + }, + { + "epoch": 0.3305512848671206, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.631405830383301, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7182406187057495, + "num_tokens": 77726517.0, + "step": 3010 + }, + { + "epoch": 0.33066110256973424, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.2443959712982178, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.7005031704902649, + "num_tokens": 77752521.0, + "step": 3011 + }, + { + "epoch": 0.3307709202723479, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.2215969562530518, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6976823806762695, + "num_tokens": 77779017.0, + "step": 3012 + }, + { + "epoch": 0.33088073797496154, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.3259315490722656, + "learning_rate": 1e-06, + "loss": 1.0642, + "mean_token_accuracy": 0.6786312460899353, + "num_tokens": 77803179.0, + "step": 3013 + }, + { + "epoch": 0.33099055567757524, + "ewc_loss": 2.4437904357910156e-05, + "grad_norm": 2.117161512374878, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.715882420539856, + "num_tokens": 77830529.0, + "step": 3014 + }, + { + "epoch": 0.3311003733801889, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.341724395751953, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6977571249008179, + "num_tokens": 77857041.0, + "step": 3015 + }, + { + "epoch": 0.33121019108280253, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.4669837951660156, + "learning_rate": 1e-06, + "loss": 1.0813, + "mean_token_accuracy": 0.6807331442832947, + "num_tokens": 77881178.0, + "step": 3016 + }, + { + "epoch": 0.33132000878541623, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.1879971027374268, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.6824103593826294, + "num_tokens": 77910440.0, + "step": 3017 + }, + { + "epoch": 0.3314298264880299, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.091562509536743, + "learning_rate": 1e-06, + "loss": 1.0852, + "mean_token_accuracy": 0.6773552298545837, + "num_tokens": 77941588.0, + "step": 3018 + }, + { + "epoch": 0.3315396441906435, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.506866216659546, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7141398191452026, + "num_tokens": 77962746.0, + "step": 3019 + }, + { + "epoch": 0.33164946189325717, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.4649977684020996, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.6885761022567749, + "num_tokens": 77985228.0, + "step": 3020 + }, + { + "epoch": 0.33175927959587087, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.1393344402313232, + "learning_rate": 1e-06, + "loss": 1.08, + "mean_token_accuracy": 0.6755198836326599, + "num_tokens": 78015834.0, + "step": 3021 + }, + { + "epoch": 0.3318690972984845, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.5018301010131836, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.7034853100776672, + "num_tokens": 78038463.0, + "step": 3022 + }, + { + "epoch": 0.33197891500109816, + "ewc_loss": 2.4557113647460938e-05, + "grad_norm": 2.061234474182129, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6957305073738098, + "num_tokens": 78068067.0, + "step": 3023 + }, + { + "epoch": 0.33208873270371186, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.123253345489502, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6987844705581665, + "num_tokens": 78099716.0, + "step": 3024 + }, + { + "epoch": 0.3321985504063255, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.186809539794922, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.729518711566925, + "num_tokens": 78123803.0, + "step": 3025 + }, + { + "epoch": 0.33230836810893916, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.3320915699005127, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7099748253822327, + "num_tokens": 78148900.0, + "step": 3026 + }, + { + "epoch": 0.3324181858115528, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.626103639602661, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7043185830116272, + "num_tokens": 78172082.0, + "step": 3027 + }, + { + "epoch": 0.3325280035141665, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.2729735374450684, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7156972885131836, + "num_tokens": 78197140.0, + "step": 3028 + }, + { + "epoch": 0.33263782121678015, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.0277137756347656, + "learning_rate": 1e-06, + "loss": 1.0782, + "mean_token_accuracy": 0.6780177354812622, + "num_tokens": 78229344.0, + "step": 3029 + }, + { + "epoch": 0.3327476389193938, + "ewc_loss": 2.467632293701172e-05, + "grad_norm": 2.1064796447753906, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6958589553833008, + "num_tokens": 78257350.0, + "step": 3030 + }, + { + "epoch": 0.33285745662200744, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.039552688598633, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7092050313949585, + "num_tokens": 78286663.0, + "step": 3031 + }, + { + "epoch": 0.33296727432462114, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.0461041927337646, + "learning_rate": 1e-06, + "loss": 1.0474, + "mean_token_accuracy": 0.6880553960800171, + "num_tokens": 78318532.0, + "step": 3032 + }, + { + "epoch": 0.3330770920272348, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.3007915019989014, + "learning_rate": 1e-06, + "loss": 1.0944, + "mean_token_accuracy": 0.6729416251182556, + "num_tokens": 78342676.0, + "step": 3033 + }, + { + "epoch": 0.33318690972984844, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.341242790222168, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7044049501419067, + "num_tokens": 78364760.0, + "step": 3034 + }, + { + "epoch": 0.33329672743246214, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.38984751701355, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6968560218811035, + "num_tokens": 78388601.0, + "step": 3035 + }, + { + "epoch": 0.3334065451350758, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.220396041870117, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6970078349113464, + "num_tokens": 78415549.0, + "step": 3036 + }, + { + "epoch": 0.33351636283768943, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.1872615814208984, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7049839496612549, + "num_tokens": 78442685.0, + "step": 3037 + }, + { + "epoch": 0.3336261805403031, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.2038049697875977, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.6972586512565613, + "num_tokens": 78469472.0, + "step": 3038 + }, + { + "epoch": 0.3337359982429168, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 1.9924390316009521, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7114841341972351, + "num_tokens": 78501465.0, + "step": 3039 + }, + { + "epoch": 0.3338458159455304, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.2097280025482178, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7169581055641174, + "num_tokens": 78525782.0, + "step": 3040 + }, + { + "epoch": 0.33395563364814407, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.2431228160858154, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.7004686594009399, + "num_tokens": 78550247.0, + "step": 3041 + }, + { + "epoch": 0.33406545135075777, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.1479499340057373, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7059544920921326, + "num_tokens": 78576229.0, + "step": 3042 + }, + { + "epoch": 0.3341752690533714, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.1114096641540527, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6963151097297668, + "num_tokens": 78603794.0, + "step": 3043 + }, + { + "epoch": 0.33428508675598506, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.4869017601013184, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7038411498069763, + "num_tokens": 78624317.0, + "step": 3044 + }, + { + "epoch": 0.3343949044585987, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.238478183746338, + "learning_rate": 1e-06, + "loss": 1.0555, + "mean_token_accuracy": 0.6830487251281738, + "num_tokens": 78649693.0, + "step": 3045 + }, + { + "epoch": 0.3345047221612124, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.0294687747955322, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7270516157150269, + "num_tokens": 78676792.0, + "step": 3046 + }, + { + "epoch": 0.33461453986382605, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.416501522064209, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7142090797424316, + "num_tokens": 78698812.0, + "step": 3047 + }, + { + "epoch": 0.3347243575664397, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.431859016418457, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6995680928230286, + "num_tokens": 78723823.0, + "step": 3048 + }, + { + "epoch": 0.33483417526905335, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.1460797786712646, + "learning_rate": 1e-06, + "loss": 1.0482, + "mean_token_accuracy": 0.6893832683563232, + "num_tokens": 78751359.0, + "step": 3049 + }, + { + "epoch": 0.33494399297166705, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.139646530151367, + "learning_rate": 1e-06, + "loss": 1.102, + "mean_token_accuracy": 0.6697202324867249, + "num_tokens": 78778325.0, + "step": 3050 + }, + { + "epoch": 0.3350538106742807, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.541280508041382, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7212978601455688, + "num_tokens": 78800096.0, + "step": 3051 + }, + { + "epoch": 0.33516362837689434, + "ewc_loss": 2.47955322265625e-05, + "grad_norm": 2.7611896991729736, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7041716575622559, + "num_tokens": 78820036.0, + "step": 3052 + }, + { + "epoch": 0.33527344607950804, + "ewc_loss": 2.491474151611328e-05, + "grad_norm": 2.2424795627593994, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6964872479438782, + "num_tokens": 78845369.0, + "step": 3053 + }, + { + "epoch": 0.3353832637821217, + "ewc_loss": 2.491474151611328e-05, + "grad_norm": 2.0101633071899414, + "learning_rate": 1e-06, + "loss": 1.0958, + "mean_token_accuracy": 0.6763148307800293, + "num_tokens": 78878373.0, + "step": 3054 + }, + { + "epoch": 0.33549308148473533, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.3697478771209717, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6964241862297058, + "num_tokens": 78900627.0, + "step": 3055 + }, + { + "epoch": 0.335602899187349, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 1.9354369640350342, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6926854252815247, + "num_tokens": 78932358.0, + "step": 3056 + }, + { + "epoch": 0.3357127168899627, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.350656032562256, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.7022315859794617, + "num_tokens": 78954929.0, + "step": 3057 + }, + { + "epoch": 0.3358225345925763, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.0590128898620605, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7139362692832947, + "num_tokens": 78984236.0, + "step": 3058 + }, + { + "epoch": 0.33593235229519, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.1426117420196533, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6957976818084717, + "num_tokens": 79013737.0, + "step": 3059 + }, + { + "epoch": 0.3360421699978036, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 1.943822979927063, + "learning_rate": 1e-06, + "loss": 1.0716, + "mean_token_accuracy": 0.6796122193336487, + "num_tokens": 79047783.0, + "step": 3060 + }, + { + "epoch": 0.3361519877004173, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.483027458190918, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7000726461410522, + "num_tokens": 79069512.0, + "step": 3061 + }, + { + "epoch": 0.33626180540303097, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.0667669773101807, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7235002517700195, + "num_tokens": 79098817.0, + "step": 3062 + }, + { + "epoch": 0.3363716231056446, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.5633063316345215, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6873142719268799, + "num_tokens": 79119523.0, + "step": 3063 + }, + { + "epoch": 0.3364814408082583, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.1811861991882324, + "learning_rate": 1e-06, + "loss": 1.0934, + "mean_token_accuracy": 0.6760958433151245, + "num_tokens": 79148617.0, + "step": 3064 + }, + { + "epoch": 0.33659125851087196, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.197017192840576, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7036738395690918, + "num_tokens": 79174100.0, + "step": 3065 + }, + { + "epoch": 0.3367010762134856, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.1889491081237793, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.710391104221344, + "num_tokens": 79200804.0, + "step": 3066 + }, + { + "epoch": 0.33681089391609925, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.3780264854431152, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7042824625968933, + "num_tokens": 79223988.0, + "step": 3067 + }, + { + "epoch": 0.33692071161871295, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.3267626762390137, + "learning_rate": 1e-06, + "loss": 1.0828, + "mean_token_accuracy": 0.6791650056838989, + "num_tokens": 79248404.0, + "step": 3068 + }, + { + "epoch": 0.3370305293213266, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.0694308280944824, + "learning_rate": 1e-06, + "loss": 1.082, + "mean_token_accuracy": 0.6850248575210571, + "num_tokens": 79277072.0, + "step": 3069 + }, + { + "epoch": 0.33714034702394025, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.169257879257202, + "learning_rate": 1e-06, + "loss": 1.0992, + "mean_token_accuracy": 0.6732178926467896, + "num_tokens": 79304757.0, + "step": 3070 + }, + { + "epoch": 0.33725016472655395, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 1.892728567123413, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.6889171600341797, + "num_tokens": 79339350.0, + "step": 3071 + }, + { + "epoch": 0.3373599824291676, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.07539439201355, + "learning_rate": 1e-06, + "loss": 1.0508, + "mean_token_accuracy": 0.6828538775444031, + "num_tokens": 79369818.0, + "step": 3072 + }, + { + "epoch": 0.33746980013178124, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.088618278503418, + "learning_rate": 1e-06, + "loss": 1.0876, + "mean_token_accuracy": 0.6817839741706848, + "num_tokens": 79401881.0, + "step": 3073 + }, + { + "epoch": 0.3375796178343949, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.4977381229400635, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.7037661671638489, + "num_tokens": 79420457.0, + "step": 3074 + }, + { + "epoch": 0.3376894355370086, + "ewc_loss": 2.5033950805664062e-05, + "grad_norm": 2.1341147422790527, + "learning_rate": 1e-06, + "loss": 1.0424, + "mean_token_accuracy": 0.6997476816177368, + "num_tokens": 79450330.0, + "step": 3075 + }, + { + "epoch": 0.33779925323962223, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.204746961593628, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6980008482933044, + "num_tokens": 79477626.0, + "step": 3076 + }, + { + "epoch": 0.3379090709422359, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.3120555877685547, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6952823400497437, + "num_tokens": 79500703.0, + "step": 3077 + }, + { + "epoch": 0.3380188886448495, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.263610363006592, + "learning_rate": 1e-06, + "loss": 1.0366, + "mean_token_accuracy": 0.684660792350769, + "num_tokens": 79526708.0, + "step": 3078 + }, + { + "epoch": 0.3381287063474632, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 1.8957897424697876, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7024965882301331, + "num_tokens": 79561806.0, + "step": 3079 + }, + { + "epoch": 0.33823852405007687, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.2321503162384033, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7053850889205933, + "num_tokens": 79587764.0, + "step": 3080 + }, + { + "epoch": 0.3383483417526905, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.091042995452881, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7117843627929688, + "num_tokens": 79616144.0, + "step": 3081 + }, + { + "epoch": 0.3384581594553042, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.380293130874634, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7080643177032471, + "num_tokens": 79638796.0, + "step": 3082 + }, + { + "epoch": 0.33856797715791787, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.2675251960754395, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7198770046234131, + "num_tokens": 79662566.0, + "step": 3083 + }, + { + "epoch": 0.3386777948605315, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.222565174102783, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.7059492468833923, + "num_tokens": 79691734.0, + "step": 3084 + }, + { + "epoch": 0.33878761256314516, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.0756523609161377, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6996222734451294, + "num_tokens": 79720234.0, + "step": 3085 + }, + { + "epoch": 0.33889743026575886, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.2600088119506836, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6906839609146118, + "num_tokens": 79744274.0, + "step": 3086 + }, + { + "epoch": 0.3390072479683725, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.142796516418457, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.6969770193099976, + "num_tokens": 79771274.0, + "step": 3087 + }, + { + "epoch": 0.33911706567098615, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.281222105026245, + "learning_rate": 1e-06, + "loss": 1.0525, + "mean_token_accuracy": 0.6837912797927856, + "num_tokens": 79797808.0, + "step": 3088 + }, + { + "epoch": 0.3392268833735998, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.382080316543579, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7052545547485352, + "num_tokens": 79818490.0, + "step": 3089 + }, + { + "epoch": 0.3393367010762135, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.2682578563690186, + "learning_rate": 1e-06, + "loss": 1.0661, + "mean_token_accuracy": 0.6803209185600281, + "num_tokens": 79846912.0, + "step": 3090 + }, + { + "epoch": 0.33944651877882714, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.935107707977295, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.717620849609375, + "num_tokens": 79862416.0, + "step": 3091 + }, + { + "epoch": 0.3395563364814408, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.273144483566284, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6877239942550659, + "num_tokens": 79888186.0, + "step": 3092 + }, + { + "epoch": 0.3396661541840545, + "ewc_loss": 2.5153160095214844e-05, + "grad_norm": 2.204253673553467, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6924184560775757, + "num_tokens": 79914312.0, + "step": 3093 + }, + { + "epoch": 0.33977597188666814, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.348405361175537, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.708831250667572, + "num_tokens": 79938231.0, + "step": 3094 + }, + { + "epoch": 0.3398857895892818, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.044023275375366, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7154412865638733, + "num_tokens": 79965742.0, + "step": 3095 + }, + { + "epoch": 0.33999560729189543, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 1.961458444595337, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.7010621428489685, + "num_tokens": 80000024.0, + "step": 3096 + }, + { + "epoch": 0.34010542499450913, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 1.9731584787368774, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6992686986923218, + "num_tokens": 80032041.0, + "step": 3097 + }, + { + "epoch": 0.3402152426971228, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.2110867500305176, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6858426332473755, + "num_tokens": 80057633.0, + "step": 3098 + }, + { + "epoch": 0.3403250603997364, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.236598014831543, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6969075202941895, + "num_tokens": 80083175.0, + "step": 3099 + }, + { + "epoch": 0.3404348781023501, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.1079699993133545, + "learning_rate": 1e-06, + "loss": 1.0866, + "mean_token_accuracy": 0.6807583570480347, + "num_tokens": 80112222.0, + "step": 3100 + }, + { + "epoch": 0.34054469580496377, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.7365825176239014, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7056310176849365, + "num_tokens": 80129502.0, + "step": 3101 + }, + { + "epoch": 0.3406545135075774, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.429633617401123, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7139758467674255, + "num_tokens": 80152719.0, + "step": 3102 + }, + { + "epoch": 0.34076433121019106, + "ewc_loss": 2.5391578674316406e-05, + "grad_norm": 2.1707358360290527, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7136632204055786, + "num_tokens": 80177076.0, + "step": 3103 + }, + { + "epoch": 0.34087414891280476, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.1391022205352783, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6969507932662964, + "num_tokens": 80204022.0, + "step": 3104 + }, + { + "epoch": 0.3409839666154184, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.201878070831299, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7119449973106384, + "num_tokens": 80229727.0, + "step": 3105 + }, + { + "epoch": 0.34109378431803206, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.305673599243164, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7203216552734375, + "num_tokens": 80251886.0, + "step": 3106 + }, + { + "epoch": 0.3412036020206457, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 1.9952605962753296, + "learning_rate": 1e-06, + "loss": 1.0604, + "mean_token_accuracy": 0.6817898154258728, + "num_tokens": 80283712.0, + "step": 3107 + }, + { + "epoch": 0.3413134197232594, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.194850444793701, + "learning_rate": 1e-06, + "loss": 1.0825, + "mean_token_accuracy": 0.6855268478393555, + "num_tokens": 80311758.0, + "step": 3108 + }, + { + "epoch": 0.34142323742587305, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.4167823791503906, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6962265968322754, + "num_tokens": 80333423.0, + "step": 3109 + }, + { + "epoch": 0.3415330551284867, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.3409409523010254, + "learning_rate": 1e-06, + "loss": 1.0494, + "mean_token_accuracy": 0.6822659969329834, + "num_tokens": 80358027.0, + "step": 3110 + }, + { + "epoch": 0.3416428728311004, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.144763708114624, + "learning_rate": 1e-06, + "loss": 1.0676, + "mean_token_accuracy": 0.6978682279586792, + "num_tokens": 80385582.0, + "step": 3111 + }, + { + "epoch": 0.34175269053371404, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.1889922618865967, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7166997194290161, + "num_tokens": 80415533.0, + "step": 3112 + }, + { + "epoch": 0.3418625082363277, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.285062551498413, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7164758443832397, + "num_tokens": 80439223.0, + "step": 3113 + }, + { + "epoch": 0.34197232593894134, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.4774460792541504, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.703404426574707, + "num_tokens": 80462023.0, + "step": 3114 + }, + { + "epoch": 0.34208214364155504, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.4832096099853516, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6962758898735046, + "num_tokens": 80484919.0, + "step": 3115 + }, + { + "epoch": 0.3421919613441687, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.214057207107544, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7245858907699585, + "num_tokens": 80508170.0, + "step": 3116 + }, + { + "epoch": 0.34230177904678233, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.156092643737793, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.6871502995491028, + "num_tokens": 80539222.0, + "step": 3117 + }, + { + "epoch": 0.34241159674939603, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.2945022583007812, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7109272480010986, + "num_tokens": 80562836.0, + "step": 3118 + }, + { + "epoch": 0.3425214144520097, + "ewc_loss": 2.5510787963867188e-05, + "grad_norm": 2.1827030181884766, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7008072137832642, + "num_tokens": 80588438.0, + "step": 3119 + }, + { + "epoch": 0.3426312321546233, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.2595653533935547, + "learning_rate": 1e-06, + "loss": 1.0827, + "mean_token_accuracy": 0.679743230342865, + "num_tokens": 80615939.0, + "step": 3120 + }, + { + "epoch": 0.34274104985723697, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.1937146186828613, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6904981136322021, + "num_tokens": 80642335.0, + "step": 3121 + }, + { + "epoch": 0.34285086755985067, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.3440144062042236, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6939209699630737, + "num_tokens": 80665781.0, + "step": 3122 + }, + { + "epoch": 0.3429606852624643, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.1248586177825928, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6976451873779297, + "num_tokens": 80696047.0, + "step": 3123 + }, + { + "epoch": 0.34307050296507796, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.5082356929779053, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7064604759216309, + "num_tokens": 80719122.0, + "step": 3124 + }, + { + "epoch": 0.3431803206676916, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.0798277854919434, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6986954212188721, + "num_tokens": 80747066.0, + "step": 3125 + }, + { + "epoch": 0.3432901383703053, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.1143698692321777, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.7030056715011597, + "num_tokens": 80775529.0, + "step": 3126 + }, + { + "epoch": 0.34339995607291895, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.0849037170410156, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6890919208526611, + "num_tokens": 80806422.0, + "step": 3127 + }, + { + "epoch": 0.3435097737755326, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.24399995803833, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7008246183395386, + "num_tokens": 80829206.0, + "step": 3128 + }, + { + "epoch": 0.3436195914781463, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.022850275039673, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6916187405586243, + "num_tokens": 80859251.0, + "step": 3129 + }, + { + "epoch": 0.34372940918075995, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.4963669776916504, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7069585919380188, + "num_tokens": 80879585.0, + "step": 3130 + }, + { + "epoch": 0.3438392268833736, + "ewc_loss": 2.562999725341797e-05, + "grad_norm": 2.7620208263397217, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7158337235450745, + "num_tokens": 80896459.0, + "step": 3131 + }, + { + "epoch": 0.34394904458598724, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.5239851474761963, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7087125778198242, + "num_tokens": 80916349.0, + "step": 3132 + }, + { + "epoch": 0.34405886228860094, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.426241874694824, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7134482860565186, + "num_tokens": 80940274.0, + "step": 3133 + }, + { + "epoch": 0.3441686799912146, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.155103921890259, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.69638991355896, + "num_tokens": 80967458.0, + "step": 3134 + }, + { + "epoch": 0.34427849769382823, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.5953176021575928, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7094255685806274, + "num_tokens": 80987311.0, + "step": 3135 + }, + { + "epoch": 0.3443883153964419, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.5033912658691406, + "learning_rate": 1e-06, + "loss": 1.1255, + "mean_token_accuracy": 0.6655994653701782, + "num_tokens": 81009634.0, + "step": 3136 + }, + { + "epoch": 0.3444981330990556, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.54533052444458, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7036327123641968, + "num_tokens": 81029324.0, + "step": 3137 + }, + { + "epoch": 0.3446079508016692, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.050396680831909, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6869379281997681, + "num_tokens": 81060117.0, + "step": 3138 + }, + { + "epoch": 0.3447177685042829, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.4001054763793945, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7054725885391235, + "num_tokens": 81084163.0, + "step": 3139 + }, + { + "epoch": 0.3448275862068966, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.126741647720337, + "learning_rate": 1e-06, + "loss": 1.0667, + "mean_token_accuracy": 0.6838445663452148, + "num_tokens": 81115321.0, + "step": 3140 + }, + { + "epoch": 0.3449374039095102, + "ewc_loss": 2.574920654296875e-05, + "grad_norm": 2.5937411785125732, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7069599628448486, + "num_tokens": 81136307.0, + "step": 3141 + }, + { + "epoch": 0.34504722161212387, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.1596579551696777, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6875742077827454, + "num_tokens": 81163479.0, + "step": 3142 + }, + { + "epoch": 0.3451570393147375, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.3905975818634033, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.686234712600708, + "num_tokens": 81188105.0, + "step": 3143 + }, + { + "epoch": 0.3452668570173512, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.0901641845703125, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6991993188858032, + "num_tokens": 81217529.0, + "step": 3144 + }, + { + "epoch": 0.34537667471996486, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.0708956718444824, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7075526118278503, + "num_tokens": 81245442.0, + "step": 3145 + }, + { + "epoch": 0.3454864924225785, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.1168935298919678, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7116808295249939, + "num_tokens": 81273374.0, + "step": 3146 + }, + { + "epoch": 0.3455963101251922, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.22536301612854, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.705102264881134, + "num_tokens": 81299150.0, + "step": 3147 + }, + { + "epoch": 0.34570612782780585, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.3394291400909424, + "learning_rate": 1e-06, + "loss": 1.0632, + "mean_token_accuracy": 0.683979868888855, + "num_tokens": 81323114.0, + "step": 3148 + }, + { + "epoch": 0.3458159455304195, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.512019634246826, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7212867736816406, + "num_tokens": 81343642.0, + "step": 3149 + }, + { + "epoch": 0.34592576323303315, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.1873550415039062, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7161272168159485, + "num_tokens": 81371186.0, + "step": 3150 + }, + { + "epoch": 0.34603558093564685, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.235267162322998, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7010118961334229, + "num_tokens": 81398073.0, + "step": 3151 + }, + { + "epoch": 0.3461453986382605, + "ewc_loss": 2.586841583251953e-05, + "grad_norm": 2.089606523513794, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7063111066818237, + "num_tokens": 81426255.0, + "step": 3152 + }, + { + "epoch": 0.34625521634087414, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.21046781539917, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6941309571266174, + "num_tokens": 81452329.0, + "step": 3153 + }, + { + "epoch": 0.3463650340434878, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.2309792041778564, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7261207699775696, + "num_tokens": 81475750.0, + "step": 3154 + }, + { + "epoch": 0.3464748517461015, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.5413379669189453, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6987191438674927, + "num_tokens": 81499098.0, + "step": 3155 + }, + { + "epoch": 0.34658466944871513, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.1864519119262695, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.6852825880050659, + "num_tokens": 81525567.0, + "step": 3156 + }, + { + "epoch": 0.3466944871513288, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.290466547012329, + "learning_rate": 1e-06, + "loss": 1.079, + "mean_token_accuracy": 0.6732644438743591, + "num_tokens": 81552757.0, + "step": 3157 + }, + { + "epoch": 0.3468043048539425, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.236752986907959, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6893554329872131, + "num_tokens": 81579060.0, + "step": 3158 + }, + { + "epoch": 0.3469141225565561, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.201228141784668, + "learning_rate": 1e-06, + "loss": 1.0569, + "mean_token_accuracy": 0.6839008927345276, + "num_tokens": 81608542.0, + "step": 3159 + }, + { + "epoch": 0.34702394025916977, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 1.9734370708465576, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.703161895275116, + "num_tokens": 81639879.0, + "step": 3160 + }, + { + "epoch": 0.3471337579617834, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.188310146331787, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7146173715591431, + "num_tokens": 81664867.0, + "step": 3161 + }, + { + "epoch": 0.3472435756643971, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 2.36213755607605, + "learning_rate": 1e-06, + "loss": 1.0785, + "mean_token_accuracy": 0.6823646426200867, + "num_tokens": 81691172.0, + "step": 3162 + }, + { + "epoch": 0.34735339336701077, + "ewc_loss": 2.5987625122070312e-05, + "grad_norm": 1.9734902381896973, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7071007490158081, + "num_tokens": 81722600.0, + "step": 3163 + }, + { + "epoch": 0.3474632110696244, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.9628241062164307, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7236347198486328, + "num_tokens": 81737797.0, + "step": 3164 + }, + { + "epoch": 0.34757302877223806, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.5337390899658203, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7127393484115601, + "num_tokens": 81757397.0, + "step": 3165 + }, + { + "epoch": 0.34768284647485176, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.3738813400268555, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7189298868179321, + "num_tokens": 81778433.0, + "step": 3166 + }, + { + "epoch": 0.3477926641774654, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.06274151802063, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6934125423431396, + "num_tokens": 81808813.0, + "step": 3167 + }, + { + "epoch": 0.34790248188007905, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.2675130367279053, + "learning_rate": 1e-06, + "loss": 1.0386, + "mean_token_accuracy": 0.6826492547988892, + "num_tokens": 81835433.0, + "step": 3168 + }, + { + "epoch": 0.34801229958269275, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.5508861541748047, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7082070708274841, + "num_tokens": 81855762.0, + "step": 3169 + }, + { + "epoch": 0.3481221172853064, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.3480656147003174, + "learning_rate": 1e-06, + "loss": 1.0879, + "mean_token_accuracy": 0.6743285655975342, + "num_tokens": 81880430.0, + "step": 3170 + }, + { + "epoch": 0.34823193498792004, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.3617336750030518, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7148662805557251, + "num_tokens": 81903500.0, + "step": 3171 + }, + { + "epoch": 0.3483417526905337, + "ewc_loss": 2.6226043701171875e-05, + "grad_norm": 2.0954227447509766, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7131146192550659, + "num_tokens": 81935057.0, + "step": 3172 + }, + { + "epoch": 0.3484515703931474, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.4184470176696777, + "learning_rate": 1e-06, + "loss": 1.0704, + "mean_token_accuracy": 0.6761521100997925, + "num_tokens": 81959076.0, + "step": 3173 + }, + { + "epoch": 0.34856138809576104, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.495467185974121, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.7084566354751587, + "num_tokens": 81981974.0, + "step": 3174 + }, + { + "epoch": 0.3486712057983747, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 1.96016263961792, + "learning_rate": 1e-06, + "loss": 1.0497, + "mean_token_accuracy": 0.6867801547050476, + "num_tokens": 82015685.0, + "step": 3175 + }, + { + "epoch": 0.3487810235009884, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.482330083847046, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.709037184715271, + "num_tokens": 82037034.0, + "step": 3176 + }, + { + "epoch": 0.34889084120360203, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.366787910461426, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7085769176483154, + "num_tokens": 82060036.0, + "step": 3177 + }, + { + "epoch": 0.3490006589062157, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.329956531524658, + "learning_rate": 1e-06, + "loss": 1.0365, + "mean_token_accuracy": 0.6980471611022949, + "num_tokens": 82086924.0, + "step": 3178 + }, + { + "epoch": 0.3491104766088293, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.4259798526763916, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.7073478102684021, + "num_tokens": 82110692.0, + "step": 3179 + }, + { + "epoch": 0.349220294311443, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.5088491439819336, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.6907400488853455, + "num_tokens": 82133124.0, + "step": 3180 + }, + { + "epoch": 0.34933011201405667, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.1157994270324707, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7350108623504639, + "num_tokens": 82160041.0, + "step": 3181 + }, + { + "epoch": 0.3494399297166703, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.4101924896240234, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7057758569717407, + "num_tokens": 82183128.0, + "step": 3182 + }, + { + "epoch": 0.34954974741928396, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.361386775970459, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7036746740341187, + "num_tokens": 82206128.0, + "step": 3183 + }, + { + "epoch": 0.34965956512189766, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.4853906631469727, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7056427597999573, + "num_tokens": 82227187.0, + "step": 3184 + }, + { + "epoch": 0.3497693828245113, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.297239065170288, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.710368275642395, + "num_tokens": 82250941.0, + "step": 3185 + }, + { + "epoch": 0.34987920052712496, + "ewc_loss": 2.6464462280273438e-05, + "grad_norm": 2.1470372676849365, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6926568746566772, + "num_tokens": 82279220.0, + "step": 3186 + }, + { + "epoch": 0.34998901822973866, + "ewc_loss": 2.6702880859375e-05, + "grad_norm": 2.1371099948883057, + "learning_rate": 1e-06, + "loss": 1.0481, + "mean_token_accuracy": 0.6878198385238647, + "num_tokens": 82308794.0, + "step": 3187 + }, + { + "epoch": 0.3500988359323523, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.1675844192504883, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.7051898837089539, + "num_tokens": 82333167.0, + "step": 3188 + }, + { + "epoch": 0.35020865363496595, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.389317274093628, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7158293724060059, + "num_tokens": 82356761.0, + "step": 3189 + }, + { + "epoch": 0.3503184713375796, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.3914692401885986, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7054588794708252, + "num_tokens": 82377746.0, + "step": 3190 + }, + { + "epoch": 0.3504282890401933, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.2285120487213135, + "learning_rate": 1e-06, + "loss": 1.1197, + "mean_token_accuracy": 0.6705701351165771, + "num_tokens": 82404722.0, + "step": 3191 + }, + { + "epoch": 0.35053810674280694, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.2176265716552734, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7055091261863708, + "num_tokens": 82431008.0, + "step": 3192 + }, + { + "epoch": 0.3506479244454206, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.2234504222869873, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6845963597297668, + "num_tokens": 82459447.0, + "step": 3193 + }, + { + "epoch": 0.3507577421480343, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.165501832962036, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6904330849647522, + "num_tokens": 82485581.0, + "step": 3194 + }, + { + "epoch": 0.35086755985064794, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.081512928009033, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7048439383506775, + "num_tokens": 82512694.0, + "step": 3195 + }, + { + "epoch": 0.3509773775532616, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.225003719329834, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7169739007949829, + "num_tokens": 82537247.0, + "step": 3196 + }, + { + "epoch": 0.35108719525587523, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.4440126419067383, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7173036932945251, + "num_tokens": 82558570.0, + "step": 3197 + }, + { + "epoch": 0.35119701295848893, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.1121082305908203, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6941624283790588, + "num_tokens": 82587299.0, + "step": 3198 + }, + { + "epoch": 0.3513068306611026, + "ewc_loss": 2.682209014892578e-05, + "grad_norm": 2.1506474018096924, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.6926082372665405, + "num_tokens": 82619644.0, + "step": 3199 + }, + { + "epoch": 0.3514166483637162, + "ewc_loss": 2.6941299438476562e-05, + "grad_norm": 2.406837224960327, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7073454260826111, + "num_tokens": 82644335.0, + "step": 3200 + }, + { + "epoch": 0.35152646606632987, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.0210819244384766, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.6931488513946533, + "num_tokens": 82674506.0, + "step": 3201 + }, + { + "epoch": 0.35163628376894357, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.159615993499756, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.6886987686157227, + "num_tokens": 82702472.0, + "step": 3202 + }, + { + "epoch": 0.3517461014715572, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.1008036136627197, + "learning_rate": 1e-06, + "loss": 1.0477, + "mean_token_accuracy": 0.6814973950386047, + "num_tokens": 82730656.0, + "step": 3203 + }, + { + "epoch": 0.35185591917417086, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.094125270843506, + "learning_rate": 1e-06, + "loss": 1.0917, + "mean_token_accuracy": 0.6746824979782104, + "num_tokens": 82761816.0, + "step": 3204 + }, + { + "epoch": 0.35196573687678456, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.25295352935791, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7133010625839233, + "num_tokens": 82785818.0, + "step": 3205 + }, + { + "epoch": 0.3520755545793982, + "ewc_loss": 2.7060508728027344e-05, + "grad_norm": 2.149547815322876, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7029067277908325, + "num_tokens": 82813666.0, + "step": 3206 + }, + { + "epoch": 0.35218537228201185, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.25137996673584, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.7062072157859802, + "num_tokens": 82840296.0, + "step": 3207 + }, + { + "epoch": 0.3522951899846255, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.848005771636963, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7045032978057861, + "num_tokens": 82857720.0, + "step": 3208 + }, + { + "epoch": 0.3524050076872392, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.3839588165283203, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6837872862815857, + "num_tokens": 82882989.0, + "step": 3209 + }, + { + "epoch": 0.35251482538985285, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.174180746078491, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7261989116668701, + "num_tokens": 82908373.0, + "step": 3210 + }, + { + "epoch": 0.3526246430924665, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 1.9311535358428955, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.704243540763855, + "num_tokens": 82939675.0, + "step": 3211 + }, + { + "epoch": 0.35273446079508014, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.415754795074463, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.7067558169364929, + "num_tokens": 82964508.0, + "step": 3212 + }, + { + "epoch": 0.35284427849769384, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.344700336456299, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7227625846862793, + "num_tokens": 82986542.0, + "step": 3213 + }, + { + "epoch": 0.3529540962003075, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.2874643802642822, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.7030328512191772, + "num_tokens": 83010692.0, + "step": 3214 + }, + { + "epoch": 0.35306391390292113, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.17740535736084, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7184364795684814, + "num_tokens": 83037047.0, + "step": 3215 + }, + { + "epoch": 0.35317373160553484, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.1345324516296387, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7001187801361084, + "num_tokens": 83063122.0, + "step": 3216 + }, + { + "epoch": 0.3532835493081485, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.1284523010253906, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6971442699432373, + "num_tokens": 83091684.0, + "step": 3217 + }, + { + "epoch": 0.3533933670107621, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.1004934310913086, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6965317726135254, + "num_tokens": 83119244.0, + "step": 3218 + }, + { + "epoch": 0.3535031847133758, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.432177782058716, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.710525631904602, + "num_tokens": 83141800.0, + "step": 3219 + }, + { + "epoch": 0.3536130024159895, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.1403415203094482, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.700439989566803, + "num_tokens": 83170081.0, + "step": 3220 + }, + { + "epoch": 0.3537228201186031, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.517366886138916, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7016741037368774, + "num_tokens": 83190689.0, + "step": 3221 + }, + { + "epoch": 0.35383263782121677, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.071794033050537, + "learning_rate": 1e-06, + "loss": 1.0765, + "mean_token_accuracy": 0.6770301461219788, + "num_tokens": 83223907.0, + "step": 3222 + }, + { + "epoch": 0.35394245552383047, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.493809223175049, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6988686323165894, + "num_tokens": 83245462.0, + "step": 3223 + }, + { + "epoch": 0.3540522732264441, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.5036191940307617, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7059216499328613, + "num_tokens": 83265293.0, + "step": 3224 + }, + { + "epoch": 0.35416209092905776, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.2392094135284424, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7131361961364746, + "num_tokens": 83291068.0, + "step": 3225 + }, + { + "epoch": 0.3542719086316714, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 2.2454402446746826, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.692191481590271, + "num_tokens": 83318510.0, + "step": 3226 + }, + { + "epoch": 0.3543817263342851, + "ewc_loss": 2.7179718017578125e-05, + "grad_norm": 1.8650810718536377, + "learning_rate": 1e-06, + "loss": 1.0539, + "mean_token_accuracy": 0.6797838807106018, + "num_tokens": 83353576.0, + "step": 3227 + }, + { + "epoch": 0.35449154403689875, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.383755683898926, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6946083307266235, + "num_tokens": 83376074.0, + "step": 3228 + }, + { + "epoch": 0.3546013617395124, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.122947931289673, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.700049638748169, + "num_tokens": 83401544.0, + "step": 3229 + }, + { + "epoch": 0.35471117944212605, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 1.9458448886871338, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6969887018203735, + "num_tokens": 83433777.0, + "step": 3230 + }, + { + "epoch": 0.35482099714473975, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.1851065158843994, + "learning_rate": 1e-06, + "loss": 1.0466, + "mean_token_accuracy": 0.6921762824058533, + "num_tokens": 83459923.0, + "step": 3231 + }, + { + "epoch": 0.3549308148473534, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.2144687175750732, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.677237868309021, + "num_tokens": 83487301.0, + "step": 3232 + }, + { + "epoch": 0.35504063254996704, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.5326128005981445, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7115740180015564, + "num_tokens": 83507683.0, + "step": 3233 + }, + { + "epoch": 0.35515045025258074, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.283623695373535, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.6840588450431824, + "num_tokens": 83533092.0, + "step": 3234 + }, + { + "epoch": 0.3552602679551944, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.1626713275909424, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6865111589431763, + "num_tokens": 83562221.0, + "step": 3235 + }, + { + "epoch": 0.35537008565780803, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.233015298843384, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7055498361587524, + "num_tokens": 83587927.0, + "step": 3236 + }, + { + "epoch": 0.3554799033604217, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 1.9553838968276978, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6958118081092834, + "num_tokens": 83619553.0, + "step": 3237 + }, + { + "epoch": 0.3555897210630354, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 1.9193068742752075, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7239327430725098, + "num_tokens": 83652859.0, + "step": 3238 + }, + { + "epoch": 0.355699538765649, + "ewc_loss": 2.7298927307128906e-05, + "grad_norm": 2.137317657470703, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6935850381851196, + "num_tokens": 83679314.0, + "step": 3239 + }, + { + "epoch": 0.35580935646826267, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.2495651245117188, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6963939666748047, + "num_tokens": 83704588.0, + "step": 3240 + }, + { + "epoch": 0.3559191741708763, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.0308990478515625, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7027209997177124, + "num_tokens": 83732073.0, + "step": 3241 + }, + { + "epoch": 0.35602899187349, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.080143690109253, + "learning_rate": 1e-06, + "loss": 1.0606, + "mean_token_accuracy": 0.6822125911712646, + "num_tokens": 83761233.0, + "step": 3242 + }, + { + "epoch": 0.35613880957610367, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.235017776489258, + "learning_rate": 1e-06, + "loss": 1.0669, + "mean_token_accuracy": 0.6748185753822327, + "num_tokens": 83789309.0, + "step": 3243 + }, + { + "epoch": 0.3562486272787173, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.0850558280944824, + "learning_rate": 1e-06, + "loss": 1.0785, + "mean_token_accuracy": 0.6855063438415527, + "num_tokens": 83822253.0, + "step": 3244 + }, + { + "epoch": 0.356358444981331, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.049639940261841, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7068080902099609, + "num_tokens": 83850718.0, + "step": 3245 + }, + { + "epoch": 0.35646826268394466, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.268838405609131, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.697667121887207, + "num_tokens": 83874658.0, + "step": 3246 + }, + { + "epoch": 0.3565780803865583, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.267157554626465, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7225481271743774, + "num_tokens": 83899291.0, + "step": 3247 + }, + { + "epoch": 0.35668789808917195, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.245087146759033, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7205830216407776, + "num_tokens": 83922590.0, + "step": 3248 + }, + { + "epoch": 0.35679771579178565, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.0796926021575928, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.7001757621765137, + "num_tokens": 83950687.0, + "step": 3249 + }, + { + "epoch": 0.3569075334943993, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.1575846672058105, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6965199708938599, + "num_tokens": 83977564.0, + "step": 3250 + }, + { + "epoch": 0.35701735119701294, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.3621795177459717, + "learning_rate": 1e-06, + "loss": 1.0551, + "mean_token_accuracy": 0.6830569505691528, + "num_tokens": 84003184.0, + "step": 3251 + }, + { + "epoch": 0.35712716889962665, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.043848752975464, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.7012459635734558, + "num_tokens": 84031308.0, + "step": 3252 + }, + { + "epoch": 0.3572369866022403, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.219670295715332, + "learning_rate": 1e-06, + "loss": 1.0641, + "mean_token_accuracy": 0.6767597198486328, + "num_tokens": 84057005.0, + "step": 3253 + }, + { + "epoch": 0.35734680430485394, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.225670576095581, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6877114176750183, + "num_tokens": 84084931.0, + "step": 3254 + }, + { + "epoch": 0.3574566220074676, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.3815970420837402, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6822527050971985, + "num_tokens": 84108380.0, + "step": 3255 + }, + { + "epoch": 0.3575664397100813, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.471945285797119, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7221952676773071, + "num_tokens": 84130087.0, + "step": 3256 + }, + { + "epoch": 0.35767625741269493, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.1902079582214355, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7060588598251343, + "num_tokens": 84154892.0, + "step": 3257 + }, + { + "epoch": 0.3577860751153086, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.4418256282806396, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.709575891494751, + "num_tokens": 84176496.0, + "step": 3258 + }, + { + "epoch": 0.3578958928179222, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.4094221591949463, + "learning_rate": 1e-06, + "loss": 1.0429, + "mean_token_accuracy": 0.6878582239151001, + "num_tokens": 84200797.0, + "step": 3259 + }, + { + "epoch": 0.3580057105205359, + "ewc_loss": 2.7418136596679688e-05, + "grad_norm": 2.2408065795898438, + "learning_rate": 1e-06, + "loss": 1.0894, + "mean_token_accuracy": 0.6879614591598511, + "num_tokens": 84226155.0, + "step": 3260 + }, + { + "epoch": 0.35811552822314957, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.4958579540252686, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7179539203643799, + "num_tokens": 84249673.0, + "step": 3261 + }, + { + "epoch": 0.3582253459257632, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.337006092071533, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.7006034851074219, + "num_tokens": 84273724.0, + "step": 3262 + }, + { + "epoch": 0.3583351636283769, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.344575881958008, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6930713653564453, + "num_tokens": 84299628.0, + "step": 3263 + }, + { + "epoch": 0.35844498133099056, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.256544828414917, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7127610445022583, + "num_tokens": 84327527.0, + "step": 3264 + }, + { + "epoch": 0.3585547990336042, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.276726245880127, + "learning_rate": 1e-06, + "loss": 1.0517, + "mean_token_accuracy": 0.6898359060287476, + "num_tokens": 84354889.0, + "step": 3265 + }, + { + "epoch": 0.35866461673621786, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.264211416244507, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6979488730430603, + "num_tokens": 84380522.0, + "step": 3266 + }, + { + "epoch": 0.35877443443883156, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.2527694702148438, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7126048803329468, + "num_tokens": 84405266.0, + "step": 3267 + }, + { + "epoch": 0.3588842521414452, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.291940927505493, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.718399703502655, + "num_tokens": 84428101.0, + "step": 3268 + }, + { + "epoch": 0.35899406984405885, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.255093812942505, + "learning_rate": 1e-06, + "loss": 1.0816, + "mean_token_accuracy": 0.6781460046768188, + "num_tokens": 84454829.0, + "step": 3269 + }, + { + "epoch": 0.35910388754667255, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 2.3709919452667236, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7260659337043762, + "num_tokens": 84475796.0, + "step": 3270 + }, + { + "epoch": 0.3592137052492862, + "ewc_loss": 2.765655517578125e-05, + "grad_norm": 1.9908318519592285, + "learning_rate": 1e-06, + "loss": 1.0755, + "mean_token_accuracy": 0.6754152178764343, + "num_tokens": 84506712.0, + "step": 3271 + }, + { + "epoch": 0.35932352295189984, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.5367987155914307, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7068836092948914, + "num_tokens": 84527613.0, + "step": 3272 + }, + { + "epoch": 0.3594333406545135, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.2630679607391357, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.696119487285614, + "num_tokens": 84552379.0, + "step": 3273 + }, + { + "epoch": 0.3595431583571272, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 1.9771698713302612, + "learning_rate": 1e-06, + "loss": 1.0614, + "mean_token_accuracy": 0.6814835667610168, + "num_tokens": 84583957.0, + "step": 3274 + }, + { + "epoch": 0.35965297605974084, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.101487159729004, + "learning_rate": 1e-06, + "loss": 1.0817, + "mean_token_accuracy": 0.6735143661499023, + "num_tokens": 84616061.0, + "step": 3275 + }, + { + "epoch": 0.3597627937623545, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.3694868087768555, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6947188377380371, + "num_tokens": 84639966.0, + "step": 3276 + }, + { + "epoch": 0.35987261146496813, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 1.999293327331543, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.7013555765151978, + "num_tokens": 84670805.0, + "step": 3277 + }, + { + "epoch": 0.35998242916758183, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.3651821613311768, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6964870691299438, + "num_tokens": 84695492.0, + "step": 3278 + }, + { + "epoch": 0.3600922468701955, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.211771249771118, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7252532243728638, + "num_tokens": 84721170.0, + "step": 3279 + }, + { + "epoch": 0.3602020645728091, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.222599506378174, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6933368444442749, + "num_tokens": 84746589.0, + "step": 3280 + }, + { + "epoch": 0.3603118822754228, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.2749838829040527, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6965934038162231, + "num_tokens": 84770710.0, + "step": 3281 + }, + { + "epoch": 0.36042169997803647, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 1.908002495765686, + "learning_rate": 1e-06, + "loss": 1.127, + "mean_token_accuracy": 0.6650789380073547, + "num_tokens": 84806101.0, + "step": 3282 + }, + { + "epoch": 0.3605315176806501, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.0689685344696045, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7108337879180908, + "num_tokens": 84835710.0, + "step": 3283 + }, + { + "epoch": 0.36064133538326376, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.257786512374878, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6938601136207581, + "num_tokens": 84863843.0, + "step": 3284 + }, + { + "epoch": 0.36075115308587746, + "ewc_loss": 2.777576446533203e-05, + "grad_norm": 2.388723373413086, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.6865659356117249, + "num_tokens": 84886682.0, + "step": 3285 + }, + { + "epoch": 0.3608609707884911, + "ewc_loss": 2.8014183044433594e-05, + "grad_norm": 2.5684986114501953, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6958260536193848, + "num_tokens": 84907655.0, + "step": 3286 + }, + { + "epoch": 0.36097078849110475, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.108707904815674, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6937495470046997, + "num_tokens": 84936194.0, + "step": 3287 + }, + { + "epoch": 0.3610806061937184, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.0414021015167236, + "learning_rate": 1e-06, + "loss": 1.0495, + "mean_token_accuracy": 0.6856281757354736, + "num_tokens": 84964512.0, + "step": 3288 + }, + { + "epoch": 0.3611904238963321, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.1967577934265137, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6989706754684448, + "num_tokens": 84991502.0, + "step": 3289 + }, + { + "epoch": 0.36130024159894575, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.44765567779541, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.6885031461715698, + "num_tokens": 85013129.0, + "step": 3290 + }, + { + "epoch": 0.3614100593015594, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.3227717876434326, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7166716456413269, + "num_tokens": 85036620.0, + "step": 3291 + }, + { + "epoch": 0.3615198770041731, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.3748199939727783, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6945933103561401, + "num_tokens": 85060171.0, + "step": 3292 + }, + { + "epoch": 0.36162969470678674, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.168541193008423, + "learning_rate": 1e-06, + "loss": 1.0425, + "mean_token_accuracy": 0.685409426689148, + "num_tokens": 85086553.0, + "step": 3293 + }, + { + "epoch": 0.3617395124094004, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.031853675842285, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6936112642288208, + "num_tokens": 85118680.0, + "step": 3294 + }, + { + "epoch": 0.36184933011201403, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.3233203887939453, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7139272689819336, + "num_tokens": 85142298.0, + "step": 3295 + }, + { + "epoch": 0.36195914781462774, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.0741353034973145, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7107623815536499, + "num_tokens": 85170484.0, + "step": 3296 + }, + { + "epoch": 0.3620689655172414, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.186800956726074, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7196822166442871, + "num_tokens": 85196017.0, + "step": 3297 + }, + { + "epoch": 0.362178783219855, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.408802032470703, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7262439727783203, + "num_tokens": 85216845.0, + "step": 3298 + }, + { + "epoch": 0.36228860092246873, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.124211072921753, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7133594751358032, + "num_tokens": 85243824.0, + "step": 3299 + }, + { + "epoch": 0.3623984186250824, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.0667710304260254, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.7054397463798523, + "num_tokens": 85274198.0, + "step": 3300 + }, + { + "epoch": 0.362508236327696, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.1539738178253174, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6989799737930298, + "num_tokens": 85303239.0, + "step": 3301 + }, + { + "epoch": 0.36261805403030967, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.04081392288208, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7077880501747131, + "num_tokens": 85330757.0, + "step": 3302 + }, + { + "epoch": 0.36272787173292337, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.2950093746185303, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7326771020889282, + "num_tokens": 85354065.0, + "step": 3303 + }, + { + "epoch": 0.362837689435537, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.27097749710083, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7395802736282349, + "num_tokens": 85377501.0, + "step": 3304 + }, + { + "epoch": 0.36294750713815066, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.1482386589050293, + "learning_rate": 1e-06, + "loss": 1.0237, + "mean_token_accuracy": 0.6943787336349487, + "num_tokens": 85405009.0, + "step": 3305 + }, + { + "epoch": 0.3630573248407643, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.4777629375457764, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7241577506065369, + "num_tokens": 85425905.0, + "step": 3306 + }, + { + "epoch": 0.363167142543378, + "ewc_loss": 2.8133392333984375e-05, + "grad_norm": 2.2468175888061523, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.7076447010040283, + "num_tokens": 85452029.0, + "step": 3307 + }, + { + "epoch": 0.36327696024599165, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.3468353748321533, + "learning_rate": 1e-06, + "loss": 1.0041, + "mean_token_accuracy": 0.7039567828178406, + "num_tokens": 85478947.0, + "step": 3308 + }, + { + "epoch": 0.3633867779486053, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.112086057662964, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.699349045753479, + "num_tokens": 85510887.0, + "step": 3309 + }, + { + "epoch": 0.363496595651219, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.187661647796631, + "learning_rate": 1e-06, + "loss": 1.0474, + "mean_token_accuracy": 0.6932066679000854, + "num_tokens": 85539105.0, + "step": 3310 + }, + { + "epoch": 0.36360641335383265, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.4052224159240723, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7173624038696289, + "num_tokens": 85561827.0, + "step": 3311 + }, + { + "epoch": 0.3637162310564463, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.001863956451416, + "learning_rate": 1e-06, + "loss": 1.0928, + "mean_token_accuracy": 0.6721547842025757, + "num_tokens": 85594193.0, + "step": 3312 + }, + { + "epoch": 0.36382604875905994, + "ewc_loss": 2.86102294921875e-05, + "grad_norm": 2.2213101387023926, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6934381723403931, + "num_tokens": 85619805.0, + "step": 3313 + }, + { + "epoch": 0.36393586646167364, + "ewc_loss": 2.8371810913085938e-05, + "grad_norm": 2.6062090396881104, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7029326558113098, + "num_tokens": 85640275.0, + "step": 3314 + }, + { + "epoch": 0.3640456841642873, + "ewc_loss": 2.86102294921875e-05, + "grad_norm": 2.313836097717285, + "learning_rate": 1e-06, + "loss": 1.0517, + "mean_token_accuracy": 0.6890003681182861, + "num_tokens": 85665126.0, + "step": 3315 + }, + { + "epoch": 0.36415550186690093, + "ewc_loss": 2.86102294921875e-05, + "grad_norm": 2.397709846496582, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.702762246131897, + "num_tokens": 85689650.0, + "step": 3316 + }, + { + "epoch": 0.3642653195695146, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.01369047164917, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6949117183685303, + "num_tokens": 85720567.0, + "step": 3317 + }, + { + "epoch": 0.3643751372721283, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.348334312438965, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6931790113449097, + "num_tokens": 85744555.0, + "step": 3318 + }, + { + "epoch": 0.3644849549747419, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.1810343265533447, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7010796070098877, + "num_tokens": 85769197.0, + "step": 3319 + }, + { + "epoch": 0.36459477267735557, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.1630287170410156, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7055119276046753, + "num_tokens": 85794927.0, + "step": 3320 + }, + { + "epoch": 0.3647045903799693, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.1780846118927, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.7056976556777954, + "num_tokens": 85820362.0, + "step": 3321 + }, + { + "epoch": 0.3648144080825829, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.3884410858154297, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7239171266555786, + "num_tokens": 85841839.0, + "step": 3322 + }, + { + "epoch": 0.36492422578519657, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.2396836280822754, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6898037791252136, + "num_tokens": 85867377.0, + "step": 3323 + }, + { + "epoch": 0.3650340434878102, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.224963903427124, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6814149022102356, + "num_tokens": 85894728.0, + "step": 3324 + }, + { + "epoch": 0.3651438611904239, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.1610593795776367, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7150071263313293, + "num_tokens": 85920460.0, + "step": 3325 + }, + { + "epoch": 0.36525367889303756, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.217256784439087, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6899310350418091, + "num_tokens": 85947417.0, + "step": 3326 + }, + { + "epoch": 0.3653634965956512, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.060859203338623, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6867595911026001, + "num_tokens": 85978475.0, + "step": 3327 + }, + { + "epoch": 0.3654733142982649, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.106440544128418, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6951050758361816, + "num_tokens": 86005268.0, + "step": 3328 + }, + { + "epoch": 0.36558313200087855, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.2052180767059326, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7132371664047241, + "num_tokens": 86032447.0, + "step": 3329 + }, + { + "epoch": 0.3656929497034922, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.380342483520508, + "learning_rate": 1e-06, + "loss": 1.0452, + "mean_token_accuracy": 0.6840659379959106, + "num_tokens": 86055965.0, + "step": 3330 + }, + { + "epoch": 0.36580276740610584, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.1303389072418213, + "learning_rate": 1e-06, + "loss": 1.0597, + "mean_token_accuracy": 0.6787580251693726, + "num_tokens": 86083407.0, + "step": 3331 + }, + { + "epoch": 0.36591258510871955, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.293543815612793, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.707216739654541, + "num_tokens": 86107077.0, + "step": 3332 + }, + { + "epoch": 0.3660224028113332, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.331364154815674, + "learning_rate": 1e-06, + "loss": 1.0565, + "mean_token_accuracy": 0.6912891864776611, + "num_tokens": 86131381.0, + "step": 3333 + }, + { + "epoch": 0.36613222051394684, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.4561493396759033, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6923230886459351, + "num_tokens": 86154122.0, + "step": 3334 + }, + { + "epoch": 0.3662420382165605, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.2616934776306152, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.6932374238967896, + "num_tokens": 86180267.0, + "step": 3335 + }, + { + "epoch": 0.3663518559191742, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 1.8875665664672852, + "learning_rate": 1e-06, + "loss": 1.0772, + "mean_token_accuracy": 0.6787443161010742, + "num_tokens": 86215594.0, + "step": 3336 + }, + { + "epoch": 0.36646167362178783, + "ewc_loss": 2.872943878173828e-05, + "grad_norm": 2.032181978225708, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6920737624168396, + "num_tokens": 86246071.0, + "step": 3337 + }, + { + "epoch": 0.3665714913244015, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.0726687908172607, + "learning_rate": 1e-06, + "loss": 1.1005, + "mean_token_accuracy": 0.672203004360199, + "num_tokens": 86277975.0, + "step": 3338 + }, + { + "epoch": 0.3666813090270152, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.1826889514923096, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6967105865478516, + "num_tokens": 86304784.0, + "step": 3339 + }, + { + "epoch": 0.3667911267296288, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.1430835723876953, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.713488757610321, + "num_tokens": 86334727.0, + "step": 3340 + }, + { + "epoch": 0.36690094443224247, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.2746429443359375, + "learning_rate": 1e-06, + "loss": 1.0854, + "mean_token_accuracy": 0.6947746276855469, + "num_tokens": 86360301.0, + "step": 3341 + }, + { + "epoch": 0.3670107621348561, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.237283945083618, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7116199135780334, + "num_tokens": 86384913.0, + "step": 3342 + }, + { + "epoch": 0.3671205798374698, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.2146918773651123, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.702502429485321, + "num_tokens": 86410825.0, + "step": 3343 + }, + { + "epoch": 0.36723039754008346, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.222038745880127, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6946749091148376, + "num_tokens": 86436639.0, + "step": 3344 + }, + { + "epoch": 0.3673402152426971, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.089099645614624, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.7003867626190186, + "num_tokens": 86465164.0, + "step": 3345 + }, + { + "epoch": 0.3674500329453108, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.3894031047821045, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6928151845932007, + "num_tokens": 86487983.0, + "step": 3346 + }, + { + "epoch": 0.36755985064792446, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.081852436065674, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7070080041885376, + "num_tokens": 86515243.0, + "step": 3347 + }, + { + "epoch": 0.3676696683505381, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.2823479175567627, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.709662675857544, + "num_tokens": 86540172.0, + "step": 3348 + }, + { + "epoch": 0.36777948605315175, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.040358781814575, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6818907260894775, + "num_tokens": 86571138.0, + "step": 3349 + }, + { + "epoch": 0.36788930375576545, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 1.9259958267211914, + "learning_rate": 1e-06, + "loss": 1.054, + "mean_token_accuracy": 0.6866825222969055, + "num_tokens": 86604797.0, + "step": 3350 + }, + { + "epoch": 0.3679991214583791, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.2657721042633057, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7137766480445862, + "num_tokens": 86628636.0, + "step": 3351 + }, + { + "epoch": 0.36810893916099274, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 1.8831573724746704, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6944026947021484, + "num_tokens": 86663120.0, + "step": 3352 + }, + { + "epoch": 0.3682187568636064, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.1452767848968506, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7199543714523315, + "num_tokens": 86689552.0, + "step": 3353 + }, + { + "epoch": 0.3683285745662201, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.400876522064209, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6995800137519836, + "num_tokens": 86712863.0, + "step": 3354 + }, + { + "epoch": 0.36843839226883374, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.304790496826172, + "learning_rate": 1e-06, + "loss": 1.0041, + "mean_token_accuracy": 0.6942247152328491, + "num_tokens": 86736563.0, + "step": 3355 + }, + { + "epoch": 0.3685482099714474, + "ewc_loss": 2.8848648071289062e-05, + "grad_norm": 2.0929341316223145, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.7047116756439209, + "num_tokens": 86764256.0, + "step": 3356 + }, + { + "epoch": 0.3686580276740611, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.3124494552612305, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7201936841011047, + "num_tokens": 86787797.0, + "step": 3357 + }, + { + "epoch": 0.36876784537667473, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.127631187438965, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7143412828445435, + "num_tokens": 86815390.0, + "step": 3358 + }, + { + "epoch": 0.3688776630792884, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 1.9661730527877808, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7336888909339905, + "num_tokens": 86846028.0, + "step": 3359 + }, + { + "epoch": 0.368987480781902, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.415271520614624, + "learning_rate": 1e-06, + "loss": 1.065, + "mean_token_accuracy": 0.6841980814933777, + "num_tokens": 86869114.0, + "step": 3360 + }, + { + "epoch": 0.3690972984845157, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.3746511936187744, + "learning_rate": 1e-06, + "loss": 1.0349, + "mean_token_accuracy": 0.6952909827232361, + "num_tokens": 86893563.0, + "step": 3361 + }, + { + "epoch": 0.36920711618712937, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.353921413421631, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7031270265579224, + "num_tokens": 86916057.0, + "step": 3362 + }, + { + "epoch": 0.369316933889743, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.1599106788635254, + "learning_rate": 1e-06, + "loss": 1.0342, + "mean_token_accuracy": 0.6881026029586792, + "num_tokens": 86944193.0, + "step": 3363 + }, + { + "epoch": 0.36942675159235666, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.0981645584106445, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6959922313690186, + "num_tokens": 86973130.0, + "step": 3364 + }, + { + "epoch": 0.36953656929497036, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.309262275695801, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7175984978675842, + "num_tokens": 86997045.0, + "step": 3365 + }, + { + "epoch": 0.369646386997584, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.3003528118133545, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6901828050613403, + "num_tokens": 87021295.0, + "step": 3366 + }, + { + "epoch": 0.36975620470019765, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.129376173019409, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.7014952898025513, + "num_tokens": 87046413.0, + "step": 3367 + }, + { + "epoch": 0.36986602240281136, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.3709259033203125, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6938062906265259, + "num_tokens": 87073882.0, + "step": 3368 + }, + { + "epoch": 0.369975840105425, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.271973133087158, + "learning_rate": 1e-06, + "loss": 1.0773, + "mean_token_accuracy": 0.6861749291419983, + "num_tokens": 87100385.0, + "step": 3369 + }, + { + "epoch": 0.37008565780803865, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.251596689224243, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7148561477661133, + "num_tokens": 87126593.0, + "step": 3370 + }, + { + "epoch": 0.3701954755106523, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 1.8543163537979126, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6891114711761475, + "num_tokens": 87161071.0, + "step": 3371 + }, + { + "epoch": 0.370305293213266, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.1891980171203613, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.697910487651825, + "num_tokens": 87186272.0, + "step": 3372 + }, + { + "epoch": 0.37041511091587964, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.185617446899414, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7095681428909302, + "num_tokens": 87212120.0, + "step": 3373 + }, + { + "epoch": 0.3705249286184933, + "ewc_loss": 2.9087066650390625e-05, + "grad_norm": 2.57179856300354, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7089284658432007, + "num_tokens": 87231883.0, + "step": 3374 + }, + { + "epoch": 0.370634746321107, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.0550618171691895, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6871925592422485, + "num_tokens": 87260370.0, + "step": 3375 + }, + { + "epoch": 0.37074456402372064, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.141247510910034, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6966917514801025, + "num_tokens": 87286894.0, + "step": 3376 + }, + { + "epoch": 0.3708543817263343, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 1.9858818054199219, + "learning_rate": 1e-06, + "loss": 1.0621, + "mean_token_accuracy": 0.6834619045257568, + "num_tokens": 87321398.0, + "step": 3377 + }, + { + "epoch": 0.3709641994289479, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.4077672958374023, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7168464660644531, + "num_tokens": 87345279.0, + "step": 3378 + }, + { + "epoch": 0.37107401713156163, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.5388450622558594, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.704887866973877, + "num_tokens": 87369573.0, + "step": 3379 + }, + { + "epoch": 0.3711838348341753, + "ewc_loss": 2.9206275939941406e-05, + "grad_norm": 2.2994191646575928, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6848235130310059, + "num_tokens": 87396176.0, + "step": 3380 + }, + { + "epoch": 0.3712936525367889, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 1.9929226636886597, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.696358323097229, + "num_tokens": 87426871.0, + "step": 3381 + }, + { + "epoch": 0.37140347023940257, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.1866040229797363, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7041622400283813, + "num_tokens": 87453349.0, + "step": 3382 + }, + { + "epoch": 0.37151328794201627, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 1.959761381149292, + "learning_rate": 1e-06, + "loss": 1.1025, + "mean_token_accuracy": 0.6968550086021423, + "num_tokens": 87484742.0, + "step": 3383 + }, + { + "epoch": 0.3716231056446299, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.4428727626800537, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7159616947174072, + "num_tokens": 87506704.0, + "step": 3384 + }, + { + "epoch": 0.37173292334724356, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.391918897628784, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6931915283203125, + "num_tokens": 87529839.0, + "step": 3385 + }, + { + "epoch": 0.37184274104985726, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.1777331829071045, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6928121447563171, + "num_tokens": 87558485.0, + "step": 3386 + }, + { + "epoch": 0.3719525587524709, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.1857213973999023, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7080755829811096, + "num_tokens": 87586352.0, + "step": 3387 + }, + { + "epoch": 0.37206237645508455, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.16508150100708, + "learning_rate": 1e-06, + "loss": 1.0868, + "mean_token_accuracy": 0.6886111497879028, + "num_tokens": 87614672.0, + "step": 3388 + }, + { + "epoch": 0.3721721941576982, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.0383975505828857, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7147619724273682, + "num_tokens": 87643313.0, + "step": 3389 + }, + { + "epoch": 0.3722820118603119, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.1442179679870605, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6983388662338257, + "num_tokens": 87670451.0, + "step": 3390 + }, + { + "epoch": 0.37239182956292555, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.4915390014648438, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6866857409477234, + "num_tokens": 87692044.0, + "step": 3391 + }, + { + "epoch": 0.3725016472655392, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 1.942763090133667, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7029673457145691, + "num_tokens": 87721089.0, + "step": 3392 + }, + { + "epoch": 0.37261146496815284, + "ewc_loss": 2.9325485229492188e-05, + "grad_norm": 2.3112399578094482, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.7025392055511475, + "num_tokens": 87743663.0, + "step": 3393 + }, + { + "epoch": 0.37272128267076654, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.49078369140625, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7115645408630371, + "num_tokens": 87764986.0, + "step": 3394 + }, + { + "epoch": 0.3728311003733802, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.2321414947509766, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7079786062240601, + "num_tokens": 87790380.0, + "step": 3395 + }, + { + "epoch": 0.37294091807599383, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.311478614807129, + "learning_rate": 1e-06, + "loss": 1.034, + "mean_token_accuracy": 0.6907909512519836, + "num_tokens": 87814012.0, + "step": 3396 + }, + { + "epoch": 0.37305073577860753, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.3355703353881836, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.690767765045166, + "num_tokens": 87838307.0, + "step": 3397 + }, + { + "epoch": 0.3731605534812212, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.2444019317626953, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.7020971775054932, + "num_tokens": 87864332.0, + "step": 3398 + }, + { + "epoch": 0.3732703711838348, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.6600258350372314, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6928670406341553, + "num_tokens": 87884903.0, + "step": 3399 + }, + { + "epoch": 0.37338018888644847, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.2895030975341797, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6971395015716553, + "num_tokens": 87908543.0, + "step": 3400 + }, + { + "epoch": 0.3734900065890622, + "ewc_loss": 2.944469451904297e-05, + "grad_norm": 2.4779584407806396, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7080029845237732, + "num_tokens": 87930022.0, + "step": 3401 + }, + { + "epoch": 0.3735998242916758, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 1.9743822813034058, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.692611813545227, + "num_tokens": 87961748.0, + "step": 3402 + }, + { + "epoch": 0.37370964199428947, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 1.8824951648712158, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7070562839508057, + "num_tokens": 87995355.0, + "step": 3403 + }, + { + "epoch": 0.37381945969690317, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 1.9594793319702148, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6810557842254639, + "num_tokens": 88026996.0, + "step": 3404 + }, + { + "epoch": 0.3739292773995168, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.169476270675659, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7221162915229797, + "num_tokens": 88052140.0, + "step": 3405 + }, + { + "epoch": 0.37403909510213046, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.4210734367370605, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6900136470794678, + "num_tokens": 88073640.0, + "step": 3406 + }, + { + "epoch": 0.3741489128047441, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.279186487197876, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.7000634074211121, + "num_tokens": 88096965.0, + "step": 3407 + }, + { + "epoch": 0.3742587305073578, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 1.9263360500335693, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.6915345191955566, + "num_tokens": 88128966.0, + "step": 3408 + }, + { + "epoch": 0.37436854820997145, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.219625234603882, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7128971815109253, + "num_tokens": 88152934.0, + "step": 3409 + }, + { + "epoch": 0.3744783659125851, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.5291173458099365, + "learning_rate": 1e-06, + "loss": 1.0578, + "mean_token_accuracy": 0.6851540207862854, + "num_tokens": 88175382.0, + "step": 3410 + }, + { + "epoch": 0.37458818361519874, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.124682903289795, + "learning_rate": 1e-06, + "loss": 1.0835, + "mean_token_accuracy": 0.6855418086051941, + "num_tokens": 88204938.0, + "step": 3411 + }, + { + "epoch": 0.37469800131781245, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.451568126678467, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7015604972839355, + "num_tokens": 88226259.0, + "step": 3412 + }, + { + "epoch": 0.3748078190204261, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.2532846927642822, + "learning_rate": 1e-06, + "loss": 1.0691, + "mean_token_accuracy": 0.6742173433303833, + "num_tokens": 88251893.0, + "step": 3413 + }, + { + "epoch": 0.37491763672303974, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.208665132522583, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7077080011367798, + "num_tokens": 88277786.0, + "step": 3414 + }, + { + "epoch": 0.37502745442565344, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.4302849769592285, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7031898498535156, + "num_tokens": 88300143.0, + "step": 3415 + }, + { + "epoch": 0.3751372721282671, + "ewc_loss": 2.956390380859375e-05, + "grad_norm": 2.4131078720092773, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7049682140350342, + "num_tokens": 88323457.0, + "step": 3416 + }, + { + "epoch": 0.37524708983088073, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.5899436473846436, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7131470441818237, + "num_tokens": 88343643.0, + "step": 3417 + }, + { + "epoch": 0.3753569075334944, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.0322513580322266, + "learning_rate": 1e-06, + "loss": 1.0837, + "mean_token_accuracy": 0.6756139993667603, + "num_tokens": 88373513.0, + "step": 3418 + }, + { + "epoch": 0.3754667252361081, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.0856192111968994, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6924819946289062, + "num_tokens": 88402245.0, + "step": 3419 + }, + { + "epoch": 0.3755765429387217, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.139800786972046, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6941859126091003, + "num_tokens": 88430919.0, + "step": 3420 + }, + { + "epoch": 0.37568636064133537, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.063953161239624, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7114682197570801, + "num_tokens": 88460174.0, + "step": 3421 + }, + { + "epoch": 0.37579617834394907, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.1291046142578125, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6906675100326538, + "num_tokens": 88488063.0, + "step": 3422 + }, + { + "epoch": 0.3759059960465627, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 2.0773372650146484, + "learning_rate": 1e-06, + "loss": 1.0335, + "mean_token_accuracy": 0.6949200630187988, + "num_tokens": 88520294.0, + "step": 3423 + }, + { + "epoch": 0.37601581374917636, + "ewc_loss": 2.968311309814453e-05, + "grad_norm": 1.934207797050476, + "learning_rate": 1e-06, + "loss": 1.067, + "mean_token_accuracy": 0.6786653995513916, + "num_tokens": 88554544.0, + "step": 3424 + }, + { + "epoch": 0.37612563145179, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2446579933166504, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7082239389419556, + "num_tokens": 88578934.0, + "step": 3425 + }, + { + "epoch": 0.3762354491544037, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.3644518852233887, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7076865434646606, + "num_tokens": 88600495.0, + "step": 3426 + }, + { + "epoch": 0.37634526685701736, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.13181734085083, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7149049043655396, + "num_tokens": 88627034.0, + "step": 3427 + }, + { + "epoch": 0.376455084559631, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.286822557449341, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7026838660240173, + "num_tokens": 88650799.0, + "step": 3428 + }, + { + "epoch": 0.37656490226224465, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.468897581100464, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7079814672470093, + "num_tokens": 88670895.0, + "step": 3429 + }, + { + "epoch": 0.37667471996485835, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2701199054718018, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6955319046974182, + "num_tokens": 88693919.0, + "step": 3430 + }, + { + "epoch": 0.376784537667472, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.159005880355835, + "learning_rate": 1e-06, + "loss": 1.0732, + "mean_token_accuracy": 0.6748132705688477, + "num_tokens": 88721418.0, + "step": 3431 + }, + { + "epoch": 0.37689435537008564, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.307530403137207, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7214497327804565, + "num_tokens": 88745403.0, + "step": 3432 + }, + { + "epoch": 0.37700417307269934, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 1.9078044891357422, + "learning_rate": 1e-06, + "loss": 1.0615, + "mean_token_accuracy": 0.6859367489814758, + "num_tokens": 88780492.0, + "step": 3433 + }, + { + "epoch": 0.377113990775313, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.092587947845459, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6895054578781128, + "num_tokens": 88811625.0, + "step": 3434 + }, + { + "epoch": 0.37722380847792664, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.046755075454712, + "learning_rate": 1e-06, + "loss": 1.1238, + "mean_token_accuracy": 0.6779818534851074, + "num_tokens": 88843377.0, + "step": 3435 + }, + { + "epoch": 0.3773336261805403, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.1651198863983154, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7045670747756958, + "num_tokens": 88871222.0, + "step": 3436 + }, + { + "epoch": 0.377443443883154, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 1.8734774589538574, + "learning_rate": 1e-06, + "loss": 1.0789, + "mean_token_accuracy": 0.6748142838478088, + "num_tokens": 88906003.0, + "step": 3437 + }, + { + "epoch": 0.37755326158576763, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.1847503185272217, + "learning_rate": 1e-06, + "loss": 1.0687, + "mean_token_accuracy": 0.6853810548782349, + "num_tokens": 88934175.0, + "step": 3438 + }, + { + "epoch": 0.3776630792883813, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 1.9809846878051758, + "learning_rate": 1e-06, + "loss": 1.0866, + "mean_token_accuracy": 0.6775432825088501, + "num_tokens": 88967305.0, + "step": 3439 + }, + { + "epoch": 0.3777728969909949, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.20078444480896, + "learning_rate": 1e-06, + "loss": 1.0898, + "mean_token_accuracy": 0.6727472543716431, + "num_tokens": 88995058.0, + "step": 3440 + }, + { + "epoch": 0.3778827146936086, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.526686429977417, + "learning_rate": 1e-06, + "loss": 1.0462, + "mean_token_accuracy": 0.6864608526229858, + "num_tokens": 89016825.0, + "step": 3441 + }, + { + "epoch": 0.37799253239622227, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.241727352142334, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6997612714767456, + "num_tokens": 89044034.0, + "step": 3442 + }, + { + "epoch": 0.3781023500988359, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.3072147369384766, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.7012745141983032, + "num_tokens": 89068052.0, + "step": 3443 + }, + { + "epoch": 0.3782121678014496, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.119511842727661, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7133539915084839, + "num_tokens": 89094888.0, + "step": 3444 + }, + { + "epoch": 0.37832198550406326, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.1585423946380615, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7021880149841309, + "num_tokens": 89120885.0, + "step": 3445 + }, + { + "epoch": 0.3784318032066769, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.0592081546783447, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7164910435676575, + "num_tokens": 89148606.0, + "step": 3446 + }, + { + "epoch": 0.37854162090929055, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.4458000659942627, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7251890897750854, + "num_tokens": 89170649.0, + "step": 3447 + }, + { + "epoch": 0.37865143861190426, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.223544120788574, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.6916453838348389, + "num_tokens": 89196599.0, + "step": 3448 + }, + { + "epoch": 0.3787612563145179, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.4007325172424316, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7104931473731995, + "num_tokens": 89218130.0, + "step": 3449 + }, + { + "epoch": 0.37887107401713155, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2569727897644043, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6971854567527771, + "num_tokens": 89242478.0, + "step": 3450 + }, + { + "epoch": 0.37898089171974525, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2839765548706055, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6953852772712708, + "num_tokens": 89267931.0, + "step": 3451 + }, + { + "epoch": 0.3790907094223589, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 1.9929730892181396, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7231872081756592, + "num_tokens": 89298706.0, + "step": 3452 + }, + { + "epoch": 0.37920052712497254, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.224915027618408, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.7110379934310913, + "num_tokens": 89325375.0, + "step": 3453 + }, + { + "epoch": 0.3793103448275862, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.4866204261779785, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7103085517883301, + "num_tokens": 89346526.0, + "step": 3454 + }, + { + "epoch": 0.3794201625301999, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.236008644104004, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6930482387542725, + "num_tokens": 89372830.0, + "step": 3455 + }, + { + "epoch": 0.37952998023281354, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2654483318328857, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.6989737749099731, + "num_tokens": 89396497.0, + "step": 3456 + }, + { + "epoch": 0.3796397979354272, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2296719551086426, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7065967321395874, + "num_tokens": 89422444.0, + "step": 3457 + }, + { + "epoch": 0.3797496156380408, + "ewc_loss": 2.9802322387695312e-05, + "grad_norm": 2.2346079349517822, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6842559576034546, + "num_tokens": 89447722.0, + "step": 3458 + }, + { + "epoch": 0.37985943334065453, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.1943609714508057, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7229613661766052, + "num_tokens": 89472695.0, + "step": 3459 + }, + { + "epoch": 0.3799692510432682, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 1.993639588356018, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6925865411758423, + "num_tokens": 89504156.0, + "step": 3460 + }, + { + "epoch": 0.3800790687458818, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.010942220687866, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6930055618286133, + "num_tokens": 89533887.0, + "step": 3461 + }, + { + "epoch": 0.3801888864484955, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.3787131309509277, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7052842974662781, + "num_tokens": 89558518.0, + "step": 3462 + }, + { + "epoch": 0.38029870415110917, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.5503194332122803, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6912588477134705, + "num_tokens": 89579890.0, + "step": 3463 + }, + { + "epoch": 0.3804085218537228, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.2769815921783447, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7050944566726685, + "num_tokens": 89604732.0, + "step": 3464 + }, + { + "epoch": 0.38051833955633646, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.505664348602295, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7127397656440735, + "num_tokens": 89625793.0, + "step": 3465 + }, + { + "epoch": 0.38062815725895016, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 2.311192274093628, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7154508829116821, + "num_tokens": 89648859.0, + "step": 3466 + }, + { + "epoch": 0.3807379749615638, + "ewc_loss": 2.9921531677246094e-05, + "grad_norm": 1.9941935539245605, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6992398500442505, + "num_tokens": 89679663.0, + "step": 3467 + }, + { + "epoch": 0.38084779266417745, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.2167398929595947, + "learning_rate": 1e-06, + "loss": 1.0698, + "mean_token_accuracy": 0.6823748350143433, + "num_tokens": 89705345.0, + "step": 3468 + }, + { + "epoch": 0.3809576103667911, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.2480578422546387, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.6928834915161133, + "num_tokens": 89730700.0, + "step": 3469 + }, + { + "epoch": 0.3810674280694048, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 1.968418836593628, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.711901843547821, + "num_tokens": 89763152.0, + "step": 3470 + }, + { + "epoch": 0.38117724577201845, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.3436238765716553, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6918954849243164, + "num_tokens": 89786840.0, + "step": 3471 + }, + { + "epoch": 0.3812870634746321, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.223938465118408, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7139431238174438, + "num_tokens": 89811531.0, + "step": 3472 + }, + { + "epoch": 0.3813968811772458, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.1874985694885254, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.6942539215087891, + "num_tokens": 89837392.0, + "step": 3473 + }, + { + "epoch": 0.38150669887985944, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.3048501014709473, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.714928388595581, + "num_tokens": 89861657.0, + "step": 3474 + }, + { + "epoch": 0.3816165165824731, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.2323057651519775, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.700562596321106, + "num_tokens": 89888089.0, + "step": 3475 + }, + { + "epoch": 0.38172633428508673, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 1.9094003438949585, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.7004682421684265, + "num_tokens": 89922136.0, + "step": 3476 + }, + { + "epoch": 0.38183615198770043, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.2227365970611572, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7069557309150696, + "num_tokens": 89946651.0, + "step": 3477 + }, + { + "epoch": 0.3819459696903141, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.414665460586548, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7043525576591492, + "num_tokens": 89970358.0, + "step": 3478 + }, + { + "epoch": 0.3820557873929277, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.4053611755371094, + "learning_rate": 1e-06, + "loss": 1.0396, + "mean_token_accuracy": 0.6924155950546265, + "num_tokens": 89994077.0, + "step": 3479 + }, + { + "epoch": 0.3821656050955414, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.4468705654144287, + "learning_rate": 1e-06, + "loss": 1.0428, + "mean_token_accuracy": 0.6877967119216919, + "num_tokens": 90017250.0, + "step": 3480 + }, + { + "epoch": 0.3822754227981551, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.1158435344696045, + "learning_rate": 1e-06, + "loss": 1.0652, + "mean_token_accuracy": 0.6811103820800781, + "num_tokens": 90046866.0, + "step": 3481 + }, + { + "epoch": 0.3823852405007687, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.4138522148132324, + "learning_rate": 1e-06, + "loss": 1.0928, + "mean_token_accuracy": 0.6805411577224731, + "num_tokens": 90070820.0, + "step": 3482 + }, + { + "epoch": 0.38249505820338237, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.0082669258117676, + "learning_rate": 1e-06, + "loss": 1.0503, + "mean_token_accuracy": 0.6839213371276855, + "num_tokens": 90103266.0, + "step": 3483 + }, + { + "epoch": 0.38260487590599607, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.313383102416992, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.7069786190986633, + "num_tokens": 90128277.0, + "step": 3484 + }, + { + "epoch": 0.3827146936086097, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.076237440109253, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6924776434898376, + "num_tokens": 90158627.0, + "step": 3485 + }, + { + "epoch": 0.38282451131122336, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.1173291206359863, + "learning_rate": 1e-06, + "loss": 1.0695, + "mean_token_accuracy": 0.6782377362251282, + "num_tokens": 90188177.0, + "step": 3486 + }, + { + "epoch": 0.382934329013837, + "ewc_loss": 3.0040740966796875e-05, + "grad_norm": 2.0136256217956543, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6896964311599731, + "num_tokens": 90219370.0, + "step": 3487 + }, + { + "epoch": 0.3830441467164507, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.5058977603912354, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7278357744216919, + "num_tokens": 90239794.0, + "step": 3488 + }, + { + "epoch": 0.38315396441906435, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.0570836067199707, + "learning_rate": 1e-06, + "loss": 1.0681, + "mean_token_accuracy": 0.6797892451286316, + "num_tokens": 90270780.0, + "step": 3489 + }, + { + "epoch": 0.383263782121678, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.046581983566284, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7171005010604858, + "num_tokens": 90299170.0, + "step": 3490 + }, + { + "epoch": 0.3833735998242917, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.0364208221435547, + "learning_rate": 1e-06, + "loss": 1.0921, + "mean_token_accuracy": 0.6766979694366455, + "num_tokens": 90330360.0, + "step": 3491 + }, + { + "epoch": 0.38348341752690535, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.554697275161743, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7086256742477417, + "num_tokens": 90350344.0, + "step": 3492 + }, + { + "epoch": 0.383593235229519, + "ewc_loss": 3.0279159545898438e-05, + "grad_norm": 2.392717123031616, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7180206775665283, + "num_tokens": 90374020.0, + "step": 3493 + }, + { + "epoch": 0.38370305293213264, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.3786122798919678, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7030118107795715, + "num_tokens": 90398968.0, + "step": 3494 + }, + { + "epoch": 0.38381287063474634, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.2890303134918213, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7345636487007141, + "num_tokens": 90424099.0, + "step": 3495 + }, + { + "epoch": 0.38392268833736, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.3188602924346924, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6981379389762878, + "num_tokens": 90447219.0, + "step": 3496 + }, + { + "epoch": 0.38403250603997363, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.1403067111968994, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.7089909315109253, + "num_tokens": 90475098.0, + "step": 3497 + }, + { + "epoch": 0.38414232374258733, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.556269884109497, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.736579179763794, + "num_tokens": 90493385.0, + "step": 3498 + }, + { + "epoch": 0.384252141445201, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.1279778480529785, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7044939994812012, + "num_tokens": 90524253.0, + "step": 3499 + }, + { + "epoch": 0.3843619591478146, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.0597646236419678, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6931544542312622, + "num_tokens": 90554973.0, + "step": 3500 + }, + { + "epoch": 0.38447177685042827, + "ewc_loss": 3.0159950256347656e-05, + "grad_norm": 2.0799241065979004, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.701548159122467, + "num_tokens": 90584515.0, + "step": 3501 + }, + { + "epoch": 0.38458159455304197, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.3625664710998535, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7006545066833496, + "num_tokens": 90608481.0, + "step": 3502 + }, + { + "epoch": 0.3846914122556556, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.1851143836975098, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6953929662704468, + "num_tokens": 90633367.0, + "step": 3503 + }, + { + "epoch": 0.38480122995826926, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.3889145851135254, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7009320259094238, + "num_tokens": 90657309.0, + "step": 3504 + }, + { + "epoch": 0.3849110476608829, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.433140277862549, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.701004147529602, + "num_tokens": 90681447.0, + "step": 3505 + }, + { + "epoch": 0.3850208653634966, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.3332302570343018, + "learning_rate": 1e-06, + "loss": 1.0454, + "mean_token_accuracy": 0.6891282796859741, + "num_tokens": 90705941.0, + "step": 3506 + }, + { + "epoch": 0.38513068306611026, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.0079574584960938, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7134698629379272, + "num_tokens": 90734216.0, + "step": 3507 + }, + { + "epoch": 0.3852405007687239, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.1374995708465576, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7057366371154785, + "num_tokens": 90761487.0, + "step": 3508 + }, + { + "epoch": 0.3853503184713376, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.1321916580200195, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.691473126411438, + "num_tokens": 90789622.0, + "step": 3509 + }, + { + "epoch": 0.38546013617395125, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.026819944381714, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6919836401939392, + "num_tokens": 90819427.0, + "step": 3510 + }, + { + "epoch": 0.3855699538765649, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.2152841091156006, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6879712343215942, + "num_tokens": 90846329.0, + "step": 3511 + }, + { + "epoch": 0.38567977157917854, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.0779879093170166, + "learning_rate": 1e-06, + "loss": 1.0963, + "mean_token_accuracy": 0.6710950136184692, + "num_tokens": 90875571.0, + "step": 3512 + }, + { + "epoch": 0.38578958928179224, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.5256123542785645, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7336070537567139, + "num_tokens": 90895193.0, + "step": 3513 + }, + { + "epoch": 0.3858994069844059, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 1.9662140607833862, + "learning_rate": 1e-06, + "loss": 1.0914, + "mean_token_accuracy": 0.6782378554344177, + "num_tokens": 90926516.0, + "step": 3514 + }, + { + "epoch": 0.38600922468701954, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.2233781814575195, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6781560182571411, + "num_tokens": 90955106.0, + "step": 3515 + }, + { + "epoch": 0.3861190423896332, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.2337894439697266, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.7021811604499817, + "num_tokens": 90980991.0, + "step": 3516 + }, + { + "epoch": 0.3862288600922469, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.442857265472412, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6984738707542419, + "num_tokens": 91004316.0, + "step": 3517 + }, + { + "epoch": 0.38633867779486053, + "ewc_loss": 3.039836883544922e-05, + "grad_norm": 2.2739596366882324, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7257874011993408, + "num_tokens": 91027761.0, + "step": 3518 + }, + { + "epoch": 0.3864484954974742, + "ewc_loss": 3.0517578125e-05, + "grad_norm": 2.369140386581421, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6872847676277161, + "num_tokens": 91050506.0, + "step": 3519 + }, + { + "epoch": 0.3865583132000879, + "ewc_loss": 3.0517578125e-05, + "grad_norm": 1.972306251525879, + "learning_rate": 1e-06, + "loss": 1.0584, + "mean_token_accuracy": 0.6783382296562195, + "num_tokens": 91082101.0, + "step": 3520 + }, + { + "epoch": 0.3866681309027015, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.4493603706359863, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7195848822593689, + "num_tokens": 91104010.0, + "step": 3521 + }, + { + "epoch": 0.38677794860531517, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.183077573776245, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.7050207257270813, + "num_tokens": 91129990.0, + "step": 3522 + }, + { + "epoch": 0.3868877663079288, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1277527809143066, + "learning_rate": 1e-06, + "loss": 1.0938, + "mean_token_accuracy": 0.6755369901657104, + "num_tokens": 91158919.0, + "step": 3523 + }, + { + "epoch": 0.3869975840105425, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.339075803756714, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7061439752578735, + "num_tokens": 91180887.0, + "step": 3524 + }, + { + "epoch": 0.38710740171315616, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.03769588470459, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6908366084098816, + "num_tokens": 91211256.0, + "step": 3525 + }, + { + "epoch": 0.3872172194157698, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.2578282356262207, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.7004801630973816, + "num_tokens": 91235786.0, + "step": 3526 + }, + { + "epoch": 0.3873270371183835, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.4000885486602783, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7001484036445618, + "num_tokens": 91259530.0, + "step": 3527 + }, + { + "epoch": 0.38743685482099716, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.3239591121673584, + "learning_rate": 1e-06, + "loss": 1.0501, + "mean_token_accuracy": 0.6824498176574707, + "num_tokens": 91284855.0, + "step": 3528 + }, + { + "epoch": 0.3875466725236108, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1097569465637207, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7229939699172974, + "num_tokens": 91311673.0, + "step": 3529 + }, + { + "epoch": 0.38765649022622445, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.2205119132995605, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6941037774085999, + "num_tokens": 91337540.0, + "step": 3530 + }, + { + "epoch": 0.38776630792883815, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.212406635284424, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6976808309555054, + "num_tokens": 91363407.0, + "step": 3531 + }, + { + "epoch": 0.3878761256314518, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.639188289642334, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7003092765808105, + "num_tokens": 91382935.0, + "step": 3532 + }, + { + "epoch": 0.38798594333406544, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.3282113075256348, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6996164321899414, + "num_tokens": 91407494.0, + "step": 3533 + }, + { + "epoch": 0.3880957610366791, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.124906063079834, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7202008962631226, + "num_tokens": 91434532.0, + "step": 3534 + }, + { + "epoch": 0.3882055787392928, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.70752215385437, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7301950454711914, + "num_tokens": 91453366.0, + "step": 3535 + }, + { + "epoch": 0.38831539644190644, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.197760581970215, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6977883577346802, + "num_tokens": 91479978.0, + "step": 3536 + }, + { + "epoch": 0.3884252141445201, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 1.9583748579025269, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6931645274162292, + "num_tokens": 91512790.0, + "step": 3537 + }, + { + "epoch": 0.3885350318471338, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.122606039047241, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7000216245651245, + "num_tokens": 91539101.0, + "step": 3538 + }, + { + "epoch": 0.38864484954974743, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.505455732345581, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.693170964717865, + "num_tokens": 91561374.0, + "step": 3539 + }, + { + "epoch": 0.3887546672523611, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 1.9766908884048462, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7075085639953613, + "num_tokens": 91589580.0, + "step": 3540 + }, + { + "epoch": 0.3888644849549747, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.366468667984009, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6913644671440125, + "num_tokens": 91612399.0, + "step": 3541 + }, + { + "epoch": 0.3889743026575884, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 1.9824796915054321, + "learning_rate": 1e-06, + "loss": 1.0764, + "mean_token_accuracy": 0.6811937093734741, + "num_tokens": 91644091.0, + "step": 3542 + }, + { + "epoch": 0.38908412036020207, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.153904438018799, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.699897289276123, + "num_tokens": 91670530.0, + "step": 3543 + }, + { + "epoch": 0.3891939380628157, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.4140918254852295, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.7070430517196655, + "num_tokens": 91694669.0, + "step": 3544 + }, + { + "epoch": 0.38930375576542936, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 1.9466321468353271, + "learning_rate": 1e-06, + "loss": 1.0422, + "mean_token_accuracy": 0.6854676008224487, + "num_tokens": 91728597.0, + "step": 3545 + }, + { + "epoch": 0.38941357346804306, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.047121524810791, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6987521648406982, + "num_tokens": 91756801.0, + "step": 3546 + }, + { + "epoch": 0.3895233911706567, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.50799298286438, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6883370876312256, + "num_tokens": 91778500.0, + "step": 3547 + }, + { + "epoch": 0.38963320887327035, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1666104793548584, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6961303949356079, + "num_tokens": 91804944.0, + "step": 3548 + }, + { + "epoch": 0.38974302657588406, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1142823696136475, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7287997603416443, + "num_tokens": 91833600.0, + "step": 3549 + }, + { + "epoch": 0.3898528442784977, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.198930501937866, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7290869951248169, + "num_tokens": 91856897.0, + "step": 3550 + }, + { + "epoch": 0.38996266198111135, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.362271547317505, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6970911622047424, + "num_tokens": 91878769.0, + "step": 3551 + }, + { + "epoch": 0.390072479683725, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.331834077835083, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7124341130256653, + "num_tokens": 91901129.0, + "step": 3552 + }, + { + "epoch": 0.3901822973863387, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.4452836513519287, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7274686694145203, + "num_tokens": 91921789.0, + "step": 3553 + }, + { + "epoch": 0.39029211508895234, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.193236827850342, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6930738091468811, + "num_tokens": 91949252.0, + "step": 3554 + }, + { + "epoch": 0.390401932791566, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.378113031387329, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7157860994338989, + "num_tokens": 91970893.0, + "step": 3555 + }, + { + "epoch": 0.3905117504941797, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.0242581367492676, + "learning_rate": 1e-06, + "loss": 1.0881, + "mean_token_accuracy": 0.6822456121444702, + "num_tokens": 92003459.0, + "step": 3556 + }, + { + "epoch": 0.39062156819679333, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1903958320617676, + "learning_rate": 1e-06, + "loss": 1.0729, + "mean_token_accuracy": 0.6841915845870972, + "num_tokens": 92032949.0, + "step": 3557 + }, + { + "epoch": 0.390731385899407, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.254606246948242, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6993691325187683, + "num_tokens": 92058753.0, + "step": 3558 + }, + { + "epoch": 0.3908412036020206, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.2846717834472656, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6764113903045654, + "num_tokens": 92086817.0, + "step": 3559 + }, + { + "epoch": 0.3909510213046343, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.148854970932007, + "learning_rate": 1e-06, + "loss": 1.0736, + "mean_token_accuracy": 0.6922265291213989, + "num_tokens": 92114863.0, + "step": 3560 + }, + { + "epoch": 0.391060839007248, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.169140577316284, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6933484077453613, + "num_tokens": 92143174.0, + "step": 3561 + }, + { + "epoch": 0.3911706567098616, + "ewc_loss": 3.0994415283203125e-05, + "grad_norm": 2.1960861682891846, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7063301801681519, + "num_tokens": 92169271.0, + "step": 3562 + }, + { + "epoch": 0.39128047441247527, + "ewc_loss": 3.123283386230469e-05, + "grad_norm": 2.0859782695770264, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.7013899683952332, + "num_tokens": 92197530.0, + "step": 3563 + }, + { + "epoch": 0.39139029211508897, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.0908963680267334, + "learning_rate": 1e-06, + "loss": 1.0885, + "mean_token_accuracy": 0.6678511500358582, + "num_tokens": 92226522.0, + "step": 3564 + }, + { + "epoch": 0.3915001098177026, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.0257558822631836, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6894223093986511, + "num_tokens": 92255728.0, + "step": 3565 + }, + { + "epoch": 0.39160992752031626, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 1.9242687225341797, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7090878486633301, + "num_tokens": 92288415.0, + "step": 3566 + }, + { + "epoch": 0.39171974522292996, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.31416392326355, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7070291042327881, + "num_tokens": 92310582.0, + "step": 3567 + }, + { + "epoch": 0.3918295629255436, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.2540156841278076, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.691403329372406, + "num_tokens": 92337192.0, + "step": 3568 + }, + { + "epoch": 0.39193938062815725, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.181940793991089, + "learning_rate": 1e-06, + "loss": 1.0791, + "mean_token_accuracy": 0.6763756275177002, + "num_tokens": 92366075.0, + "step": 3569 + }, + { + "epoch": 0.3920491983307709, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.2829227447509766, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6958070993423462, + "num_tokens": 92391433.0, + "step": 3570 + }, + { + "epoch": 0.3921590160333846, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.2515602111816406, + "learning_rate": 1e-06, + "loss": 1.0995, + "mean_token_accuracy": 0.6754815578460693, + "num_tokens": 92419463.0, + "step": 3571 + }, + { + "epoch": 0.39226883373599825, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.3802640438079834, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7048171758651733, + "num_tokens": 92441456.0, + "step": 3572 + }, + { + "epoch": 0.3923786514386119, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.111558437347412, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7224040627479553, + "num_tokens": 92466494.0, + "step": 3573 + }, + { + "epoch": 0.3924884691412256, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.184544086456299, + "learning_rate": 1e-06, + "loss": 1.0488, + "mean_token_accuracy": 0.6866445541381836, + "num_tokens": 92491099.0, + "step": 3574 + }, + { + "epoch": 0.39259828684383924, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.2519187927246094, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7236877083778381, + "num_tokens": 92518006.0, + "step": 3575 + }, + { + "epoch": 0.3927081045464529, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.1991114616394043, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7071535587310791, + "num_tokens": 92542700.0, + "step": 3576 + }, + { + "epoch": 0.39281792224906653, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.301208019256592, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6942030191421509, + "num_tokens": 92566279.0, + "step": 3577 + }, + { + "epoch": 0.39292773995168023, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 1.996405839920044, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6871830224990845, + "num_tokens": 92599121.0, + "step": 3578 + }, + { + "epoch": 0.3930375576542939, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.4411461353302, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.701668381690979, + "num_tokens": 92624001.0, + "step": 3579 + }, + { + "epoch": 0.3931473753569075, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1141679286956787, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.702548623085022, + "num_tokens": 92652425.0, + "step": 3580 + }, + { + "epoch": 0.39325719305952117, + "ewc_loss": 3.147125244140625e-05, + "grad_norm": 2.4235892295837402, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6972701549530029, + "num_tokens": 92674724.0, + "step": 3581 + }, + { + "epoch": 0.39336701076213487, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.2535240650177, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.6960185766220093, + "num_tokens": 92699667.0, + "step": 3582 + }, + { + "epoch": 0.3934768284647485, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1583142280578613, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6902619004249573, + "num_tokens": 92727517.0, + "step": 3583 + }, + { + "epoch": 0.39358664616736216, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.5995218753814697, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7112071514129639, + "num_tokens": 92747373.0, + "step": 3584 + }, + { + "epoch": 0.39369646386997587, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0945208072662354, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.7084662914276123, + "num_tokens": 92775832.0, + "step": 3585 + }, + { + "epoch": 0.3938062815725895, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 1.9080400466918945, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6987740993499756, + "num_tokens": 92807378.0, + "step": 3586 + }, + { + "epoch": 0.39391609927520316, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.3147294521331787, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7059783935546875, + "num_tokens": 92832155.0, + "step": 3587 + }, + { + "epoch": 0.3940259169778168, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.321568727493286, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7034095525741577, + "num_tokens": 92856290.0, + "step": 3588 + }, + { + "epoch": 0.3941357346804305, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.2723658084869385, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6971200704574585, + "num_tokens": 92881250.0, + "step": 3589 + }, + { + "epoch": 0.39424555238304415, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.3730578422546387, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6935300230979919, + "num_tokens": 92904024.0, + "step": 3590 + }, + { + "epoch": 0.3943553700856578, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0527360439300537, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7031009197235107, + "num_tokens": 92932387.0, + "step": 3591 + }, + { + "epoch": 0.39446518778827144, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1967227458953857, + "learning_rate": 1e-06, + "loss": 1.0595, + "mean_token_accuracy": 0.6817752122879028, + "num_tokens": 92959209.0, + "step": 3592 + }, + { + "epoch": 0.39457500549088514, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.73185396194458, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7174087166786194, + "num_tokens": 92977796.0, + "step": 3593 + }, + { + "epoch": 0.3946848231934988, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.3054370880126953, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7190459966659546, + "num_tokens": 93002954.0, + "step": 3594 + }, + { + "epoch": 0.39479464089611244, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0056583881378174, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6908135414123535, + "num_tokens": 93033062.0, + "step": 3595 + }, + { + "epoch": 0.39490445859872614, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.2445075511932373, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6923947930335999, + "num_tokens": 93059527.0, + "step": 3596 + }, + { + "epoch": 0.3950142763013398, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.157696008682251, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7170752286911011, + "num_tokens": 93085994.0, + "step": 3597 + }, + { + "epoch": 0.39512409400395343, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1873531341552734, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6916606426239014, + "num_tokens": 93110793.0, + "step": 3598 + }, + { + "epoch": 0.3952339117065671, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1895668506622314, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7123430967330933, + "num_tokens": 93137040.0, + "step": 3599 + }, + { + "epoch": 0.3953437294091808, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.3377323150634766, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7025011777877808, + "num_tokens": 93160651.0, + "step": 3600 + }, + { + "epoch": 0.3954535471117944, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0827958583831787, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7088170051574707, + "num_tokens": 93189203.0, + "step": 3601 + }, + { + "epoch": 0.39556336481440807, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.239365339279175, + "learning_rate": 1e-06, + "loss": 1.0688, + "mean_token_accuracy": 0.6850008964538574, + "num_tokens": 93215152.0, + "step": 3602 + }, + { + "epoch": 0.39567318251702177, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0634877681732178, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7106143236160278, + "num_tokens": 93242481.0, + "step": 3603 + }, + { + "epoch": 0.3957830002196354, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 1.9387199878692627, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6810941696166992, + "num_tokens": 93273852.0, + "step": 3604 + }, + { + "epoch": 0.39589281792224906, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.2931439876556396, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.7008227705955505, + "num_tokens": 93297600.0, + "step": 3605 + }, + { + "epoch": 0.3960026356248627, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.2046866416931152, + "learning_rate": 1e-06, + "loss": 1.1064, + "mean_token_accuracy": 0.6709674000740051, + "num_tokens": 93326205.0, + "step": 3606 + }, + { + "epoch": 0.3961124533274764, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.139495611190796, + "learning_rate": 1e-06, + "loss": 1.0906, + "mean_token_accuracy": 0.6764697432518005, + "num_tokens": 93356719.0, + "step": 3607 + }, + { + "epoch": 0.39622227103009006, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.350666046142578, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7205768823623657, + "num_tokens": 93379681.0, + "step": 3608 + }, + { + "epoch": 0.3963320887327037, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.145538568496704, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6946749091148376, + "num_tokens": 93407884.0, + "step": 3609 + }, + { + "epoch": 0.39644190643531735, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.105776309967041, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.69972825050354, + "num_tokens": 93434797.0, + "step": 3610 + }, + { + "epoch": 0.39655172413793105, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.1133153438568115, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7009111642837524, + "num_tokens": 93461110.0, + "step": 3611 + }, + { + "epoch": 0.3966615418405447, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.0441243648529053, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6930866837501526, + "num_tokens": 93493577.0, + "step": 3612 + }, + { + "epoch": 0.39677135954315834, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.3759937286376953, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7109077572822571, + "num_tokens": 93516458.0, + "step": 3613 + }, + { + "epoch": 0.39688117724577204, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.105778455734253, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7125383615493774, + "num_tokens": 93543958.0, + "step": 3614 + }, + { + "epoch": 0.3969909949483857, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.9408960342407227, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.709431529045105, + "num_tokens": 93560899.0, + "step": 3615 + }, + { + "epoch": 0.39710081265099934, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.5990078449249268, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7161009311676025, + "num_tokens": 93579658.0, + "step": 3616 + }, + { + "epoch": 0.397210630353613, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.081080675125122, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7452524900436401, + "num_tokens": 93605222.0, + "step": 3617 + }, + { + "epoch": 0.3973204480562267, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 2.251706600189209, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7055338621139526, + "num_tokens": 93631550.0, + "step": 3618 + }, + { + "epoch": 0.39743026575884033, + "ewc_loss": 3.1948089599609375e-05, + "grad_norm": 1.9968230724334717, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6922414898872375, + "num_tokens": 93661110.0, + "step": 3619 + }, + { + "epoch": 0.397540083461454, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.611656904220581, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7037657499313354, + "num_tokens": 93680548.0, + "step": 3620 + }, + { + "epoch": 0.3976499011640676, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0943045616149902, + "learning_rate": 1e-06, + "loss": 1.1552, + "mean_token_accuracy": 0.6580560207366943, + "num_tokens": 93710057.0, + "step": 3621 + }, + { + "epoch": 0.3977597188666813, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0322911739349365, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6936609745025635, + "num_tokens": 93742099.0, + "step": 3622 + }, + { + "epoch": 0.39786953656929497, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1210193634033203, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7240947484970093, + "num_tokens": 93767255.0, + "step": 3623 + }, + { + "epoch": 0.3979793542719086, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5795369148254395, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7063030004501343, + "num_tokens": 93787179.0, + "step": 3624 + }, + { + "epoch": 0.3980891719745223, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9882965087890625, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7190505266189575, + "num_tokens": 93818656.0, + "step": 3625 + }, + { + "epoch": 0.39819898967713596, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2959768772125244, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6981693506240845, + "num_tokens": 93842689.0, + "step": 3626 + }, + { + "epoch": 0.3983088073797496, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.370976686477661, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7031807899475098, + "num_tokens": 93866051.0, + "step": 3627 + }, + { + "epoch": 0.39841862508236325, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9927778244018555, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7112385034561157, + "num_tokens": 93894339.0, + "step": 3628 + }, + { + "epoch": 0.39852844278497696, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4642622470855713, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.6870768070220947, + "num_tokens": 93917128.0, + "step": 3629 + }, + { + "epoch": 0.3986382604875906, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4517879486083984, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7204371094703674, + "num_tokens": 93937068.0, + "step": 3630 + }, + { + "epoch": 0.39874807819020425, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3384268283843994, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7081590294837952, + "num_tokens": 93961825.0, + "step": 3631 + }, + { + "epoch": 0.39885789589281795, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4373114109039307, + "learning_rate": 1e-06, + "loss": 1.0792, + "mean_token_accuracy": 0.6743800640106201, + "num_tokens": 93986820.0, + "step": 3632 + }, + { + "epoch": 0.3989677135954316, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.472175121307373, + "learning_rate": 1e-06, + "loss": 1.0625, + "mean_token_accuracy": 0.6852244138717651, + "num_tokens": 94009183.0, + "step": 3633 + }, + { + "epoch": 0.39907753129804524, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2417831420898438, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7072494029998779, + "num_tokens": 94035169.0, + "step": 3634 + }, + { + "epoch": 0.3991873490006589, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1770191192626953, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7012941241264343, + "num_tokens": 94058915.0, + "step": 3635 + }, + { + "epoch": 0.3992971667032726, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0981974601745605, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7049908638000488, + "num_tokens": 94087673.0, + "step": 3636 + }, + { + "epoch": 0.39940698440588623, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0398483276367188, + "learning_rate": 1e-06, + "loss": 1.0745, + "mean_token_accuracy": 0.6792251467704773, + "num_tokens": 94117111.0, + "step": 3637 + }, + { + "epoch": 0.3995168021084999, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.344498872756958, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7128553986549377, + "num_tokens": 94142061.0, + "step": 3638 + }, + { + "epoch": 0.3996266198111135, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2611355781555176, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7253462076187134, + "num_tokens": 94166473.0, + "step": 3639 + }, + { + "epoch": 0.3997364375137272, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9514058828353882, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7217276096343994, + "num_tokens": 94195339.0, + "step": 3640 + }, + { + "epoch": 0.3998462552163409, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.218106508255005, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.705672025680542, + "num_tokens": 94220333.0, + "step": 3641 + }, + { + "epoch": 0.3999560729189545, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.043642520904541, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7015026211738586, + "num_tokens": 94247191.0, + "step": 3642 + }, + { + "epoch": 0.4000658906215682, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2757961750030518, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.680901050567627, + "num_tokens": 94272074.0, + "step": 3643 + }, + { + "epoch": 0.40017570832418187, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4499757289886475, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.723806619644165, + "num_tokens": 94292812.0, + "step": 3644 + }, + { + "epoch": 0.4002855260267955, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.440232515335083, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6939444541931152, + "num_tokens": 94317623.0, + "step": 3645 + }, + { + "epoch": 0.40039534372940916, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.067803382873535, + "learning_rate": 1e-06, + "loss": 1.0828, + "mean_token_accuracy": 0.692186713218689, + "num_tokens": 94348024.0, + "step": 3646 + }, + { + "epoch": 0.40050516143202286, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.403557300567627, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6960696578025818, + "num_tokens": 94370827.0, + "step": 3647 + }, + { + "epoch": 0.4006149791346365, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4632725715637207, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6871198415756226, + "num_tokens": 94395141.0, + "step": 3648 + }, + { + "epoch": 0.40072479683725015, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9282288551330566, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.688060998916626, + "num_tokens": 94427195.0, + "step": 3649 + }, + { + "epoch": 0.4008346145398638, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.084199905395508, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6852614879608154, + "num_tokens": 94456986.0, + "step": 3650 + }, + { + "epoch": 0.4009444322424775, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.167414426803589, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6938400268554688, + "num_tokens": 94486412.0, + "step": 3651 + }, + { + "epoch": 0.40105424994509115, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1890451908111572, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7015132308006287, + "num_tokens": 94513341.0, + "step": 3652 + }, + { + "epoch": 0.4011640676477048, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3011600971221924, + "learning_rate": 1e-06, + "loss": 1.0729, + "mean_token_accuracy": 0.6919276714324951, + "num_tokens": 94539325.0, + "step": 3653 + }, + { + "epoch": 0.4012738853503185, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1017069816589355, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6913229823112488, + "num_tokens": 94569341.0, + "step": 3654 + }, + { + "epoch": 0.40138370305293214, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4003655910491943, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7044911980628967, + "num_tokens": 94590125.0, + "step": 3655 + }, + { + "epoch": 0.4014935207555458, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2108154296875, + "learning_rate": 1e-06, + "loss": 1.0364, + "mean_token_accuracy": 0.6876410245895386, + "num_tokens": 94618712.0, + "step": 3656 + }, + { + "epoch": 0.40160333845815943, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.331192970275879, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7216428518295288, + "num_tokens": 94639634.0, + "step": 3657 + }, + { + "epoch": 0.40171315616077313, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2666473388671875, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.6827995181083679, + "num_tokens": 94663874.0, + "step": 3658 + }, + { + "epoch": 0.4018229738633868, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1446533203125, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6870648264884949, + "num_tokens": 94692854.0, + "step": 3659 + }, + { + "epoch": 0.4019327915660004, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.631592035293579, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7073495388031006, + "num_tokens": 94711951.0, + "step": 3660 + }, + { + "epoch": 0.4020426092686141, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2781574726104736, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6987959146499634, + "num_tokens": 94738139.0, + "step": 3661 + }, + { + "epoch": 0.40215242697122777, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.084970712661743, + "learning_rate": 1e-06, + "loss": 1.116, + "mean_token_accuracy": 0.6690732836723328, + "num_tokens": 94769276.0, + "step": 3662 + }, + { + "epoch": 0.4022622446738414, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.336212635040283, + "learning_rate": 1e-06, + "loss": 1.1068, + "mean_token_accuracy": 0.6734224557876587, + "num_tokens": 94793987.0, + "step": 3663 + }, + { + "epoch": 0.40237206237645506, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.399181842803955, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7068522572517395, + "num_tokens": 94815453.0, + "step": 3664 + }, + { + "epoch": 0.40248188007906877, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.003418445587158, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.716069221496582, + "num_tokens": 94842720.0, + "step": 3665 + }, + { + "epoch": 0.4025916977816824, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2629737854003906, + "learning_rate": 1e-06, + "loss": 1.0252, + "mean_token_accuracy": 0.6901212930679321, + "num_tokens": 94869987.0, + "step": 3666 + }, + { + "epoch": 0.40270151548429606, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.569007635116577, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7023345232009888, + "num_tokens": 94891386.0, + "step": 3667 + }, + { + "epoch": 0.4028113331869097, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.321321487426758, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.696883499622345, + "num_tokens": 94915405.0, + "step": 3668 + }, + { + "epoch": 0.4029211508895234, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2480525970458984, + "learning_rate": 1e-06, + "loss": 1.0433, + "mean_token_accuracy": 0.6996535062789917, + "num_tokens": 94940529.0, + "step": 3669 + }, + { + "epoch": 0.40303096859213705, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3172318935394287, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7048890590667725, + "num_tokens": 94963891.0, + "step": 3670 + }, + { + "epoch": 0.4031407862947507, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3725063800811768, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.6958924531936646, + "num_tokens": 94986480.0, + "step": 3671 + }, + { + "epoch": 0.4032506039973644, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2569172382354736, + "learning_rate": 1e-06, + "loss": 1.0459, + "mean_token_accuracy": 0.6838442087173462, + "num_tokens": 95012838.0, + "step": 3672 + }, + { + "epoch": 0.40336042169997804, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3919970989227295, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6960920095443726, + "num_tokens": 95034266.0, + "step": 3673 + }, + { + "epoch": 0.4034702394025917, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.453646659851074, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.689838171005249, + "num_tokens": 95055943.0, + "step": 3674 + }, + { + "epoch": 0.40358005710520534, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1949706077575684, + "learning_rate": 1e-06, + "loss": 1.0941, + "mean_token_accuracy": 0.6807938814163208, + "num_tokens": 95083037.0, + "step": 3675 + }, + { + "epoch": 0.40368987480781904, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.253018379211426, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6901596784591675, + "num_tokens": 95109674.0, + "step": 3676 + }, + { + "epoch": 0.4037996925104327, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.594181776046753, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7030155658721924, + "num_tokens": 95128170.0, + "step": 3677 + }, + { + "epoch": 0.40390951021304633, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0694527626037598, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6921796798706055, + "num_tokens": 95158536.0, + "step": 3678 + }, + { + "epoch": 0.40401932791566003, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9500761032104492, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.7059171199798584, + "num_tokens": 95188713.0, + "step": 3679 + }, + { + "epoch": 0.4041291456182737, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.329883575439453, + "learning_rate": 1e-06, + "loss": 1.0832, + "mean_token_accuracy": 0.6736583709716797, + "num_tokens": 95214405.0, + "step": 3680 + }, + { + "epoch": 0.4042389633208873, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9673532247543335, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7204892039299011, + "num_tokens": 95244165.0, + "step": 3681 + }, + { + "epoch": 0.40434878102350097, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.235403299331665, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.714987576007843, + "num_tokens": 95267477.0, + "step": 3682 + }, + { + "epoch": 0.40445859872611467, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1511189937591553, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6890313029289246, + "num_tokens": 95295044.0, + "step": 3683 + }, + { + "epoch": 0.4045684164287283, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3731977939605713, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.6982256174087524, + "num_tokens": 95319298.0, + "step": 3684 + }, + { + "epoch": 0.40467823413134196, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2078731060028076, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7115451097488403, + "num_tokens": 95344447.0, + "step": 3685 + }, + { + "epoch": 0.4047880518339556, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5378971099853516, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7030568718910217, + "num_tokens": 95364424.0, + "step": 3686 + }, + { + "epoch": 0.4048978695365693, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4333269596099854, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6934439539909363, + "num_tokens": 95387166.0, + "step": 3687 + }, + { + "epoch": 0.40500768723918296, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1699161529541016, + "learning_rate": 1e-06, + "loss": 1.0286, + "mean_token_accuracy": 0.6920262575149536, + "num_tokens": 95414783.0, + "step": 3688 + }, + { + "epoch": 0.4051175049417966, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2545135021209717, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6992532014846802, + "num_tokens": 95440817.0, + "step": 3689 + }, + { + "epoch": 0.4052273226444103, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.405317544937134, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.703123927116394, + "num_tokens": 95463800.0, + "step": 3690 + }, + { + "epoch": 0.40533714034702395, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9987508058547974, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7094130516052246, + "num_tokens": 95493112.0, + "step": 3691 + }, + { + "epoch": 0.4054469580496376, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1402699947357178, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7070710062980652, + "num_tokens": 95519677.0, + "step": 3692 + }, + { + "epoch": 0.40555677575225124, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.026960849761963, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.695313036441803, + "num_tokens": 95549291.0, + "step": 3693 + }, + { + "epoch": 0.40566659345486494, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2093260288238525, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7170318365097046, + "num_tokens": 95575323.0, + "step": 3694 + }, + { + "epoch": 0.4057764111574786, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.176819324493408, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7106120586395264, + "num_tokens": 95601929.0, + "step": 3695 + }, + { + "epoch": 0.40588622886009224, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0282247066497803, + "learning_rate": 1e-06, + "loss": 1.0721, + "mean_token_accuracy": 0.6836768388748169, + "num_tokens": 95632256.0, + "step": 3696 + }, + { + "epoch": 0.4059960465627059, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2101476192474365, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7120361328125, + "num_tokens": 95657228.0, + "step": 3697 + }, + { + "epoch": 0.4061058642653196, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2466135025024414, + "learning_rate": 1e-06, + "loss": 1.0708, + "mean_token_accuracy": 0.6811386346817017, + "num_tokens": 95683329.0, + "step": 3698 + }, + { + "epoch": 0.40621568196793323, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0990331172943115, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.6928564310073853, + "num_tokens": 95712366.0, + "step": 3699 + }, + { + "epoch": 0.4063254996705469, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.011587381362915, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.6922030448913574, + "num_tokens": 95741651.0, + "step": 3700 + }, + { + "epoch": 0.4064353173731606, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9852185249328613, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.7022972106933594, + "num_tokens": 95773832.0, + "step": 3701 + }, + { + "epoch": 0.4065451350757742, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.166964054107666, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.7020366787910461, + "num_tokens": 95801416.0, + "step": 3702 + }, + { + "epoch": 0.40665495277838787, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.265202283859253, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7020190954208374, + "num_tokens": 95826058.0, + "step": 3703 + }, + { + "epoch": 0.4067647704810015, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9323328733444214, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6913357973098755, + "num_tokens": 95859076.0, + "step": 3704 + }, + { + "epoch": 0.4068745881836152, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2654075622558594, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7036288380622864, + "num_tokens": 95884058.0, + "step": 3705 + }, + { + "epoch": 0.40698440588622886, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4371910095214844, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.714911937713623, + "num_tokens": 95906494.0, + "step": 3706 + }, + { + "epoch": 0.4070942235888425, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.098628044128418, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6904289126396179, + "num_tokens": 95933524.0, + "step": 3707 + }, + { + "epoch": 0.4072040412914562, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2649104595184326, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7081891298294067, + "num_tokens": 95957723.0, + "step": 3708 + }, + { + "epoch": 0.40731385899406986, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3361451625823975, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7004191875457764, + "num_tokens": 95980974.0, + "step": 3709 + }, + { + "epoch": 0.4074236766966835, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0692765712738037, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6990758180618286, + "num_tokens": 96010658.0, + "step": 3710 + }, + { + "epoch": 0.40753349439929715, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2683730125427246, + "learning_rate": 1e-06, + "loss": 1.044, + "mean_token_accuracy": 0.6815443634986877, + "num_tokens": 96036358.0, + "step": 3711 + }, + { + "epoch": 0.40764331210191085, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.151453733444214, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.6960209608078003, + "num_tokens": 96062798.0, + "step": 3712 + }, + { + "epoch": 0.4077531298045245, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.139084815979004, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6911068558692932, + "num_tokens": 96089971.0, + "step": 3713 + }, + { + "epoch": 0.40786294750713814, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.166639566421509, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.688400149345398, + "num_tokens": 96116747.0, + "step": 3714 + }, + { + "epoch": 0.4079727652097518, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2550034523010254, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.6877336502075195, + "num_tokens": 96142612.0, + "step": 3715 + }, + { + "epoch": 0.4080825829123655, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.406895399093628, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6933209300041199, + "num_tokens": 96167683.0, + "step": 3716 + }, + { + "epoch": 0.40819240061497913, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1038191318511963, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7126270532608032, + "num_tokens": 96195221.0, + "step": 3717 + }, + { + "epoch": 0.4083022183175928, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.144055128097534, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7145735025405884, + "num_tokens": 96221480.0, + "step": 3718 + }, + { + "epoch": 0.4084120360202065, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5043563842773438, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6993461847305298, + "num_tokens": 96242839.0, + "step": 3719 + }, + { + "epoch": 0.4085218537228201, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5241425037384033, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6921460628509521, + "num_tokens": 96264354.0, + "step": 3720 + }, + { + "epoch": 0.4086316714254338, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0640275478363037, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7167946100234985, + "num_tokens": 96292974.0, + "step": 3721 + }, + { + "epoch": 0.4087414891280474, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3387269973754883, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7060821056365967, + "num_tokens": 96317311.0, + "step": 3722 + }, + { + "epoch": 0.4088513068306611, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.192427396774292, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6778292655944824, + "num_tokens": 96344098.0, + "step": 3723 + }, + { + "epoch": 0.40896112453327477, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2723238468170166, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6994686722755432, + "num_tokens": 96368361.0, + "step": 3724 + }, + { + "epoch": 0.4090709422358884, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.140249729156494, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7114088535308838, + "num_tokens": 96394469.0, + "step": 3725 + }, + { + "epoch": 0.40918075993850206, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.22225022315979, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6915214657783508, + "num_tokens": 96419691.0, + "step": 3726 + }, + { + "epoch": 0.40929057764111576, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9247139692306519, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6992979645729065, + "num_tokens": 96453773.0, + "step": 3727 + }, + { + "epoch": 0.4094003953437294, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1961443424224854, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7203023433685303, + "num_tokens": 96478387.0, + "step": 3728 + }, + { + "epoch": 0.40951021304634305, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1932380199432373, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6967803239822388, + "num_tokens": 96504213.0, + "step": 3729 + }, + { + "epoch": 0.40962003074895675, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9644742012023926, + "learning_rate": 1e-06, + "loss": 1.0514, + "mean_token_accuracy": 0.6874935626983643, + "num_tokens": 96536107.0, + "step": 3730 + }, + { + "epoch": 0.4097298484515704, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2020761966705322, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.7104641795158386, + "num_tokens": 96561775.0, + "step": 3731 + }, + { + "epoch": 0.40983966615418405, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3470680713653564, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.719113290309906, + "num_tokens": 96584036.0, + "step": 3732 + }, + { + "epoch": 0.4099494838567977, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4639103412628174, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7165132164955139, + "num_tokens": 96605831.0, + "step": 3733 + }, + { + "epoch": 0.4100593015594114, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.459918737411499, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.704829216003418, + "num_tokens": 96627335.0, + "step": 3734 + }, + { + "epoch": 0.41016911926202504, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.180659055709839, + "learning_rate": 1e-06, + "loss": 1.0706, + "mean_token_accuracy": 0.6905312538146973, + "num_tokens": 96655776.0, + "step": 3735 + }, + { + "epoch": 0.4102789369646387, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0086781978607178, + "learning_rate": 1e-06, + "loss": 1.0901, + "mean_token_accuracy": 0.6856579184532166, + "num_tokens": 96686329.0, + "step": 3736 + }, + { + "epoch": 0.4103887546672524, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9229868650436401, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7122128009796143, + "num_tokens": 96717402.0, + "step": 3737 + }, + { + "epoch": 0.41049857236986603, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0937061309814453, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7147511839866638, + "num_tokens": 96743834.0, + "step": 3738 + }, + { + "epoch": 0.4106083900724797, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2865097522735596, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.6977702379226685, + "num_tokens": 96766567.0, + "step": 3739 + }, + { + "epoch": 0.4107182077750933, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1928493976593018, + "learning_rate": 1e-06, + "loss": 1.0375, + "mean_token_accuracy": 0.6940335035324097, + "num_tokens": 96793264.0, + "step": 3740 + }, + { + "epoch": 0.410828025477707, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.022218942642212, + "learning_rate": 1e-06, + "loss": 1.0439, + "mean_token_accuracy": 0.692162036895752, + "num_tokens": 96825333.0, + "step": 3741 + }, + { + "epoch": 0.41093784318032067, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0395188331604004, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7213010787963867, + "num_tokens": 96853811.0, + "step": 3742 + }, + { + "epoch": 0.4110476608829343, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2705886363983154, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7014756202697754, + "num_tokens": 96878822.0, + "step": 3743 + }, + { + "epoch": 0.41115747858554796, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4106943607330322, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7288172245025635, + "num_tokens": 96901200.0, + "step": 3744 + }, + { + "epoch": 0.41126729628816167, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.549534320831299, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7025210857391357, + "num_tokens": 96920484.0, + "step": 3745 + }, + { + "epoch": 0.4113771139907753, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2765934467315674, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7041853666305542, + "num_tokens": 96944095.0, + "step": 3746 + }, + { + "epoch": 0.41148693169338896, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5156872272491455, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7024403214454651, + "num_tokens": 96964442.0, + "step": 3747 + }, + { + "epoch": 0.41159674939600266, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.0535407066345215, + "learning_rate": 1e-06, + "loss": 1.0501, + "mean_token_accuracy": 0.6881483197212219, + "num_tokens": 96994310.0, + "step": 3748 + }, + { + "epoch": 0.4117065670986163, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4950060844421387, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.715207040309906, + "num_tokens": 97015086.0, + "step": 3749 + }, + { + "epoch": 0.41181638480122995, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.4733264446258545, + "learning_rate": 1e-06, + "loss": 1.0681, + "mean_token_accuracy": 0.685388445854187, + "num_tokens": 97038406.0, + "step": 3750 + }, + { + "epoch": 0.4119262025038436, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2628660202026367, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6913652420043945, + "num_tokens": 97063504.0, + "step": 3751 + }, + { + "epoch": 0.4120360202064573, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.472092390060425, + "learning_rate": 1e-06, + "loss": 1.054, + "mean_token_accuracy": 0.6790432929992676, + "num_tokens": 97087001.0, + "step": 3752 + }, + { + "epoch": 0.41214583790907094, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.055402994155884, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7116641402244568, + "num_tokens": 97114967.0, + "step": 3753 + }, + { + "epoch": 0.4122556556116846, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.365368127822876, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7211164832115173, + "num_tokens": 97137484.0, + "step": 3754 + }, + { + "epoch": 0.4123654733142983, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1193349361419678, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7036134004592896, + "num_tokens": 97164287.0, + "step": 3755 + }, + { + "epoch": 0.41247529101691194, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.497849464416504, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7238180041313171, + "num_tokens": 97185153.0, + "step": 3756 + }, + { + "epoch": 0.4125851087195256, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.173333168029785, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6968449354171753, + "num_tokens": 97211905.0, + "step": 3757 + }, + { + "epoch": 0.41269492642213923, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1706736087799072, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.6991881132125854, + "num_tokens": 97238603.0, + "step": 3758 + }, + { + "epoch": 0.41280474412475293, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.8979778289794922, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6787805557250977, + "num_tokens": 97271073.0, + "step": 3759 + }, + { + "epoch": 0.4129145618273666, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2102303504943848, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.700951337814331, + "num_tokens": 97297350.0, + "step": 3760 + }, + { + "epoch": 0.4130243795299802, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.230527639389038, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7040603756904602, + "num_tokens": 97321488.0, + "step": 3761 + }, + { + "epoch": 0.41313419723259387, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1270194053649902, + "learning_rate": 1e-06, + "loss": 1.1053, + "mean_token_accuracy": 0.680962324142456, + "num_tokens": 97349133.0, + "step": 3762 + }, + { + "epoch": 0.41324401493520757, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2630317211151123, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6991115808486938, + "num_tokens": 97373337.0, + "step": 3763 + }, + { + "epoch": 0.4133538326378212, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5313239097595215, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7334216833114624, + "num_tokens": 97393305.0, + "step": 3764 + }, + { + "epoch": 0.41346365034043486, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.181405544281006, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7006290555000305, + "num_tokens": 97420834.0, + "step": 3765 + }, + { + "epoch": 0.41357346804304856, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2309110164642334, + "learning_rate": 1e-06, + "loss": 1.014, + "mean_token_accuracy": 0.696706235408783, + "num_tokens": 97447875.0, + "step": 3766 + }, + { + "epoch": 0.4136832857456622, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.438664197921753, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.717957615852356, + "num_tokens": 97468715.0, + "step": 3767 + }, + { + "epoch": 0.41379310344827586, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.418804407119751, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.699292778968811, + "num_tokens": 97491436.0, + "step": 3768 + }, + { + "epoch": 0.4139029211508895, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.066047430038452, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6924442648887634, + "num_tokens": 97521771.0, + "step": 3769 + }, + { + "epoch": 0.4140127388535032, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.194558620452881, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6970529556274414, + "num_tokens": 97548123.0, + "step": 3770 + }, + { + "epoch": 0.41412255655611685, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3694376945495605, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7205997705459595, + "num_tokens": 97569764.0, + "step": 3771 + }, + { + "epoch": 0.4142323742587305, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.1012375354766846, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.6985759139060974, + "num_tokens": 97598522.0, + "step": 3772 + }, + { + "epoch": 0.41434219196134414, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2566075325012207, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7351466417312622, + "num_tokens": 97623658.0, + "step": 3773 + }, + { + "epoch": 0.41445200966395784, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 1.9730994701385498, + "learning_rate": 1e-06, + "loss": 1.0584, + "mean_token_accuracy": 0.6888983845710754, + "num_tokens": 97654071.0, + "step": 3774 + }, + { + "epoch": 0.4145618273665715, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3510074615478516, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6958547234535217, + "num_tokens": 97677831.0, + "step": 3775 + }, + { + "epoch": 0.41467164506918514, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.365203857421875, + "learning_rate": 1e-06, + "loss": 1.0459, + "mean_token_accuracy": 0.6858949661254883, + "num_tokens": 97702503.0, + "step": 3776 + }, + { + "epoch": 0.41478146277179884, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.436246156692505, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6894078254699707, + "num_tokens": 97724802.0, + "step": 3777 + }, + { + "epoch": 0.4148912804744125, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.344130516052246, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.712938666343689, + "num_tokens": 97749234.0, + "step": 3778 + }, + { + "epoch": 0.41500109817702613, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.341456651687622, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7126307487487793, + "num_tokens": 97770698.0, + "step": 3779 + }, + { + "epoch": 0.4151109158796398, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.255913019180298, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.7002893686294556, + "num_tokens": 97794172.0, + "step": 3780 + }, + { + "epoch": 0.4152207335822535, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.45377516746521, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.7022662162780762, + "num_tokens": 97815066.0, + "step": 3781 + }, + { + "epoch": 0.4153305512848671, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.10443377494812, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7269622683525085, + "num_tokens": 97841199.0, + "step": 3782 + }, + { + "epoch": 0.41544036898748077, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.5010769367218018, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6928003430366516, + "num_tokens": 97865472.0, + "step": 3783 + }, + { + "epoch": 0.41555018669009447, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.3939507007598877, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7130417823791504, + "num_tokens": 97886053.0, + "step": 3784 + }, + { + "epoch": 0.4156600043927081, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.6011414527893066, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7041133642196655, + "num_tokens": 97908195.0, + "step": 3785 + }, + { + "epoch": 0.41576982209532176, + "ewc_loss": 3.218650817871094e-05, + "grad_norm": 2.2520647048950195, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7077429294586182, + "num_tokens": 97935081.0, + "step": 3786 + }, + { + "epoch": 0.4158796397979354, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 1.9929847717285156, + "learning_rate": 1e-06, + "loss": 1.0486, + "mean_token_accuracy": 0.6794606447219849, + "num_tokens": 97966325.0, + "step": 3787 + }, + { + "epoch": 0.4159894575005491, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.3883776664733887, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6996141672134399, + "num_tokens": 97991457.0, + "step": 3788 + }, + { + "epoch": 0.41609927520316276, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.2042574882507324, + "learning_rate": 1e-06, + "loss": 1.0444, + "mean_token_accuracy": 0.6871472001075745, + "num_tokens": 98018341.0, + "step": 3789 + }, + { + "epoch": 0.4162090929057764, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.129802703857422, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7085773348808289, + "num_tokens": 98045098.0, + "step": 3790 + }, + { + "epoch": 0.41631891060839005, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.3296191692352295, + "learning_rate": 1e-06, + "loss": 1.0427, + "mean_token_accuracy": 0.6844907999038696, + "num_tokens": 98069357.0, + "step": 3791 + }, + { + "epoch": 0.41642872831100375, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.3248870372772217, + "learning_rate": 1e-06, + "loss": 1.0712, + "mean_token_accuracy": 0.6815646886825562, + "num_tokens": 98094848.0, + "step": 3792 + }, + { + "epoch": 0.4165385460136174, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.3909220695495605, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6858272552490234, + "num_tokens": 98118179.0, + "step": 3793 + }, + { + "epoch": 0.41664836371623104, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.303018093109131, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.69706130027771, + "num_tokens": 98142211.0, + "step": 3794 + }, + { + "epoch": 0.41675818141884474, + "ewc_loss": 3.24249267578125e-05, + "grad_norm": 2.2625396251678467, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6957385540008545, + "num_tokens": 98168425.0, + "step": 3795 + }, + { + "epoch": 0.4168679991214584, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.4425854682922363, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7139147520065308, + "num_tokens": 98191140.0, + "step": 3796 + }, + { + "epoch": 0.41697781682407203, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.0228214263916016, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7118602991104126, + "num_tokens": 98221268.0, + "step": 3797 + }, + { + "epoch": 0.4170876345266857, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.401278018951416, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7115724086761475, + "num_tokens": 98243642.0, + "step": 3798 + }, + { + "epoch": 0.4171974522292994, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.3280980587005615, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7315441370010376, + "num_tokens": 98265808.0, + "step": 3799 + }, + { + "epoch": 0.417307269931913, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.0694174766540527, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7110799551010132, + "num_tokens": 98293589.0, + "step": 3800 + }, + { + "epoch": 0.4174170876345267, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.1525816917419434, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6957432627677917, + "num_tokens": 98321281.0, + "step": 3801 + }, + { + "epoch": 0.4175269053371403, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.1860718727111816, + "learning_rate": 1e-06, + "loss": 1.0742, + "mean_token_accuracy": 0.6784920692443848, + "num_tokens": 98347302.0, + "step": 3802 + }, + { + "epoch": 0.417636723039754, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.181870937347412, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6769418716430664, + "num_tokens": 98375310.0, + "step": 3803 + }, + { + "epoch": 0.41774654074236767, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 1.915300726890564, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.716178297996521, + "num_tokens": 98405962.0, + "step": 3804 + }, + { + "epoch": 0.4178563584449813, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.2647697925567627, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.734574556350708, + "num_tokens": 98429344.0, + "step": 3805 + }, + { + "epoch": 0.417966176147595, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.2159881591796875, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7022337913513184, + "num_tokens": 98455018.0, + "step": 3806 + }, + { + "epoch": 0.41807599385020866, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.080695390701294, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.7081925272941589, + "num_tokens": 98484221.0, + "step": 3807 + }, + { + "epoch": 0.4181858115528223, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 1.9950095415115356, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6961613297462463, + "num_tokens": 98515513.0, + "step": 3808 + }, + { + "epoch": 0.41829562925543595, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.4402878284454346, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7152677774429321, + "num_tokens": 98537078.0, + "step": 3809 + }, + { + "epoch": 0.41840544695804965, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.509437084197998, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7061318159103394, + "num_tokens": 98556703.0, + "step": 3810 + }, + { + "epoch": 0.4185152646606633, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.0934152603149414, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6900064945220947, + "num_tokens": 98586006.0, + "step": 3811 + }, + { + "epoch": 0.41862508236327695, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.345019578933716, + "learning_rate": 1e-06, + "loss": 1.061, + "mean_token_accuracy": 0.6842256784439087, + "num_tokens": 98609590.0, + "step": 3812 + }, + { + "epoch": 0.41873490006589065, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.5876550674438477, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7050259113311768, + "num_tokens": 98632955.0, + "step": 3813 + }, + { + "epoch": 0.4188447177685043, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.5007710456848145, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.6991441249847412, + "num_tokens": 98654276.0, + "step": 3814 + }, + { + "epoch": 0.41895453547111794, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.4350178241729736, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.7017624378204346, + "num_tokens": 98676588.0, + "step": 3815 + }, + { + "epoch": 0.4190643531737316, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.0865063667297363, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6933125853538513, + "num_tokens": 98704836.0, + "step": 3816 + }, + { + "epoch": 0.4191741708763453, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.1225078105926514, + "learning_rate": 1e-06, + "loss": 1.102, + "mean_token_accuracy": 0.6707767844200134, + "num_tokens": 98736500.0, + "step": 3817 + }, + { + "epoch": 0.41928398857895893, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.539705514907837, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.698712944984436, + "num_tokens": 98758217.0, + "step": 3818 + }, + { + "epoch": 0.4193938062815726, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.208134174346924, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.6799926161766052, + "num_tokens": 98782642.0, + "step": 3819 + }, + { + "epoch": 0.4195036239841862, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.1380250453948975, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6998016834259033, + "num_tokens": 98809715.0, + "step": 3820 + }, + { + "epoch": 0.4196134416867999, + "ewc_loss": 3.266334533691406e-05, + "grad_norm": 2.3215880393981934, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7082202434539795, + "num_tokens": 98832396.0, + "step": 3821 + }, + { + "epoch": 0.41972325938941357, + "ewc_loss": 3.2901763916015625e-05, + "grad_norm": 2.3864023685455322, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.7069352865219116, + "num_tokens": 98855032.0, + "step": 3822 + }, + { + "epoch": 0.4198330770920272, + "ewc_loss": 3.2901763916015625e-05, + "grad_norm": 2.0385282039642334, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6839783191680908, + "num_tokens": 98885502.0, + "step": 3823 + }, + { + "epoch": 0.4199428947946409, + "ewc_loss": 3.2901763916015625e-05, + "grad_norm": 2.2830982208251953, + "learning_rate": 1e-06, + "loss": 1.0759, + "mean_token_accuracy": 0.6825591921806335, + "num_tokens": 98910833.0, + "step": 3824 + }, + { + "epoch": 0.42005271249725457, + "ewc_loss": 3.2901763916015625e-05, + "grad_norm": 2.1461424827575684, + "learning_rate": 1e-06, + "loss": 1.0757, + "mean_token_accuracy": 0.67815101146698, + "num_tokens": 98939922.0, + "step": 3825 + }, + { + "epoch": 0.4201625301998682, + "ewc_loss": 3.2901763916015625e-05, + "grad_norm": 2.5009381771087646, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.697890043258667, + "num_tokens": 98966447.0, + "step": 3826 + }, + { + "epoch": 0.42027234790248186, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.2074172496795654, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.702206015586853, + "num_tokens": 98990069.0, + "step": 3827 + }, + { + "epoch": 0.42038216560509556, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.0055153369903564, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7122116088867188, + "num_tokens": 99018779.0, + "step": 3828 + }, + { + "epoch": 0.4204919833077092, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.149510145187378, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6843990087509155, + "num_tokens": 99046989.0, + "step": 3829 + }, + { + "epoch": 0.42060180101032285, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.207829713821411, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6910614967346191, + "num_tokens": 99075830.0, + "step": 3830 + }, + { + "epoch": 0.42071161871293655, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.220247507095337, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6927893161773682, + "num_tokens": 99104043.0, + "step": 3831 + }, + { + "epoch": 0.4208214364155502, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.0723214149475098, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6911184787750244, + "num_tokens": 99134589.0, + "step": 3832 + }, + { + "epoch": 0.42093125411816384, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.319716453552246, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6881811618804932, + "num_tokens": 99159009.0, + "step": 3833 + }, + { + "epoch": 0.4210410718207775, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.1203668117523193, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6872402429580688, + "num_tokens": 99188690.0, + "step": 3834 + }, + { + "epoch": 0.4211508895233912, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.4251809120178223, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7199928760528564, + "num_tokens": 99209626.0, + "step": 3835 + }, + { + "epoch": 0.42126070722600484, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.154550552368164, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6924425363540649, + "num_tokens": 99237195.0, + "step": 3836 + }, + { + "epoch": 0.4213705249286185, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.115199565887451, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7174345254898071, + "num_tokens": 99265505.0, + "step": 3837 + }, + { + "epoch": 0.42148034263123213, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.172778606414795, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7180519104003906, + "num_tokens": 99290547.0, + "step": 3838 + }, + { + "epoch": 0.42159016033384583, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.36641001701355, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7039057016372681, + "num_tokens": 99313701.0, + "step": 3839 + }, + { + "epoch": 0.4216999780364595, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.1305861473083496, + "learning_rate": 1e-06, + "loss": 1.0783, + "mean_token_accuracy": 0.6776902079582214, + "num_tokens": 99344109.0, + "step": 3840 + }, + { + "epoch": 0.4218097957390731, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.215162992477417, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7072422504425049, + "num_tokens": 99370893.0, + "step": 3841 + }, + { + "epoch": 0.4219196134416868, + "ewc_loss": 3.314018249511719e-05, + "grad_norm": 2.2887983322143555, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7020872831344604, + "num_tokens": 99393832.0, + "step": 3842 + }, + { + "epoch": 0.42202943114430047, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.380906820297241, + "learning_rate": 1e-06, + "loss": 1.1195, + "mean_token_accuracy": 0.667231559753418, + "num_tokens": 99418060.0, + "step": 3843 + }, + { + "epoch": 0.4221392488469141, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.3400485515594482, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7109548449516296, + "num_tokens": 99441684.0, + "step": 3844 + }, + { + "epoch": 0.42224906654952776, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.1265780925750732, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7030354142189026, + "num_tokens": 99469380.0, + "step": 3845 + }, + { + "epoch": 0.42235888425214146, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.7922732830047607, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7288722991943359, + "num_tokens": 99487015.0, + "step": 3846 + }, + { + "epoch": 0.4224687019547551, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.2572338581085205, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6892129182815552, + "num_tokens": 99513084.0, + "step": 3847 + }, + { + "epoch": 0.42257851965736876, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.2948570251464844, + "learning_rate": 1e-06, + "loss": 1.0386, + "mean_token_accuracy": 0.6908326148986816, + "num_tokens": 99538790.0, + "step": 3848 + }, + { + "epoch": 0.4226883373599824, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.1965839862823486, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7109509110450745, + "num_tokens": 99563439.0, + "step": 3849 + }, + { + "epoch": 0.4227981550625961, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.2893993854522705, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7254644632339478, + "num_tokens": 99586655.0, + "step": 3850 + }, + { + "epoch": 0.42290797276520975, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.256964683532715, + "learning_rate": 1e-06, + "loss": 1.0428, + "mean_token_accuracy": 0.6860439777374268, + "num_tokens": 99611237.0, + "step": 3851 + }, + { + "epoch": 0.4230177904678234, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 1.9115315675735474, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6944204568862915, + "num_tokens": 99645810.0, + "step": 3852 + }, + { + "epoch": 0.4231276081704371, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.655937433242798, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.699508011341095, + "num_tokens": 99664956.0, + "step": 3853 + }, + { + "epoch": 0.42323742587305074, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.047305107116699, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7180183529853821, + "num_tokens": 99693325.0, + "step": 3854 + }, + { + "epoch": 0.4233472435756644, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.436603307723999, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7183563709259033, + "num_tokens": 99715867.0, + "step": 3855 + }, + { + "epoch": 0.42345706127827804, + "ewc_loss": 3.337860107421875e-05, + "grad_norm": 2.2162256240844727, + "learning_rate": 1e-06, + "loss": 1.1024, + "mean_token_accuracy": 0.6730474233627319, + "num_tokens": 99745782.0, + "step": 3856 + }, + { + "epoch": 0.42356687898089174, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.365112066268921, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6837989687919617, + "num_tokens": 99768440.0, + "step": 3857 + }, + { + "epoch": 0.4236766966835054, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.5053770542144775, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6941306591033936, + "num_tokens": 99789971.0, + "step": 3858 + }, + { + "epoch": 0.42378651438611903, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3934624195098877, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6995861530303955, + "num_tokens": 99813570.0, + "step": 3859 + }, + { + "epoch": 0.42389633208873273, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.391977548599243, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6873598098754883, + "num_tokens": 99837657.0, + "step": 3860 + }, + { + "epoch": 0.4240061497913464, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.0967087745666504, + "learning_rate": 1e-06, + "loss": 1.1153, + "mean_token_accuracy": 0.6661520004272461, + "num_tokens": 99868728.0, + "step": 3861 + }, + { + "epoch": 0.42411596749396, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.210777521133423, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.6825281381607056, + "num_tokens": 99896969.0, + "step": 3862 + }, + { + "epoch": 0.42422578519657367, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.372251033782959, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.709784209728241, + "num_tokens": 99919296.0, + "step": 3863 + }, + { + "epoch": 0.42433560289918737, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.503314971923828, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7104978561401367, + "num_tokens": 99941534.0, + "step": 3864 + }, + { + "epoch": 0.424445420601801, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1926159858703613, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.706652045249939, + "num_tokens": 99968505.0, + "step": 3865 + }, + { + "epoch": 0.42455523830441466, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1582090854644775, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7311224937438965, + "num_tokens": 99993455.0, + "step": 3866 + }, + { + "epoch": 0.4246650560070283, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.354827880859375, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.7000464200973511, + "num_tokens": 100018287.0, + "step": 3867 + }, + { + "epoch": 0.424774873709642, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3985531330108643, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7102991342544556, + "num_tokens": 100040541.0, + "step": 3868 + }, + { + "epoch": 0.42488469141225566, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1467738151550293, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7169952392578125, + "num_tokens": 100066442.0, + "step": 3869 + }, + { + "epoch": 0.4249945091148693, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.358741044998169, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7505717277526855, + "num_tokens": 100086709.0, + "step": 3870 + }, + { + "epoch": 0.425104326817483, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.352834939956665, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7129815220832825, + "num_tokens": 100110542.0, + "step": 3871 + }, + { + "epoch": 0.42521414452009665, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.0362842082977295, + "learning_rate": 1e-06, + "loss": 1.1064, + "mean_token_accuracy": 0.6743057370185852, + "num_tokens": 100141320.0, + "step": 3872 + }, + { + "epoch": 0.4253239622227103, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.4070205688476562, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6855964064598083, + "num_tokens": 100164223.0, + "step": 3873 + }, + { + "epoch": 0.42543377992532394, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.4164788722991943, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7133917808532715, + "num_tokens": 100185941.0, + "step": 3874 + }, + { + "epoch": 0.42554359762793764, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.203003168106079, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7095794677734375, + "num_tokens": 100210732.0, + "step": 3875 + }, + { + "epoch": 0.4256534153305513, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1380372047424316, + "learning_rate": 1e-06, + "loss": 1.0669, + "mean_token_accuracy": 0.6832040548324585, + "num_tokens": 100238137.0, + "step": 3876 + }, + { + "epoch": 0.42576323303316493, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1572837829589844, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.6911168098449707, + "num_tokens": 100264186.0, + "step": 3877 + }, + { + "epoch": 0.4258730507357786, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 1.9537841081619263, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6989474296569824, + "num_tokens": 100294331.0, + "step": 3878 + }, + { + "epoch": 0.4259828684383923, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.2869863510131836, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7337950468063354, + "num_tokens": 100316130.0, + "step": 3879 + }, + { + "epoch": 0.4260926861410059, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3355114459991455, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7006436586380005, + "num_tokens": 100340052.0, + "step": 3880 + }, + { + "epoch": 0.4262025038436196, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.363661289215088, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.699320912361145, + "num_tokens": 100365026.0, + "step": 3881 + }, + { + "epoch": 0.4263123215462333, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3968145847320557, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6956859827041626, + "num_tokens": 100386885.0, + "step": 3882 + }, + { + "epoch": 0.4264221392488469, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.2282230854034424, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.6829497814178467, + "num_tokens": 100413857.0, + "step": 3883 + }, + { + "epoch": 0.42653195695146057, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 1.823996901512146, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7020500302314758, + "num_tokens": 100448443.0, + "step": 3884 + }, + { + "epoch": 0.4266417746540742, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.216695785522461, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7283515334129333, + "num_tokens": 100473036.0, + "step": 3885 + }, + { + "epoch": 0.4267515923566879, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.5894315242767334, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6982683539390564, + "num_tokens": 100492808.0, + "step": 3886 + }, + { + "epoch": 0.42686141005930156, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3308298587799072, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.686470627784729, + "num_tokens": 100516357.0, + "step": 3887 + }, + { + "epoch": 0.4269712277619152, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.258119821548462, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7085201740264893, + "num_tokens": 100539907.0, + "step": 3888 + }, + { + "epoch": 0.4270810454645289, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.0066542625427246, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7162115573883057, + "num_tokens": 100570127.0, + "step": 3889 + }, + { + "epoch": 0.42719086316714255, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.8249878883361816, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7222144603729248, + "num_tokens": 100587307.0, + "step": 3890 + }, + { + "epoch": 0.4273006808697562, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.414499521255493, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6973108053207397, + "num_tokens": 100610619.0, + "step": 3891 + }, + { + "epoch": 0.42741049857236985, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.1227073669433594, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6945083737373352, + "num_tokens": 100639567.0, + "step": 3892 + }, + { + "epoch": 0.42752031627498355, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.4734277725219727, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7044969797134399, + "num_tokens": 100659524.0, + "step": 3893 + }, + { + "epoch": 0.4276301339775972, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3613638877868652, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7084023356437683, + "num_tokens": 100681819.0, + "step": 3894 + }, + { + "epoch": 0.42773995168021084, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 1.9113514423370361, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.7008652687072754, + "num_tokens": 100715905.0, + "step": 3895 + }, + { + "epoch": 0.4278497693828245, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 1.984635829925537, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6948019862174988, + "num_tokens": 100751228.0, + "step": 3896 + }, + { + "epoch": 0.4279595870854382, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.4360175132751465, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7038204669952393, + "num_tokens": 100773363.0, + "step": 3897 + }, + { + "epoch": 0.42806940478805183, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.0388667583465576, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6892449855804443, + "num_tokens": 100801727.0, + "step": 3898 + }, + { + "epoch": 0.4281792224906655, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 1.9249143600463867, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7042582035064697, + "num_tokens": 100834173.0, + "step": 3899 + }, + { + "epoch": 0.4282890401932792, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.51749849319458, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6994425058364868, + "num_tokens": 100855435.0, + "step": 3900 + }, + { + "epoch": 0.4283988578958928, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.0700466632843018, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.691155195236206, + "num_tokens": 100884327.0, + "step": 3901 + }, + { + "epoch": 0.42850867559850647, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.182952404022217, + "learning_rate": 1e-06, + "loss": 1.0607, + "mean_token_accuracy": 0.6778642535209656, + "num_tokens": 100910211.0, + "step": 3902 + }, + { + "epoch": 0.4286184933011201, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3368256092071533, + "learning_rate": 1e-06, + "loss": 1.0634, + "mean_token_accuracy": 0.6839659214019775, + "num_tokens": 100935401.0, + "step": 3903 + }, + { + "epoch": 0.4287283110037338, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.3190231323242188, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7136778831481934, + "num_tokens": 100958877.0, + "step": 3904 + }, + { + "epoch": 0.42883812870634747, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.321009397506714, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7117908000946045, + "num_tokens": 100980312.0, + "step": 3905 + }, + { + "epoch": 0.4289479464089611, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.022951602935791, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7175766229629517, + "num_tokens": 101008596.0, + "step": 3906 + }, + { + "epoch": 0.4290577641115748, + "ewc_loss": 3.361701965332031e-05, + "grad_norm": 2.198728084564209, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7182689905166626, + "num_tokens": 101033996.0, + "step": 3907 + }, + { + "epoch": 0.42916758181418846, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.265319347381592, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7056986093521118, + "num_tokens": 101057963.0, + "step": 3908 + }, + { + "epoch": 0.4292773995168021, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.6152968406677246, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6957229375839233, + "num_tokens": 101077804.0, + "step": 3909 + }, + { + "epoch": 0.42938721721941575, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.5358002185821533, + "learning_rate": 1e-06, + "loss": 1.0738, + "mean_token_accuracy": 0.6874211430549622, + "num_tokens": 101100214.0, + "step": 3910 + }, + { + "epoch": 0.42949703492202945, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.3321094512939453, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7268959879875183, + "num_tokens": 101124305.0, + "step": 3911 + }, + { + "epoch": 0.4296068526246431, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.151280641555786, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7215920686721802, + "num_tokens": 101150289.0, + "step": 3912 + }, + { + "epoch": 0.42971667032725674, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.207982301712036, + "learning_rate": 1e-06, + "loss": 1.0545, + "mean_token_accuracy": 0.6906369924545288, + "num_tokens": 101178339.0, + "step": 3913 + }, + { + "epoch": 0.4298264880298704, + "ewc_loss": 3.409385681152344e-05, + "grad_norm": 2.2821552753448486, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7149993777275085, + "num_tokens": 101202614.0, + "step": 3914 + }, + { + "epoch": 0.4299363057324841, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.288161277770996, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7213543653488159, + "num_tokens": 101224528.0, + "step": 3915 + }, + { + "epoch": 0.43004612343509774, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.110330581665039, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.732822060585022, + "num_tokens": 101249037.0, + "step": 3916 + }, + { + "epoch": 0.4301559411377114, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 1.9770586490631104, + "learning_rate": 1e-06, + "loss": 1.0309, + "mean_token_accuracy": 0.6954679489135742, + "num_tokens": 101282444.0, + "step": 3917 + }, + { + "epoch": 0.4302657588403251, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.097707748413086, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7083572149276733, + "num_tokens": 101313358.0, + "step": 3918 + }, + { + "epoch": 0.43037557654293873, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.0833487510681152, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.6994529366493225, + "num_tokens": 101340581.0, + "step": 3919 + }, + { + "epoch": 0.4304853942455524, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.5143730640411377, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7166863083839417, + "num_tokens": 101361533.0, + "step": 3920 + }, + { + "epoch": 0.430595211948166, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.3967416286468506, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7105029225349426, + "num_tokens": 101383326.0, + "step": 3921 + }, + { + "epoch": 0.4307050296507797, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2746598720550537, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.705539882183075, + "num_tokens": 101409081.0, + "step": 3922 + }, + { + "epoch": 0.43081484735339337, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.104651927947998, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6986355781555176, + "num_tokens": 101438699.0, + "step": 3923 + }, + { + "epoch": 0.430924665056007, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.079155921936035, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7137940526008606, + "num_tokens": 101464864.0, + "step": 3924 + }, + { + "epoch": 0.43103448275862066, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.4553472995758057, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.6900265216827393, + "num_tokens": 101488078.0, + "step": 3925 + }, + { + "epoch": 0.43114430046123436, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.283102035522461, + "learning_rate": 1e-06, + "loss": 1.0891, + "mean_token_accuracy": 0.6814504265785217, + "num_tokens": 101514050.0, + "step": 3926 + }, + { + "epoch": 0.431254118163848, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.4090449810028076, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7224522233009338, + "num_tokens": 101536723.0, + "step": 3927 + }, + { + "epoch": 0.43136393586646166, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2855682373046875, + "learning_rate": 1e-06, + "loss": 1.0508, + "mean_token_accuracy": 0.6883044242858887, + "num_tokens": 101561970.0, + "step": 3928 + }, + { + "epoch": 0.43147375356907536, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1015114784240723, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7036205530166626, + "num_tokens": 101589583.0, + "step": 3929 + }, + { + "epoch": 0.431583571271689, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.020313024520874, + "learning_rate": 1e-06, + "loss": 1.0424, + "mean_token_accuracy": 0.6839434504508972, + "num_tokens": 101621673.0, + "step": 3930 + }, + { + "epoch": 0.43169338897430265, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.4024109840393066, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6926475763320923, + "num_tokens": 101644065.0, + "step": 3931 + }, + { + "epoch": 0.4318032066769163, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.9033563137054443, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7302423715591431, + "num_tokens": 101662750.0, + "step": 3932 + }, + { + "epoch": 0.43191302437953, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1241912841796875, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7092865109443665, + "num_tokens": 101688990.0, + "step": 3933 + }, + { + "epoch": 0.43202284208214364, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.427773952484131, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6968847513198853, + "num_tokens": 101710686.0, + "step": 3934 + }, + { + "epoch": 0.4321326597847573, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2848522663116455, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.6904339790344238, + "num_tokens": 101737899.0, + "step": 3935 + }, + { + "epoch": 0.432242477487371, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1349544525146484, + "learning_rate": 1e-06, + "loss": 1.0494, + "mean_token_accuracy": 0.6848559379577637, + "num_tokens": 101766454.0, + "step": 3936 + }, + { + "epoch": 0.43235229518998464, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.254920244216919, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7148135900497437, + "num_tokens": 101791654.0, + "step": 3937 + }, + { + "epoch": 0.4324621128925983, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2056052684783936, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6987264156341553, + "num_tokens": 101817589.0, + "step": 3938 + }, + { + "epoch": 0.43257193059521193, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.018254280090332, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6934717297554016, + "num_tokens": 101847859.0, + "step": 3939 + }, + { + "epoch": 0.43268174829782563, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2867236137390137, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7012535333633423, + "num_tokens": 101871930.0, + "step": 3940 + }, + { + "epoch": 0.4327915660004393, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.073876142501831, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7037041187286377, + "num_tokens": 101898638.0, + "step": 3941 + }, + { + "epoch": 0.4329013837030529, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.221137285232544, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.692874550819397, + "num_tokens": 101926998.0, + "step": 3942 + }, + { + "epoch": 0.43301120140566657, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.47019624710083, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6917335987091064, + "num_tokens": 101949123.0, + "step": 3943 + }, + { + "epoch": 0.43312101910828027, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.3160154819488525, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7143582701683044, + "num_tokens": 101971407.0, + "step": 3944 + }, + { + "epoch": 0.4332308368108939, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.191455125808716, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.7048406600952148, + "num_tokens": 101997943.0, + "step": 3945 + }, + { + "epoch": 0.43334065451350756, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.0587875843048096, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7078044414520264, + "num_tokens": 102025822.0, + "step": 3946 + }, + { + "epoch": 0.43345047221612126, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.0862631797790527, + "learning_rate": 1e-06, + "loss": 1.05, + "mean_token_accuracy": 0.6896654367446899, + "num_tokens": 102053305.0, + "step": 3947 + }, + { + "epoch": 0.4335602899187349, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.246969699859619, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.7032192945480347, + "num_tokens": 102078836.0, + "step": 3948 + }, + { + "epoch": 0.43367010762134856, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2133281230926514, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7234398126602173, + "num_tokens": 102103001.0, + "step": 3949 + }, + { + "epoch": 0.4337799253239622, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2062859535217285, + "learning_rate": 1e-06, + "loss": 1.047, + "mean_token_accuracy": 0.6840453147888184, + "num_tokens": 102131895.0, + "step": 3950 + }, + { + "epoch": 0.4338897430265759, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.4448652267456055, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7200635075569153, + "num_tokens": 102156023.0, + "step": 3951 + }, + { + "epoch": 0.43399956072918955, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.6534969806671143, + "learning_rate": 1e-06, + "loss": 1.0656, + "mean_token_accuracy": 0.6878711581230164, + "num_tokens": 102176947.0, + "step": 3952 + }, + { + "epoch": 0.4341093784318032, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2515361309051514, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6981886625289917, + "num_tokens": 102203157.0, + "step": 3953 + }, + { + "epoch": 0.43421919613441684, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.422124147415161, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7004445791244507, + "num_tokens": 102226106.0, + "step": 3954 + }, + { + "epoch": 0.43432901383703054, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.356011390686035, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6957216262817383, + "num_tokens": 102250220.0, + "step": 3955 + }, + { + "epoch": 0.4344388315396442, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1923630237579346, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6791038513183594, + "num_tokens": 102276373.0, + "step": 3956 + }, + { + "epoch": 0.43454864924225783, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.359729290008545, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7086364030838013, + "num_tokens": 102298062.0, + "step": 3957 + }, + { + "epoch": 0.43465846694487154, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2027926445007324, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7004148960113525, + "num_tokens": 102324012.0, + "step": 3958 + }, + { + "epoch": 0.4347682846474852, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.173508882522583, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.687057375907898, + "num_tokens": 102352815.0, + "step": 3959 + }, + { + "epoch": 0.4348781023500988, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.228525161743164, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6990388631820679, + "num_tokens": 102378014.0, + "step": 3960 + }, + { + "epoch": 0.4349879200527125, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.2353146076202393, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7292653322219849, + "num_tokens": 102402223.0, + "step": 3961 + }, + { + "epoch": 0.4350977377553262, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.0840983390808105, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7348037958145142, + "num_tokens": 102429827.0, + "step": 3962 + }, + { + "epoch": 0.4352075554579398, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.057950258255005, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6954990029335022, + "num_tokens": 102458769.0, + "step": 3963 + }, + { + "epoch": 0.43531737316055347, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.0808310508728027, + "learning_rate": 1e-06, + "loss": 1.025, + "mean_token_accuracy": 0.6965688467025757, + "num_tokens": 102487677.0, + "step": 3964 + }, + { + "epoch": 0.43542719086316717, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1222455501556396, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.7050328254699707, + "num_tokens": 102516194.0, + "step": 3965 + }, + { + "epoch": 0.4355370085657808, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.834211587905884, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7197731733322144, + "num_tokens": 102534115.0, + "step": 3966 + }, + { + "epoch": 0.43564682626839446, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.3537254333496094, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7147495746612549, + "num_tokens": 102557009.0, + "step": 3967 + }, + { + "epoch": 0.4357566439710081, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.184208869934082, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6967376470565796, + "num_tokens": 102581438.0, + "step": 3968 + }, + { + "epoch": 0.4358664616736218, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.3893091678619385, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7175694704055786, + "num_tokens": 102603822.0, + "step": 3969 + }, + { + "epoch": 0.43597627937623545, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.3694121837615967, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7067912817001343, + "num_tokens": 102625582.0, + "step": 3970 + }, + { + "epoch": 0.4360860970788491, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.1689138412475586, + "learning_rate": 1e-06, + "loss": 1.0527, + "mean_token_accuracy": 0.6888568997383118, + "num_tokens": 102655237.0, + "step": 3971 + }, + { + "epoch": 0.43619591478146275, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.1973559856414795, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7078888416290283, + "num_tokens": 102681610.0, + "step": 3972 + }, + { + "epoch": 0.43630573248407645, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.0644564628601074, + "learning_rate": 1e-06, + "loss": 1.0384, + "mean_token_accuracy": 0.6808838248252869, + "num_tokens": 102711636.0, + "step": 3973 + }, + { + "epoch": 0.4364155501866901, + "ewc_loss": 3.4332275390625e-05, + "grad_norm": 2.212644100189209, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6795355677604675, + "num_tokens": 102740269.0, + "step": 3974 + }, + { + "epoch": 0.43652536788930374, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.200772762298584, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7302883863449097, + "num_tokens": 102765140.0, + "step": 3975 + }, + { + "epoch": 0.43663518559191744, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.3915607929229736, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6948684453964233, + "num_tokens": 102790550.0, + "step": 3976 + }, + { + "epoch": 0.4367450032945311, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.113774299621582, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6922264099121094, + "num_tokens": 102818283.0, + "step": 3977 + }, + { + "epoch": 0.43685482099714473, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.416980028152466, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6976807713508606, + "num_tokens": 102839697.0, + "step": 3978 + }, + { + "epoch": 0.4369646386997584, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.0754990577697754, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6920691728591919, + "num_tokens": 102868038.0, + "step": 3979 + }, + { + "epoch": 0.4370744564023721, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.4270005226135254, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6975419521331787, + "num_tokens": 102890261.0, + "step": 3980 + }, + { + "epoch": 0.4371842741049857, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.4928767681121826, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6850936412811279, + "num_tokens": 102912602.0, + "step": 3981 + }, + { + "epoch": 0.43729409180759937, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.290285587310791, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7195032238960266, + "num_tokens": 102936652.0, + "step": 3982 + }, + { + "epoch": 0.4374039095102131, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.0663506984710693, + "learning_rate": 1e-06, + "loss": 1.0614, + "mean_token_accuracy": 0.679653525352478, + "num_tokens": 102964564.0, + "step": 3983 + }, + { + "epoch": 0.4375137272128267, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.156017303466797, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.72368323802948, + "num_tokens": 102991026.0, + "step": 3984 + }, + { + "epoch": 0.43762354491544037, + "ewc_loss": 3.457069396972656e-05, + "grad_norm": 2.191960096359253, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6947180032730103, + "num_tokens": 103017666.0, + "step": 3985 + }, + { + "epoch": 0.437733362618054, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.498939275741577, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7195633053779602, + "num_tokens": 103038445.0, + "step": 3986 + }, + { + "epoch": 0.4378431803206677, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.116016387939453, + "learning_rate": 1e-06, + "loss": 1.0301, + "mean_token_accuracy": 0.6874677538871765, + "num_tokens": 103069019.0, + "step": 3987 + }, + { + "epoch": 0.43795299802328136, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.516275405883789, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7077988982200623, + "num_tokens": 103090281.0, + "step": 3988 + }, + { + "epoch": 0.438062815725895, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.0454788208007812, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7305054068565369, + "num_tokens": 103118962.0, + "step": 3989 + }, + { + "epoch": 0.43817263342850865, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.161447286605835, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7189794182777405, + "num_tokens": 103145401.0, + "step": 3990 + }, + { + "epoch": 0.43828245113112235, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.1652681827545166, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.7008552551269531, + "num_tokens": 103172448.0, + "step": 3991 + }, + { + "epoch": 0.438392268833736, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.2318520545959473, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.707549512386322, + "num_tokens": 103198006.0, + "step": 3992 + }, + { + "epoch": 0.43850208653634964, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.2938618659973145, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7067276835441589, + "num_tokens": 103222454.0, + "step": 3993 + }, + { + "epoch": 0.43861190423896335, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.1945273876190186, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7096673846244812, + "num_tokens": 103249089.0, + "step": 3994 + }, + { + "epoch": 0.438721721941577, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.3857412338256836, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7291244268417358, + "num_tokens": 103270197.0, + "step": 3995 + }, + { + "epoch": 0.43883153964419064, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.1708266735076904, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.6906670331954956, + "num_tokens": 103297499.0, + "step": 3996 + }, + { + "epoch": 0.4389413573468043, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.185495138168335, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7141422033309937, + "num_tokens": 103324855.0, + "step": 3997 + }, + { + "epoch": 0.439051175049418, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.173046350479126, + "learning_rate": 1e-06, + "loss": 1.0964, + "mean_token_accuracy": 0.6807482838630676, + "num_tokens": 103352855.0, + "step": 3998 + }, + { + "epoch": 0.43916099275203163, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.290578603744507, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.698172926902771, + "num_tokens": 103375974.0, + "step": 3999 + }, + { + "epoch": 0.4392708104546453, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.002512216567993, + "learning_rate": 1e-06, + "loss": 1.1056, + "mean_token_accuracy": 0.6678575277328491, + "num_tokens": 103407738.0, + "step": 4000 + }, + { + "epoch": 0.4393806281572589, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.0955166816711426, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6962682604789734, + "num_tokens": 103437027.0, + "step": 4001 + }, + { + "epoch": 0.4394904458598726, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.3301937580108643, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7055538892745972, + "num_tokens": 103460046.0, + "step": 4002 + }, + { + "epoch": 0.43960026356248627, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.243835687637329, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7053027153015137, + "num_tokens": 103485621.0, + "step": 4003 + }, + { + "epoch": 0.4397100812650999, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 1.8825019598007202, + "learning_rate": 1e-06, + "loss": 1.0317, + "mean_token_accuracy": 0.6845036149024963, + "num_tokens": 103518984.0, + "step": 4004 + }, + { + "epoch": 0.4398198989677136, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.404649019241333, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6935676336288452, + "num_tokens": 103541595.0, + "step": 4005 + }, + { + "epoch": 0.43992971667032726, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.340932846069336, + "learning_rate": 1e-06, + "loss": 1.03, + "mean_token_accuracy": 0.6878314018249512, + "num_tokens": 103565995.0, + "step": 4006 + }, + { + "epoch": 0.4400395343729409, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.4744210243225098, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6891990900039673, + "num_tokens": 103589961.0, + "step": 4007 + }, + { + "epoch": 0.44014935207555456, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 1.9361494779586792, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6876498460769653, + "num_tokens": 103624015.0, + "step": 4008 + }, + { + "epoch": 0.44025916977816826, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.1853187084198, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7183946371078491, + "num_tokens": 103648981.0, + "step": 4009 + }, + { + "epoch": 0.4403689874807819, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.319080114364624, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7133411169052124, + "num_tokens": 103672992.0, + "step": 4010 + }, + { + "epoch": 0.44047880518339555, + "ewc_loss": 3.528594970703125e-05, + "grad_norm": 2.476409912109375, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7102288007736206, + "num_tokens": 103693983.0, + "step": 4011 + }, + { + "epoch": 0.44058862288600925, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.3238096237182617, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6875125169754028, + "num_tokens": 103717413.0, + "step": 4012 + }, + { + "epoch": 0.4406984405886229, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4322798252105713, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7066633701324463, + "num_tokens": 103739793.0, + "step": 4013 + }, + { + "epoch": 0.44080825829123654, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.322715997695923, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7140040397644043, + "num_tokens": 103763425.0, + "step": 4014 + }, + { + "epoch": 0.4409180759938502, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4774677753448486, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7281639575958252, + "num_tokens": 103784225.0, + "step": 4015 + }, + { + "epoch": 0.4410278936964639, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.1001453399658203, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7047761678695679, + "num_tokens": 103811751.0, + "step": 4016 + }, + { + "epoch": 0.44113771139907754, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 1.932503342628479, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6916027069091797, + "num_tokens": 103843323.0, + "step": 4017 + }, + { + "epoch": 0.4412475291016912, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2674434185028076, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.7098586559295654, + "num_tokens": 103869122.0, + "step": 4018 + }, + { + "epoch": 0.44135734680430483, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2661619186401367, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6915791034698486, + "num_tokens": 103892813.0, + "step": 4019 + }, + { + "epoch": 0.44146716450691853, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 1.8979582786560059, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.6914352178573608, + "num_tokens": 103925498.0, + "step": 4020 + }, + { + "epoch": 0.4415769822095322, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.324333429336548, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7151509523391724, + "num_tokens": 103948200.0, + "step": 4021 + }, + { + "epoch": 0.4416867999121458, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.3073129653930664, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.731122612953186, + "num_tokens": 103969306.0, + "step": 4022 + }, + { + "epoch": 0.4417966176147595, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4286184310913086, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7125368714332581, + "num_tokens": 103992620.0, + "step": 4023 + }, + { + "epoch": 0.44190643531737317, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.154259443283081, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6975213885307312, + "num_tokens": 104022069.0, + "step": 4024 + }, + { + "epoch": 0.4420162530199868, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.087393045425415, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6973406076431274, + "num_tokens": 104052384.0, + "step": 4025 + }, + { + "epoch": 0.44212607072260046, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.146162271499634, + "learning_rate": 1e-06, + "loss": 1.0574, + "mean_token_accuracy": 0.6862144470214844, + "num_tokens": 104080332.0, + "step": 4026 + }, + { + "epoch": 0.44223588842521416, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.1820337772369385, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6916032433509827, + "num_tokens": 104106260.0, + "step": 4027 + }, + { + "epoch": 0.4423457061278278, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 1.974373459815979, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6963911056518555, + "num_tokens": 104135474.0, + "step": 4028 + }, + { + "epoch": 0.44245552383044146, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.375401258468628, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.701628565788269, + "num_tokens": 104157966.0, + "step": 4029 + }, + { + "epoch": 0.4425653415330551, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4271533489227295, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6936160326004028, + "num_tokens": 104180901.0, + "step": 4030 + }, + { + "epoch": 0.4426751592356688, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.09921932220459, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7024881839752197, + "num_tokens": 104210427.0, + "step": 4031 + }, + { + "epoch": 0.44278497693828245, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.0105273723602295, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7114754915237427, + "num_tokens": 104240781.0, + "step": 4032 + }, + { + "epoch": 0.4428947946408961, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.26601505279541, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.6987152099609375, + "num_tokens": 104265444.0, + "step": 4033 + }, + { + "epoch": 0.4430046123435098, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2825987339019775, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7092905044555664, + "num_tokens": 104289851.0, + "step": 4034 + }, + { + "epoch": 0.44311443004612344, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.076734781265259, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7072117924690247, + "num_tokens": 104318477.0, + "step": 4035 + }, + { + "epoch": 0.4432242477487371, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2162392139434814, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.6885561347007751, + "num_tokens": 104344467.0, + "step": 4036 + }, + { + "epoch": 0.44333406545135073, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2584941387176514, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.6881479024887085, + "num_tokens": 104370130.0, + "step": 4037 + }, + { + "epoch": 0.44344388315396444, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.259122848510742, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7226405739784241, + "num_tokens": 104393844.0, + "step": 4038 + }, + { + "epoch": 0.4435537008565781, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.5897629261016846, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6976185441017151, + "num_tokens": 104413570.0, + "step": 4039 + }, + { + "epoch": 0.4436635185591917, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.1274116039276123, + "learning_rate": 1e-06, + "loss": 1.1068, + "mean_token_accuracy": 0.667695164680481, + "num_tokens": 104443227.0, + "step": 4040 + }, + { + "epoch": 0.44377333626180543, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.30299973487854, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7044210433959961, + "num_tokens": 104467348.0, + "step": 4041 + }, + { + "epoch": 0.4438831539644191, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4818005561828613, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7172569036483765, + "num_tokens": 104487199.0, + "step": 4042 + }, + { + "epoch": 0.4439929716670327, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4103238582611084, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7032229900360107, + "num_tokens": 104508389.0, + "step": 4043 + }, + { + "epoch": 0.44410278936964637, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 1.9747028350830078, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7084635496139526, + "num_tokens": 104541411.0, + "step": 4044 + }, + { + "epoch": 0.44421260707226007, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.0873076915740967, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.713667631149292, + "num_tokens": 104568153.0, + "step": 4045 + }, + { + "epoch": 0.4443224247748737, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.4313411712646484, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.6915024518966675, + "num_tokens": 104591936.0, + "step": 4046 + }, + { + "epoch": 0.44443224247748736, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.3100528717041016, + "learning_rate": 1e-06, + "loss": 1.0543, + "mean_token_accuracy": 0.6857475638389587, + "num_tokens": 104621841.0, + "step": 4047 + }, + { + "epoch": 0.444542060180101, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.147102117538452, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6983243227005005, + "num_tokens": 104651647.0, + "step": 4048 + }, + { + "epoch": 0.4446518778827147, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.3815598487854004, + "learning_rate": 1e-06, + "loss": 1.0567, + "mean_token_accuracy": 0.6880399584770203, + "num_tokens": 104677572.0, + "step": 4049 + }, + { + "epoch": 0.44476169558532835, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.0778229236602783, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7161363363265991, + "num_tokens": 104706092.0, + "step": 4050 + }, + { + "epoch": 0.444871513287942, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.237084150314331, + "learning_rate": 1e-06, + "loss": 1.0584, + "mean_token_accuracy": 0.6849140524864197, + "num_tokens": 104732079.0, + "step": 4051 + }, + { + "epoch": 0.4449813309905557, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.165898084640503, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6877171993255615, + "num_tokens": 104760767.0, + "step": 4052 + }, + { + "epoch": 0.44509114869316935, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.184687852859497, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7156790494918823, + "num_tokens": 104788430.0, + "step": 4053 + }, + { + "epoch": 0.445200966395783, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.224942922592163, + "learning_rate": 1e-06, + "loss": 1.0315, + "mean_token_accuracy": 0.6903026103973389, + "num_tokens": 104817342.0, + "step": 4054 + }, + { + "epoch": 0.44531078409839664, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.1543145179748535, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.6999555826187134, + "num_tokens": 104843385.0, + "step": 4055 + }, + { + "epoch": 0.44542060180101034, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.2680516242980957, + "learning_rate": 1e-06, + "loss": 1.0744, + "mean_token_accuracy": 0.6762055158615112, + "num_tokens": 104868452.0, + "step": 4056 + }, + { + "epoch": 0.445530419503624, + "ewc_loss": 3.552436828613281e-05, + "grad_norm": 2.3294687271118164, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7235028147697449, + "num_tokens": 104890186.0, + "step": 4057 + }, + { + "epoch": 0.44564023720623763, + "ewc_loss": 3.5762786865234375e-05, + "grad_norm": 2.247602701187134, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6930155754089355, + "num_tokens": 104914247.0, + "step": 4058 + }, + { + "epoch": 0.44575005490885133, + "ewc_loss": 3.600120544433594e-05, + "grad_norm": 2.183114767074585, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.6932398080825806, + "num_tokens": 104939335.0, + "step": 4059 + }, + { + "epoch": 0.445859872611465, + "ewc_loss": 3.600120544433594e-05, + "grad_norm": 2.3471624851226807, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7105492949485779, + "num_tokens": 104962635.0, + "step": 4060 + }, + { + "epoch": 0.4459696903140786, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3936328887939453, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7247816920280457, + "num_tokens": 104983584.0, + "step": 4061 + }, + { + "epoch": 0.44607950801669227, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3968422412872314, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.71983802318573, + "num_tokens": 105004819.0, + "step": 4062 + }, + { + "epoch": 0.446189325719306, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.368394374847412, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7328091859817505, + "num_tokens": 105026420.0, + "step": 4063 + }, + { + "epoch": 0.4462991434219196, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3713178634643555, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7054502367973328, + "num_tokens": 105048663.0, + "step": 4064 + }, + { + "epoch": 0.44640896112453327, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.4663517475128174, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7044864892959595, + "num_tokens": 105072117.0, + "step": 4065 + }, + { + "epoch": 0.4465187788271469, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.170537233352661, + "learning_rate": 1e-06, + "loss": 1.1159, + "mean_token_accuracy": 0.6722071170806885, + "num_tokens": 105100997.0, + "step": 4066 + }, + { + "epoch": 0.4466285965297606, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.236600875854492, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.690908670425415, + "num_tokens": 105126987.0, + "step": 4067 + }, + { + "epoch": 0.44673841423237426, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.2753708362579346, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7098655700683594, + "num_tokens": 105151037.0, + "step": 4068 + }, + { + "epoch": 0.4468482319349879, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.151968002319336, + "learning_rate": 1e-06, + "loss": 1.0641, + "mean_token_accuracy": 0.6810765862464905, + "num_tokens": 105179052.0, + "step": 4069 + }, + { + "epoch": 0.4469580496376016, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.2984321117401123, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7122949361801147, + "num_tokens": 105202762.0, + "step": 4070 + }, + { + "epoch": 0.44706786734021525, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.06913161277771, + "learning_rate": 1e-06, + "loss": 1.0629, + "mean_token_accuracy": 0.6818516254425049, + "num_tokens": 105231743.0, + "step": 4071 + }, + { + "epoch": 0.4471776850428289, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.141475200653076, + "learning_rate": 1e-06, + "loss": 1.0664, + "mean_token_accuracy": 0.6775689125061035, + "num_tokens": 105258755.0, + "step": 4072 + }, + { + "epoch": 0.44728750274544254, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3208675384521484, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6928385496139526, + "num_tokens": 105283134.0, + "step": 4073 + }, + { + "epoch": 0.44739732044805625, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.5672545433044434, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7350034713745117, + "num_tokens": 105302217.0, + "step": 4074 + }, + { + "epoch": 0.4475071381506699, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.37675142288208, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7009704113006592, + "num_tokens": 105324669.0, + "step": 4075 + }, + { + "epoch": 0.44761695585328354, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.198605537414551, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6885559558868408, + "num_tokens": 105351150.0, + "step": 4076 + }, + { + "epoch": 0.4477267735558972, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 1.897396206855774, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.7016245126724243, + "num_tokens": 105382344.0, + "step": 4077 + }, + { + "epoch": 0.4478365912585109, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 1.9681313037872314, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.714953601360321, + "num_tokens": 105412565.0, + "step": 4078 + }, + { + "epoch": 0.44794640896112453, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.4484691619873047, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.6978963613510132, + "num_tokens": 105434373.0, + "step": 4079 + }, + { + "epoch": 0.4480562266637382, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.38096022605896, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6935542821884155, + "num_tokens": 105457771.0, + "step": 4080 + }, + { + "epoch": 0.4481660443663519, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.0691347122192383, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7188533544540405, + "num_tokens": 105483652.0, + "step": 4081 + }, + { + "epoch": 0.4482758620689655, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.297532558441162, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7219923734664917, + "num_tokens": 105506506.0, + "step": 4082 + }, + { + "epoch": 0.44838567977157917, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.196566581726074, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7140382528305054, + "num_tokens": 105535039.0, + "step": 4083 + }, + { + "epoch": 0.4484954974741928, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.144881248474121, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6971482038497925, + "num_tokens": 105561456.0, + "step": 4084 + }, + { + "epoch": 0.4486053151768065, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.236686944961548, + "learning_rate": 1e-06, + "loss": 1.0422, + "mean_token_accuracy": 0.6838381290435791, + "num_tokens": 105587627.0, + "step": 4085 + }, + { + "epoch": 0.44871513287942016, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3492400646209717, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7078431248664856, + "num_tokens": 105613040.0, + "step": 4086 + }, + { + "epoch": 0.4488249505820338, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.3192267417907715, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.7021442651748657, + "num_tokens": 105638821.0, + "step": 4087 + }, + { + "epoch": 0.4489347682846475, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3551855087280273, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.700295627117157, + "num_tokens": 105661894.0, + "step": 4088 + }, + { + "epoch": 0.44904458598726116, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 1.8661134243011475, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6905200481414795, + "num_tokens": 105697736.0, + "step": 4089 + }, + { + "epoch": 0.4491544036898748, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.414685010910034, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7254139184951782, + "num_tokens": 105717630.0, + "step": 4090 + }, + { + "epoch": 0.44926422139248845, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.34159517288208, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7084519863128662, + "num_tokens": 105739780.0, + "step": 4091 + }, + { + "epoch": 0.44937403909510215, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.108898878097534, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7155154943466187, + "num_tokens": 105767513.0, + "step": 4092 + }, + { + "epoch": 0.4494838567977158, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.233260154724121, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.700946569442749, + "num_tokens": 105793045.0, + "step": 4093 + }, + { + "epoch": 0.44959367450032944, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.140967845916748, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6948060989379883, + "num_tokens": 105818510.0, + "step": 4094 + }, + { + "epoch": 0.4497034922029431, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.603919267654419, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7105605006217957, + "num_tokens": 105836643.0, + "step": 4095 + }, + { + "epoch": 0.4498133099055568, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 1.9763901233673096, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6849038600921631, + "num_tokens": 105865818.0, + "step": 4096 + }, + { + "epoch": 0.44992312760817044, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.1958868503570557, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.713183581829071, + "num_tokens": 105891557.0, + "step": 4097 + }, + { + "epoch": 0.4500329453107841, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.3837926387786865, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6914096474647522, + "num_tokens": 105914959.0, + "step": 4098 + }, + { + "epoch": 0.4501427630133978, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.2452549934387207, + "learning_rate": 1e-06, + "loss": 1.036, + "mean_token_accuracy": 0.6971089839935303, + "num_tokens": 105940788.0, + "step": 4099 + }, + { + "epoch": 0.45025258071601143, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.6676483154296875, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7273382544517517, + "num_tokens": 105960218.0, + "step": 4100 + }, + { + "epoch": 0.4503623984186251, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 2.280465602874756, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7088404893875122, + "num_tokens": 105985451.0, + "step": 4101 + }, + { + "epoch": 0.4504722161212387, + "ewc_loss": 3.62396240234375e-05, + "grad_norm": 1.993990182876587, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7100880742073059, + "num_tokens": 106015411.0, + "step": 4102 + }, + { + "epoch": 0.4505820338238524, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.2492237091064453, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6964877843856812, + "num_tokens": 106040921.0, + "step": 4103 + }, + { + "epoch": 0.45069185152646607, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.2555370330810547, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7073535919189453, + "num_tokens": 106065656.0, + "step": 4104 + }, + { + "epoch": 0.4508016692290797, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.0097341537475586, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7083432674407959, + "num_tokens": 106097280.0, + "step": 4105 + }, + { + "epoch": 0.45091148693169336, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.082653284072876, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.729464590549469, + "num_tokens": 106123047.0, + "step": 4106 + }, + { + "epoch": 0.45102130463430706, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.1154065132141113, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7126116752624512, + "num_tokens": 106151119.0, + "step": 4107 + }, + { + "epoch": 0.4511311223369207, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 1.9487850666046143, + "learning_rate": 1e-06, + "loss": 1.0422, + "mean_token_accuracy": 0.6835658550262451, + "num_tokens": 106183173.0, + "step": 4108 + }, + { + "epoch": 0.45124094003953436, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.3759055137634277, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7099895477294922, + "num_tokens": 106206388.0, + "step": 4109 + }, + { + "epoch": 0.45135075774214806, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.072192907333374, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.7088727951049805, + "num_tokens": 106235476.0, + "step": 4110 + }, + { + "epoch": 0.4514605754447617, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.1683952808380127, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7095663547515869, + "num_tokens": 106261962.0, + "step": 4111 + }, + { + "epoch": 0.45157039314737535, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.313886880874634, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.68295818567276, + "num_tokens": 106288609.0, + "step": 4112 + }, + { + "epoch": 0.451680210849989, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.0934720039367676, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.714111328125, + "num_tokens": 106316201.0, + "step": 4113 + }, + { + "epoch": 0.4517900285526027, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 1.9610544443130493, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6866084337234497, + "num_tokens": 106349835.0, + "step": 4114 + }, + { + "epoch": 0.45189984625521634, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.1609785556793213, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6895552277565002, + "num_tokens": 106376203.0, + "step": 4115 + }, + { + "epoch": 0.45200966395783, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.1278393268585205, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.701683759689331, + "num_tokens": 106404493.0, + "step": 4116 + }, + { + "epoch": 0.4521194816604437, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.0371503829956055, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7013422250747681, + "num_tokens": 106432988.0, + "step": 4117 + }, + { + "epoch": 0.45222929936305734, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.566666841506958, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6954630017280579, + "num_tokens": 106453120.0, + "step": 4118 + }, + { + "epoch": 0.452339117065671, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.2960853576660156, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.715580940246582, + "num_tokens": 106475765.0, + "step": 4119 + }, + { + "epoch": 0.4524489347682846, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.2078189849853516, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7114771604537964, + "num_tokens": 106501456.0, + "step": 4120 + }, + { + "epoch": 0.45255875247089833, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.0252952575683594, + "learning_rate": 1e-06, + "loss": 1.0934, + "mean_token_accuracy": 0.6695916056632996, + "num_tokens": 106534135.0, + "step": 4121 + }, + { + "epoch": 0.452668570173512, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.2971808910369873, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7163212895393372, + "num_tokens": 106558166.0, + "step": 4122 + }, + { + "epoch": 0.4527783878761256, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.141216993331909, + "learning_rate": 1e-06, + "loss": 1.0213, + "mean_token_accuracy": 0.6896596550941467, + "num_tokens": 106585004.0, + "step": 4123 + }, + { + "epoch": 0.45288820557873927, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.1804096698760986, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.728270411491394, + "num_tokens": 106609383.0, + "step": 4124 + }, + { + "epoch": 0.45299802328135297, + "ewc_loss": 3.6716461181640625e-05, + "grad_norm": 2.044187545776367, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.697899341583252, + "num_tokens": 106641061.0, + "step": 4125 + }, + { + "epoch": 0.4531078409839666, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 1.9060132503509521, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7098032236099243, + "num_tokens": 106672179.0, + "step": 4126 + }, + { + "epoch": 0.45321765868658026, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.391477108001709, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6935962438583374, + "num_tokens": 106695512.0, + "step": 4127 + }, + { + "epoch": 0.45332747638919396, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 1.9503953456878662, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.728298544883728, + "num_tokens": 106723126.0, + "step": 4128 + }, + { + "epoch": 0.4534372940918076, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.3113672733306885, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7056623697280884, + "num_tokens": 106747737.0, + "step": 4129 + }, + { + "epoch": 0.45354711179442125, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 1.9979848861694336, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6932365298271179, + "num_tokens": 106778830.0, + "step": 4130 + }, + { + "epoch": 0.4536569294970349, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.211596965789795, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6926873922348022, + "num_tokens": 106803825.0, + "step": 4131 + }, + { + "epoch": 0.4537667471996486, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 1.912476658821106, + "learning_rate": 1e-06, + "loss": 1.0965, + "mean_token_accuracy": 0.6772127151489258, + "num_tokens": 106836896.0, + "step": 4132 + }, + { + "epoch": 0.45387656490226225, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.181748867034912, + "learning_rate": 1e-06, + "loss": 1.077, + "mean_token_accuracy": 0.6767751574516296, + "num_tokens": 106862131.0, + "step": 4133 + }, + { + "epoch": 0.4539863826048759, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.1919162273406982, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.7020277976989746, + "num_tokens": 106888132.0, + "step": 4134 + }, + { + "epoch": 0.4540962003074896, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.0998263359069824, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7111511826515198, + "num_tokens": 106916637.0, + "step": 4135 + }, + { + "epoch": 0.45420601801010324, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.1718950271606445, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.70073401927948, + "num_tokens": 106943004.0, + "step": 4136 + }, + { + "epoch": 0.4543158357127169, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.0851006507873535, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6966901421546936, + "num_tokens": 106972634.0, + "step": 4137 + }, + { + "epoch": 0.45442565341533053, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.391390323638916, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6886963844299316, + "num_tokens": 106995022.0, + "step": 4138 + }, + { + "epoch": 0.45453547111794423, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.5534827709198, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7200702428817749, + "num_tokens": 107014998.0, + "step": 4139 + }, + { + "epoch": 0.4546452888205579, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 1.9881662130355835, + "learning_rate": 1e-06, + "loss": 1.0671, + "mean_token_accuracy": 0.6780519485473633, + "num_tokens": 107047636.0, + "step": 4140 + }, + { + "epoch": 0.4547551065231715, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.3783068656921387, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7075541615486145, + "num_tokens": 107073075.0, + "step": 4141 + }, + { + "epoch": 0.45486492422578517, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.4320523738861084, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.7092053294181824, + "num_tokens": 107095074.0, + "step": 4142 + }, + { + "epoch": 0.4549747419283989, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.2286481857299805, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.6965938210487366, + "num_tokens": 107120478.0, + "step": 4143 + }, + { + "epoch": 0.4550845596310125, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.1142067909240723, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.7007012367248535, + "num_tokens": 107149935.0, + "step": 4144 + }, + { + "epoch": 0.45519437733362617, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.318263053894043, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7020424008369446, + "num_tokens": 107176438.0, + "step": 4145 + }, + { + "epoch": 0.45530419503623987, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.0730481147766113, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7062639594078064, + "num_tokens": 107202811.0, + "step": 4146 + }, + { + "epoch": 0.4554140127388535, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.082475423812866, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6931398510932922, + "num_tokens": 107233423.0, + "step": 4147 + }, + { + "epoch": 0.45552383044146716, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.150984525680542, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.7008135318756104, + "num_tokens": 107259979.0, + "step": 4148 + }, + { + "epoch": 0.4556336481440808, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.280532121658325, + "learning_rate": 1e-06, + "loss": 1.0577, + "mean_token_accuracy": 0.6861335039138794, + "num_tokens": 107289683.0, + "step": 4149 + }, + { + "epoch": 0.4557434658466945, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.27982497215271, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7059948444366455, + "num_tokens": 107315025.0, + "step": 4150 + }, + { + "epoch": 0.45585328354930815, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.460371732711792, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7018184661865234, + "num_tokens": 107337420.0, + "step": 4151 + }, + { + "epoch": 0.4559631012519218, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.4439046382904053, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6974179744720459, + "num_tokens": 107360034.0, + "step": 4152 + }, + { + "epoch": 0.45607291895453544, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.06453275680542, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6859322786331177, + "num_tokens": 107389164.0, + "step": 4153 + }, + { + "epoch": 0.45618273665714915, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.223036050796509, + "learning_rate": 1e-06, + "loss": 1.1103, + "mean_token_accuracy": 0.6750933527946472, + "num_tokens": 107417239.0, + "step": 4154 + }, + { + "epoch": 0.4562925543597628, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.3158836364746094, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6946208477020264, + "num_tokens": 107442241.0, + "step": 4155 + }, + { + "epoch": 0.45640237206237644, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.2728283405303955, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7021058797836304, + "num_tokens": 107464525.0, + "step": 4156 + }, + { + "epoch": 0.45651218976499014, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.4196853637695312, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7179950475692749, + "num_tokens": 107487863.0, + "step": 4157 + }, + { + "epoch": 0.4566220074676038, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.0792160034179688, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6963045597076416, + "num_tokens": 107516463.0, + "step": 4158 + }, + { + "epoch": 0.45673182517021743, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.044788360595703, + "learning_rate": 1e-06, + "loss": 1.0498, + "mean_token_accuracy": 0.6838216185569763, + "num_tokens": 107547372.0, + "step": 4159 + }, + { + "epoch": 0.4568416428728311, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.364994764328003, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7077746391296387, + "num_tokens": 107570039.0, + "step": 4160 + }, + { + "epoch": 0.4569514605754448, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.2968361377716064, + "learning_rate": 1e-06, + "loss": 1.0351, + "mean_token_accuracy": 0.6917917728424072, + "num_tokens": 107595999.0, + "step": 4161 + }, + { + "epoch": 0.4570612782780584, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.0670595169067383, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6975655555725098, + "num_tokens": 107624303.0, + "step": 4162 + }, + { + "epoch": 0.45717109598067207, + "ewc_loss": 3.695487976074219e-05, + "grad_norm": 2.1572844982147217, + "learning_rate": 1e-06, + "loss": 1.0659, + "mean_token_accuracy": 0.6863045692443848, + "num_tokens": 107654015.0, + "step": 4163 + }, + { + "epoch": 0.4572809136832858, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.04624605178833, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7328891754150391, + "num_tokens": 107681066.0, + "step": 4164 + }, + { + "epoch": 0.4573907313858994, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.3611385822296143, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7166971564292908, + "num_tokens": 107704738.0, + "step": 4165 + }, + { + "epoch": 0.45750054908851306, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.019911766052246, + "learning_rate": 1e-06, + "loss": 1.0723, + "mean_token_accuracy": 0.6779236197471619, + "num_tokens": 107738308.0, + "step": 4166 + }, + { + "epoch": 0.4576103667911267, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.016526699066162, + "learning_rate": 1e-06, + "loss": 1.0366, + "mean_token_accuracy": 0.6849952340126038, + "num_tokens": 107770284.0, + "step": 4167 + }, + { + "epoch": 0.4577201844937404, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.0811927318573, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6890861988067627, + "num_tokens": 107800283.0, + "step": 4168 + }, + { + "epoch": 0.45783000219635406, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.2668495178222656, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7206459045410156, + "num_tokens": 107824988.0, + "step": 4169 + }, + { + "epoch": 0.4579398198989677, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.3559107780456543, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7229634523391724, + "num_tokens": 107847507.0, + "step": 4170 + }, + { + "epoch": 0.45804963760158135, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 1.9977773427963257, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6947669982910156, + "num_tokens": 107878874.0, + "step": 4171 + }, + { + "epoch": 0.45815945530419505, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.0466949939727783, + "learning_rate": 1e-06, + "loss": 1.066, + "mean_token_accuracy": 0.6915486454963684, + "num_tokens": 107907987.0, + "step": 4172 + }, + { + "epoch": 0.4582692730068087, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 1.978710651397705, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6935545802116394, + "num_tokens": 107935921.0, + "step": 4173 + }, + { + "epoch": 0.45837909070942234, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 1.9679661989212036, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.700840950012207, + "num_tokens": 107969167.0, + "step": 4174 + }, + { + "epoch": 0.45848890841203604, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.2577314376831055, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7126626968383789, + "num_tokens": 107993529.0, + "step": 4175 + }, + { + "epoch": 0.4585987261146497, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.293287992477417, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6844985485076904, + "num_tokens": 108019805.0, + "step": 4176 + }, + { + "epoch": 0.45870854381726334, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.019474983215332, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.692138671875, + "num_tokens": 108048592.0, + "step": 4177 + }, + { + "epoch": 0.458818361519877, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 1.9673713445663452, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7172253131866455, + "num_tokens": 108079410.0, + "step": 4178 + }, + { + "epoch": 0.4589281792224907, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.188220977783203, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.7133074402809143, + "num_tokens": 108106842.0, + "step": 4179 + }, + { + "epoch": 0.45903799692510433, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.1336522102355957, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.699947714805603, + "num_tokens": 108134819.0, + "step": 4180 + }, + { + "epoch": 0.459147814627718, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.235029935836792, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7117476463317871, + "num_tokens": 108160211.0, + "step": 4181 + }, + { + "epoch": 0.4592576323303316, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.3238344192504883, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7056379914283752, + "num_tokens": 108184231.0, + "step": 4182 + }, + { + "epoch": 0.4593674500329453, + "ewc_loss": 3.719329833984375e-05, + "grad_norm": 2.3287954330444336, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.7002325057983398, + "num_tokens": 108206916.0, + "step": 4183 + }, + { + "epoch": 0.45947726773555897, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.357285499572754, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7145084142684937, + "num_tokens": 108227851.0, + "step": 4184 + }, + { + "epoch": 0.4595870854381726, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.190018892288208, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6950103044509888, + "num_tokens": 108254622.0, + "step": 4185 + }, + { + "epoch": 0.4596969031407863, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1148035526275635, + "learning_rate": 1e-06, + "loss": 1.041, + "mean_token_accuracy": 0.698986291885376, + "num_tokens": 108282366.0, + "step": 4186 + }, + { + "epoch": 0.45980672084339996, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 1.9960558414459229, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.6896374821662903, + "num_tokens": 108314359.0, + "step": 4187 + }, + { + "epoch": 0.4599165385460136, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.0005862712860107, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7010515928268433, + "num_tokens": 108342960.0, + "step": 4188 + }, + { + "epoch": 0.46002635624862726, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.103463649749756, + "learning_rate": 1e-06, + "loss": 1.0345, + "mean_token_accuracy": 0.6888071298599243, + "num_tokens": 108369224.0, + "step": 4189 + }, + { + "epoch": 0.46013617395124096, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.140760898590088, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6927758455276489, + "num_tokens": 108397077.0, + "step": 4190 + }, + { + "epoch": 0.4602459916538546, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.216534376144409, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6964472532272339, + "num_tokens": 108421690.0, + "step": 4191 + }, + { + "epoch": 0.46035580935646825, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1474759578704834, + "learning_rate": 1e-06, + "loss": 1.0636, + "mean_token_accuracy": 0.6792463660240173, + "num_tokens": 108449768.0, + "step": 4192 + }, + { + "epoch": 0.46046562705908195, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.8650171756744385, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7065478563308716, + "num_tokens": 108466772.0, + "step": 4193 + }, + { + "epoch": 0.4605754447616956, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.4123148918151855, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7220505475997925, + "num_tokens": 108488184.0, + "step": 4194 + }, + { + "epoch": 0.46068526246430924, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1372034549713135, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6982893347740173, + "num_tokens": 108514984.0, + "step": 4195 + }, + { + "epoch": 0.4607950801669229, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1342997550964355, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7113569974899292, + "num_tokens": 108540998.0, + "step": 4196 + }, + { + "epoch": 0.4609048978695366, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.0083675384521484, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7071474194526672, + "num_tokens": 108570211.0, + "step": 4197 + }, + { + "epoch": 0.46101471557215024, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.0228729248046875, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6786991953849792, + "num_tokens": 108599725.0, + "step": 4198 + }, + { + "epoch": 0.4611245332747639, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.29091477394104, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6930767893791199, + "num_tokens": 108624771.0, + "step": 4199 + }, + { + "epoch": 0.4612343509773775, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1230180263519287, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6902244091033936, + "num_tokens": 108655965.0, + "step": 4200 + }, + { + "epoch": 0.46134416867999123, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3643887042999268, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7018927335739136, + "num_tokens": 108678648.0, + "step": 4201 + }, + { + "epoch": 0.4614539863826049, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1231045722961426, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6920835971832275, + "num_tokens": 108708088.0, + "step": 4202 + }, + { + "epoch": 0.4615638040852185, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 1.9491595029830933, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.712820291519165, + "num_tokens": 108737497.0, + "step": 4203 + }, + { + "epoch": 0.4616736217878322, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.186948537826538, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7090804576873779, + "num_tokens": 108764609.0, + "step": 4204 + }, + { + "epoch": 0.46178343949044587, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.0539238452911377, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.702345073223114, + "num_tokens": 108790774.0, + "step": 4205 + }, + { + "epoch": 0.4618932571930595, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 1.9963420629501343, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7177252769470215, + "num_tokens": 108820188.0, + "step": 4206 + }, + { + "epoch": 0.46200307489567316, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.4371142387390137, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7253418564796448, + "num_tokens": 108841518.0, + "step": 4207 + }, + { + "epoch": 0.46211289259828686, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3934733867645264, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6932276487350464, + "num_tokens": 108865351.0, + "step": 4208 + }, + { + "epoch": 0.4622227103009005, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.5539164543151855, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.689866304397583, + "num_tokens": 108887902.0, + "step": 4209 + }, + { + "epoch": 0.46233252800351415, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 1.9928144216537476, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7061688303947449, + "num_tokens": 108917553.0, + "step": 4210 + }, + { + "epoch": 0.46244234570612786, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 1.9509997367858887, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6989691257476807, + "num_tokens": 108949999.0, + "step": 4211 + }, + { + "epoch": 0.4625521634087415, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.2407732009887695, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.7051246166229248, + "num_tokens": 108976054.0, + "step": 4212 + }, + { + "epoch": 0.46266198111135515, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1694748401641846, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7157412767410278, + "num_tokens": 109001853.0, + "step": 4213 + }, + { + "epoch": 0.4627717988139688, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.7402701377868652, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7332061529159546, + "num_tokens": 109018515.0, + "step": 4214 + }, + { + "epoch": 0.4628816165165825, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3189492225646973, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7070077657699585, + "num_tokens": 109043524.0, + "step": 4215 + }, + { + "epoch": 0.46299143421919614, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3703062534332275, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6943097114562988, + "num_tokens": 109065870.0, + "step": 4216 + }, + { + "epoch": 0.4631012519218098, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.099001884460449, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6900616884231567, + "num_tokens": 109092819.0, + "step": 4217 + }, + { + "epoch": 0.46321106962442343, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.138526201248169, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7164175510406494, + "num_tokens": 109119366.0, + "step": 4218 + }, + { + "epoch": 0.46332088732703713, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.5105464458465576, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6988271474838257, + "num_tokens": 109142027.0, + "step": 4219 + }, + { + "epoch": 0.4634307050296508, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3897361755371094, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7029115557670593, + "num_tokens": 109164566.0, + "step": 4220 + }, + { + "epoch": 0.4635405227322644, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.5182580947875977, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7255845665931702, + "num_tokens": 109185471.0, + "step": 4221 + }, + { + "epoch": 0.4636503404348781, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3231842517852783, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7017303109169006, + "num_tokens": 109209593.0, + "step": 4222 + }, + { + "epoch": 0.4637601581374918, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3273983001708984, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6860071420669556, + "num_tokens": 109233085.0, + "step": 4223 + }, + { + "epoch": 0.4638699758401054, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.3820812702178955, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7104381322860718, + "num_tokens": 109255439.0, + "step": 4224 + }, + { + "epoch": 0.46397979354271907, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.1543588638305664, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6935780048370361, + "num_tokens": 109285240.0, + "step": 4225 + }, + { + "epoch": 0.46408961124533277, + "ewc_loss": 3.7670135498046875e-05, + "grad_norm": 2.522550344467163, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7175118327140808, + "num_tokens": 109306534.0, + "step": 4226 + }, + { + "epoch": 0.4641994289479464, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.1621811389923096, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6921627521514893, + "num_tokens": 109335144.0, + "step": 4227 + }, + { + "epoch": 0.46430924665056006, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.0339484214782715, + "learning_rate": 1e-06, + "loss": 1.0394, + "mean_token_accuracy": 0.6946717500686646, + "num_tokens": 109363223.0, + "step": 4228 + }, + { + "epoch": 0.4644190643531737, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.35575532913208, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7100361585617065, + "num_tokens": 109385976.0, + "step": 4229 + }, + { + "epoch": 0.4645288820557874, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.1885695457458496, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6941289901733398, + "num_tokens": 109413991.0, + "step": 4230 + }, + { + "epoch": 0.46463869975840105, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.4036333560943604, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7222874760627747, + "num_tokens": 109434389.0, + "step": 4231 + }, + { + "epoch": 0.4647485174610147, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 1.9544371366500854, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6901635527610779, + "num_tokens": 109466391.0, + "step": 4232 + }, + { + "epoch": 0.4648583351636284, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.335456609725952, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7113730907440186, + "num_tokens": 109489623.0, + "step": 4233 + }, + { + "epoch": 0.46496815286624205, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.1894643306732178, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6897612810134888, + "num_tokens": 109515780.0, + "step": 4234 + }, + { + "epoch": 0.4650779705688557, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.2910516262054443, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7069146633148193, + "num_tokens": 109539212.0, + "step": 4235 + }, + { + "epoch": 0.46518778827146934, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.5044455528259277, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.6880414485931396, + "num_tokens": 109560287.0, + "step": 4236 + }, + { + "epoch": 0.46529760597408304, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.057643413543701, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7050485610961914, + "num_tokens": 109588456.0, + "step": 4237 + }, + { + "epoch": 0.4654074236766967, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.1226084232330322, + "learning_rate": 1e-06, + "loss": 1.0699, + "mean_token_accuracy": 0.6823610067367554, + "num_tokens": 109616581.0, + "step": 4238 + }, + { + "epoch": 0.46551724137931033, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.336634635925293, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7195102572441101, + "num_tokens": 109641331.0, + "step": 4239 + }, + { + "epoch": 0.46562705908192403, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.030266523361206, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6907185316085815, + "num_tokens": 109670886.0, + "step": 4240 + }, + { + "epoch": 0.4657368767845377, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.149405002593994, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6941584348678589, + "num_tokens": 109697801.0, + "step": 4241 + }, + { + "epoch": 0.4658466944871513, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.163093328475952, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7041727304458618, + "num_tokens": 109724159.0, + "step": 4242 + }, + { + "epoch": 0.46595651218976497, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.317896604537964, + "learning_rate": 1e-06, + "loss": 1.0672, + "mean_token_accuracy": 0.671873152256012, + "num_tokens": 109749383.0, + "step": 4243 + }, + { + "epoch": 0.4660663298923787, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.22537899017334, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7086695432662964, + "num_tokens": 109776098.0, + "step": 4244 + }, + { + "epoch": 0.4661761475949923, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.35994291305542, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7114249467849731, + "num_tokens": 109797559.0, + "step": 4245 + }, + { + "epoch": 0.46628596529760596, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.004262685775757, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7297143936157227, + "num_tokens": 109828678.0, + "step": 4246 + }, + { + "epoch": 0.4663957830002196, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.101538896560669, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6827788949012756, + "num_tokens": 109857652.0, + "step": 4247 + }, + { + "epoch": 0.4665056007028333, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.15682315826416, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6995038390159607, + "num_tokens": 109884554.0, + "step": 4248 + }, + { + "epoch": 0.46661541840544696, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.112407922744751, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7169936895370483, + "num_tokens": 109911682.0, + "step": 4249 + }, + { + "epoch": 0.4667252361080606, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.3167436122894287, + "learning_rate": 1e-06, + "loss": 1.0696, + "mean_token_accuracy": 0.6851204633712769, + "num_tokens": 109936157.0, + "step": 4250 + }, + { + "epoch": 0.4668350538106743, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.045732259750366, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.6949356198310852, + "num_tokens": 109963461.0, + "step": 4251 + }, + { + "epoch": 0.46694487151328795, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.3567497730255127, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7266371846199036, + "num_tokens": 109985239.0, + "step": 4252 + }, + { + "epoch": 0.4670546892159016, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.2352936267852783, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6897786259651184, + "num_tokens": 110009774.0, + "step": 4253 + }, + { + "epoch": 0.46716450691851524, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.4881341457366943, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.7020034790039062, + "num_tokens": 110030532.0, + "step": 4254 + }, + { + "epoch": 0.46727432462112894, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 1.910196304321289, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6881440877914429, + "num_tokens": 110064065.0, + "step": 4255 + }, + { + "epoch": 0.4673841423237426, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.1794559955596924, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.703517496585846, + "num_tokens": 110089461.0, + "step": 4256 + }, + { + "epoch": 0.46749396002635624, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 1.8760653734207153, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6960418224334717, + "num_tokens": 110124072.0, + "step": 4257 + }, + { + "epoch": 0.4676037777289699, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.4922285079956055, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.7051723599433899, + "num_tokens": 110145663.0, + "step": 4258 + }, + { + "epoch": 0.4677135954315836, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.116750478744507, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6880713701248169, + "num_tokens": 110175157.0, + "step": 4259 + }, + { + "epoch": 0.46782341313419723, + "ewc_loss": 3.790855407714844e-05, + "grad_norm": 2.367486000061035, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7067064642906189, + "num_tokens": 110198393.0, + "step": 4260 + }, + { + "epoch": 0.4679332308368109, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 1.9186244010925293, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6961449384689331, + "num_tokens": 110231770.0, + "step": 4261 + }, + { + "epoch": 0.4680430485394246, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.105619430541992, + "learning_rate": 1e-06, + "loss": 1.0668, + "mean_token_accuracy": 0.6853175163269043, + "num_tokens": 110261755.0, + "step": 4262 + }, + { + "epoch": 0.4681528662420382, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.5265586376190186, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6881347894668579, + "num_tokens": 110281687.0, + "step": 4263 + }, + { + "epoch": 0.46826268394465187, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.3573882579803467, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7142349481582642, + "num_tokens": 110305069.0, + "step": 4264 + }, + { + "epoch": 0.4683725016472655, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.3931806087493896, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7380629181861877, + "num_tokens": 110324935.0, + "step": 4265 + }, + { + "epoch": 0.4684823193498792, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.4404122829437256, + "learning_rate": 1e-06, + "loss": 1.0277, + "mean_token_accuracy": 0.6951260566711426, + "num_tokens": 110346955.0, + "step": 4266 + }, + { + "epoch": 0.46859213705249286, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.075697660446167, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7145625352859497, + "num_tokens": 110373544.0, + "step": 4267 + }, + { + "epoch": 0.4687019547551065, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.041215419769287, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6848775148391724, + "num_tokens": 110405117.0, + "step": 4268 + }, + { + "epoch": 0.4688117724577202, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.4027440547943115, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6962673664093018, + "num_tokens": 110426767.0, + "step": 4269 + }, + { + "epoch": 0.46892159016033386, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.3417460918426514, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6909312009811401, + "num_tokens": 110450394.0, + "step": 4270 + }, + { + "epoch": 0.4690314078629475, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.0862207412719727, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6850938200950623, + "num_tokens": 110479240.0, + "step": 4271 + }, + { + "epoch": 0.46914122556556115, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.2410190105438232, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7092710733413696, + "num_tokens": 110503201.0, + "step": 4272 + }, + { + "epoch": 0.46925104326817485, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.1204378604888916, + "learning_rate": 1e-06, + "loss": 1.0461, + "mean_token_accuracy": 0.6871898770332336, + "num_tokens": 110533873.0, + "step": 4273 + }, + { + "epoch": 0.4693608609707885, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.5840587615966797, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7125317454338074, + "num_tokens": 110553795.0, + "step": 4274 + }, + { + "epoch": 0.46947067867340214, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.140530824661255, + "learning_rate": 1e-06, + "loss": 1.0535, + "mean_token_accuracy": 0.6841377019882202, + "num_tokens": 110582884.0, + "step": 4275 + }, + { + "epoch": 0.4695804963760158, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.113258123397827, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7067351341247559, + "num_tokens": 110608619.0, + "step": 4276 + }, + { + "epoch": 0.4696903140786295, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.26918888092041, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7081034183502197, + "num_tokens": 110633119.0, + "step": 4277 + }, + { + "epoch": 0.46980013178124314, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.2148916721343994, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6886869072914124, + "num_tokens": 110658497.0, + "step": 4278 + }, + { + "epoch": 0.4699099494838568, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.246870517730713, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7119728922843933, + "num_tokens": 110682326.0, + "step": 4279 + }, + { + "epoch": 0.4700197671864705, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.116325855255127, + "learning_rate": 1e-06, + "loss": 1.0623, + "mean_token_accuracy": 0.6883890628814697, + "num_tokens": 110712977.0, + "step": 4280 + }, + { + "epoch": 0.47012958488908413, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.3858718872070312, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.693504810333252, + "num_tokens": 110735079.0, + "step": 4281 + }, + { + "epoch": 0.4702394025916978, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.128829002380371, + "learning_rate": 1e-06, + "loss": 1.0349, + "mean_token_accuracy": 0.6896340250968933, + "num_tokens": 110763270.0, + "step": 4282 + }, + { + "epoch": 0.4703492202943114, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.251051425933838, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6939364671707153, + "num_tokens": 110789102.0, + "step": 4283 + }, + { + "epoch": 0.4704590379969251, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 1.989501714706421, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7179679870605469, + "num_tokens": 110820768.0, + "step": 4284 + }, + { + "epoch": 0.47056885569953877, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.024365186691284, + "learning_rate": 1e-06, + "loss": 1.0882, + "mean_token_accuracy": 0.6741117238998413, + "num_tokens": 110852042.0, + "step": 4285 + }, + { + "epoch": 0.4706786734021524, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.2690045833587646, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.71802818775177, + "num_tokens": 110877114.0, + "step": 4286 + }, + { + "epoch": 0.4707884911047661, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 1.9963812828063965, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7409091591835022, + "num_tokens": 110905762.0, + "step": 4287 + }, + { + "epoch": 0.47089830880737976, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.419027805328369, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.6981850862503052, + "num_tokens": 110926649.0, + "step": 4288 + }, + { + "epoch": 0.4710081265099934, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.3127593994140625, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.693042516708374, + "num_tokens": 110952555.0, + "step": 4289 + }, + { + "epoch": 0.47111794421260705, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.2517073154449463, + "learning_rate": 1e-06, + "loss": 1.0385, + "mean_token_accuracy": 0.6904481649398804, + "num_tokens": 110978068.0, + "step": 4290 + }, + { + "epoch": 0.47122776191522076, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.48885178565979, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.7116732597351074, + "num_tokens": 111000492.0, + "step": 4291 + }, + { + "epoch": 0.4713375796178344, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.0685176849365234, + "learning_rate": 1e-06, + "loss": 1.0351, + "mean_token_accuracy": 0.6880811452865601, + "num_tokens": 111031966.0, + "step": 4292 + }, + { + "epoch": 0.47144739732044805, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.1659622192382812, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6841436624526978, + "num_tokens": 111060055.0, + "step": 4293 + }, + { + "epoch": 0.4715572150230617, + "ewc_loss": 3.838539123535156e-05, + "grad_norm": 2.0777838230133057, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6911485195159912, + "num_tokens": 111088646.0, + "step": 4294 + }, + { + "epoch": 0.4716670327256754, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.2787957191467285, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7067333459854126, + "num_tokens": 111113189.0, + "step": 4295 + }, + { + "epoch": 0.47177685042828904, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.3082289695739746, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7045808434486389, + "num_tokens": 111136436.0, + "step": 4296 + }, + { + "epoch": 0.4718866681309027, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.198166847229004, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.710761547088623, + "num_tokens": 111162183.0, + "step": 4297 + }, + { + "epoch": 0.4719964858335164, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.104301691055298, + "learning_rate": 1e-06, + "loss": 1.0317, + "mean_token_accuracy": 0.6879202127456665, + "num_tokens": 111191417.0, + "step": 4298 + }, + { + "epoch": 0.47210630353613003, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.045203685760498, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.704825758934021, + "num_tokens": 111219309.0, + "step": 4299 + }, + { + "epoch": 0.4722161212387437, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.1090967655181885, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7176120281219482, + "num_tokens": 111247213.0, + "step": 4300 + }, + { + "epoch": 0.4723259389413573, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.4101402759552, + "learning_rate": 1e-06, + "loss": 1.0345, + "mean_token_accuracy": 0.6882244944572449, + "num_tokens": 111270004.0, + "step": 4301 + }, + { + "epoch": 0.472435756643971, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.313021659851074, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7100227475166321, + "num_tokens": 111294602.0, + "step": 4302 + }, + { + "epoch": 0.4725455743465847, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.2630841732025146, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6889104843139648, + "num_tokens": 111321839.0, + "step": 4303 + }, + { + "epoch": 0.4726553920491983, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.2555084228515625, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6889288425445557, + "num_tokens": 111348923.0, + "step": 4304 + }, + { + "epoch": 0.47276520975181197, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 1.9577218294143677, + "learning_rate": 1e-06, + "loss": 1.0264, + "mean_token_accuracy": 0.6912393569946289, + "num_tokens": 111378979.0, + "step": 4305 + }, + { + "epoch": 0.47287502745442567, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.428119659423828, + "learning_rate": 1e-06, + "loss": 1.0497, + "mean_token_accuracy": 0.6989374160766602, + "num_tokens": 111402072.0, + "step": 4306 + }, + { + "epoch": 0.4729848451570393, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.331239938735962, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7088629603385925, + "num_tokens": 111424366.0, + "step": 4307 + }, + { + "epoch": 0.47309466285965296, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 1.9124259948730469, + "learning_rate": 1e-06, + "loss": 1.0715, + "mean_token_accuracy": 0.6810123324394226, + "num_tokens": 111457727.0, + "step": 4308 + }, + { + "epoch": 0.47320448056226666, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.3792309761047363, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7143774628639221, + "num_tokens": 111479473.0, + "step": 4309 + }, + { + "epoch": 0.4733142982648803, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.199950933456421, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6863826513290405, + "num_tokens": 111505459.0, + "step": 4310 + }, + { + "epoch": 0.47342411596749395, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.1679415702819824, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6954754590988159, + "num_tokens": 111529820.0, + "step": 4311 + }, + { + "epoch": 0.4735339336701076, + "ewc_loss": 3.8623809814453125e-05, + "grad_norm": 2.2782905101776123, + "learning_rate": 1e-06, + "loss": 1.0433, + "mean_token_accuracy": 0.6995633244514465, + "num_tokens": 111556109.0, + "step": 4312 + }, + { + "epoch": 0.4736437513727213, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.2180497646331787, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7095385789871216, + "num_tokens": 111580721.0, + "step": 4313 + }, + { + "epoch": 0.47375356907533495, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2889835834503174, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7144883871078491, + "num_tokens": 111604459.0, + "step": 4314 + }, + { + "epoch": 0.4738633867779486, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.3162145614624023, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7040425539016724, + "num_tokens": 111629930.0, + "step": 4315 + }, + { + "epoch": 0.4739732044805623, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.100637674331665, + "learning_rate": 1e-06, + "loss": 1.0631, + "mean_token_accuracy": 0.6841330528259277, + "num_tokens": 111657613.0, + "step": 4316 + }, + { + "epoch": 0.47408302218317594, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.5246851444244385, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7271764874458313, + "num_tokens": 111677344.0, + "step": 4317 + }, + { + "epoch": 0.4741928398857896, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.2840256690979004, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7151528596878052, + "num_tokens": 111701017.0, + "step": 4318 + }, + { + "epoch": 0.47430265758840323, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.4989383220672607, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7182756662368774, + "num_tokens": 111721446.0, + "step": 4319 + }, + { + "epoch": 0.47441247529101693, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.140122413635254, + "learning_rate": 1e-06, + "loss": 1.056, + "mean_token_accuracy": 0.6863926649093628, + "num_tokens": 111749768.0, + "step": 4320 + }, + { + "epoch": 0.4745222929936306, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.121157646179199, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6990385055541992, + "num_tokens": 111780595.0, + "step": 4321 + }, + { + "epoch": 0.4746321106962442, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.0477871894836426, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.69626384973526, + "num_tokens": 111808453.0, + "step": 4322 + }, + { + "epoch": 0.47474192839885787, + "ewc_loss": 3.886222839355469e-05, + "grad_norm": 2.073554039001465, + "learning_rate": 1e-06, + "loss": 1.0685, + "mean_token_accuracy": 0.678704559803009, + "num_tokens": 111838333.0, + "step": 4323 + }, + { + "epoch": 0.4748517461014716, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2751123905181885, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7027711868286133, + "num_tokens": 111862836.0, + "step": 4324 + }, + { + "epoch": 0.4749615638040852, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.248811960220337, + "learning_rate": 1e-06, + "loss": 1.0973, + "mean_token_accuracy": 0.6742401123046875, + "num_tokens": 111889887.0, + "step": 4325 + }, + { + "epoch": 0.47507138150669886, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2796006202697754, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6901350021362305, + "num_tokens": 111916363.0, + "step": 4326 + }, + { + "epoch": 0.47518119920931257, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.0225017070770264, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7152667045593262, + "num_tokens": 111947248.0, + "step": 4327 + }, + { + "epoch": 0.4752910169119262, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.9999421834945679, + "learning_rate": 1e-06, + "loss": 1.0685, + "mean_token_accuracy": 0.6873008012771606, + "num_tokens": 111979006.0, + "step": 4328 + }, + { + "epoch": 0.47540083461453986, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2659177780151367, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7195661067962646, + "num_tokens": 112001590.0, + "step": 4329 + }, + { + "epoch": 0.4755106523171535, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.5152556896209717, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7112252712249756, + "num_tokens": 112022137.0, + "step": 4330 + }, + { + "epoch": 0.4756204700197672, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.211418390274048, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7285749316215515, + "num_tokens": 112047019.0, + "step": 4331 + }, + { + "epoch": 0.47573028772238085, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.1293487548828125, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6927272081375122, + "num_tokens": 112073720.0, + "step": 4332 + }, + { + "epoch": 0.4758401054249945, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.225536346435547, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.703392505645752, + "num_tokens": 112100905.0, + "step": 4333 + }, + { + "epoch": 0.47594992312760814, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.0681567192077637, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.704390287399292, + "num_tokens": 112127185.0, + "step": 4334 + }, + { + "epoch": 0.47605974083022184, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.3718252182006836, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7166380286216736, + "num_tokens": 112148344.0, + "step": 4335 + }, + { + "epoch": 0.4761695585328355, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2951982021331787, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6972013711929321, + "num_tokens": 112172753.0, + "step": 4336 + }, + { + "epoch": 0.47627937623544914, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2434394359588623, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.707769513130188, + "num_tokens": 112195835.0, + "step": 4337 + }, + { + "epoch": 0.47638919393806284, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.013049840927124, + "learning_rate": 1e-06, + "loss": 1.0778, + "mean_token_accuracy": 0.6767091751098633, + "num_tokens": 112225281.0, + "step": 4338 + }, + { + "epoch": 0.4764990116406765, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2583506107330322, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7028716206550598, + "num_tokens": 112249829.0, + "step": 4339 + }, + { + "epoch": 0.47660882934329013, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.994676947593689, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6928465962409973, + "num_tokens": 112279445.0, + "step": 4340 + }, + { + "epoch": 0.4767186470459038, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.0529422760009766, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6916342377662659, + "num_tokens": 112309786.0, + "step": 4341 + }, + { + "epoch": 0.4768284647485175, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.1140964031219482, + "learning_rate": 1e-06, + "loss": 1.0576, + "mean_token_accuracy": 0.6849100589752197, + "num_tokens": 112338737.0, + "step": 4342 + }, + { + "epoch": 0.4769382824511311, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.8124172687530518, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6854345798492432, + "num_tokens": 112376012.0, + "step": 4343 + }, + { + "epoch": 0.47704810015374477, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.5097482204437256, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6913289427757263, + "num_tokens": 112398477.0, + "step": 4344 + }, + { + "epoch": 0.47715791785635847, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.3254306316375732, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7039117813110352, + "num_tokens": 112421259.0, + "step": 4345 + }, + { + "epoch": 0.4772677355589721, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.322352647781372, + "learning_rate": 1e-06, + "loss": 1.039, + "mean_token_accuracy": 0.686295747756958, + "num_tokens": 112445827.0, + "step": 4346 + }, + { + "epoch": 0.47737755326158576, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.495351552963257, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.6835429072380066, + "num_tokens": 112468017.0, + "step": 4347 + }, + { + "epoch": 0.4774873709641994, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2436182498931885, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6801198720932007, + "num_tokens": 112493136.0, + "step": 4348 + }, + { + "epoch": 0.4775971886668131, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.1301279067993164, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6913471221923828, + "num_tokens": 112517619.0, + "step": 4349 + }, + { + "epoch": 0.47770700636942676, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.9839565753936768, + "learning_rate": 1e-06, + "loss": 1.1238, + "mean_token_accuracy": 0.6697536706924438, + "num_tokens": 112553523.0, + "step": 4350 + }, + { + "epoch": 0.4778168240720404, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2235257625579834, + "learning_rate": 1e-06, + "loss": 1.1596, + "mean_token_accuracy": 0.6577771902084351, + "num_tokens": 112581580.0, + "step": 4351 + }, + { + "epoch": 0.47792664177465405, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.059413433074951, + "learning_rate": 1e-06, + "loss": 1.0683, + "mean_token_accuracy": 0.6858820915222168, + "num_tokens": 112612031.0, + "step": 4352 + }, + { + "epoch": 0.47803645947726775, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.9413939714431763, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7076013088226318, + "num_tokens": 112642814.0, + "step": 4353 + }, + { + "epoch": 0.4781462771798814, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.4475796222686768, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7046886682510376, + "num_tokens": 112665706.0, + "step": 4354 + }, + { + "epoch": 0.47825609488249504, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2609167098999023, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6923750638961792, + "num_tokens": 112690358.0, + "step": 4355 + }, + { + "epoch": 0.47836591258510874, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 1.9841878414154053, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6963317394256592, + "num_tokens": 112719946.0, + "step": 4356 + }, + { + "epoch": 0.4784757302877224, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.2984259128570557, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6942697763442993, + "num_tokens": 112744322.0, + "step": 4357 + }, + { + "epoch": 0.47858554799033604, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.062296152114868, + "learning_rate": 1e-06, + "loss": 1.1016, + "mean_token_accuracy": 0.6730762720108032, + "num_tokens": 112775838.0, + "step": 4358 + }, + { + "epoch": 0.4786953656929497, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.063044548034668, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7128862738609314, + "num_tokens": 112803649.0, + "step": 4359 + }, + { + "epoch": 0.4788051833955634, + "ewc_loss": 3.910064697265625e-05, + "grad_norm": 2.161665678024292, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6842910647392273, + "num_tokens": 112831812.0, + "step": 4360 + }, + { + "epoch": 0.47891500109817703, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.110443592071533, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7242774367332458, + "num_tokens": 112858557.0, + "step": 4361 + }, + { + "epoch": 0.4790248188007907, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.0822808742523193, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7067243456840515, + "num_tokens": 112884800.0, + "step": 4362 + }, + { + "epoch": 0.4791346365034044, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1190412044525146, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7007167339324951, + "num_tokens": 112913848.0, + "step": 4363 + }, + { + "epoch": 0.479244454206018, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.219355583190918, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7060254812240601, + "num_tokens": 112939866.0, + "step": 4364 + }, + { + "epoch": 0.47935427190863167, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1437783241271973, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7361232042312622, + "num_tokens": 112965056.0, + "step": 4365 + }, + { + "epoch": 0.4794640896112453, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.0987210273742676, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.688052773475647, + "num_tokens": 112993346.0, + "step": 4366 + }, + { + "epoch": 0.479573907313859, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.212552547454834, + "learning_rate": 1e-06, + "loss": 1.0786, + "mean_token_accuracy": 0.6836715936660767, + "num_tokens": 113020155.0, + "step": 4367 + }, + { + "epoch": 0.47968372501647266, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1945090293884277, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.6963376402854919, + "num_tokens": 113044250.0, + "step": 4368 + }, + { + "epoch": 0.4797935427190863, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.120527505874634, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6963754296302795, + "num_tokens": 113072237.0, + "step": 4369 + }, + { + "epoch": 0.47990336042169995, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.2637248039245605, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7076111435890198, + "num_tokens": 113096600.0, + "step": 4370 + }, + { + "epoch": 0.48001317812431366, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 1.9162476062774658, + "learning_rate": 1e-06, + "loss": 1.0277, + "mean_token_accuracy": 0.6938979029655457, + "num_tokens": 113131476.0, + "step": 4371 + }, + { + "epoch": 0.4801229958269273, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.0857021808624268, + "learning_rate": 1e-06, + "loss": 1.0264, + "mean_token_accuracy": 0.6880303621292114, + "num_tokens": 113161614.0, + "step": 4372 + }, + { + "epoch": 0.48023281352954095, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.359870195388794, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6997870206832886, + "num_tokens": 113184723.0, + "step": 4373 + }, + { + "epoch": 0.48034263123215465, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.229313611984253, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.696390688419342, + "num_tokens": 113207737.0, + "step": 4374 + }, + { + "epoch": 0.4804524489347683, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1731512546539307, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6990652680397034, + "num_tokens": 113236517.0, + "step": 4375 + }, + { + "epoch": 0.48056226663738194, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.276571035385132, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7146413922309875, + "num_tokens": 113258955.0, + "step": 4376 + }, + { + "epoch": 0.4806720843399956, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.135267734527588, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.7009678483009338, + "num_tokens": 113287178.0, + "step": 4377 + }, + { + "epoch": 0.4807819020426093, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1686794757843018, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7066186666488647, + "num_tokens": 113314384.0, + "step": 4378 + }, + { + "epoch": 0.48089171974522293, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.486135482788086, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.6955490708351135, + "num_tokens": 113335624.0, + "step": 4379 + }, + { + "epoch": 0.4810015374478366, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.164703130722046, + "learning_rate": 1e-06, + "loss": 1.0622, + "mean_token_accuracy": 0.6820751428604126, + "num_tokens": 113364171.0, + "step": 4380 + }, + { + "epoch": 0.4811113551504502, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.4957385063171387, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7141746878623962, + "num_tokens": 113385080.0, + "step": 4381 + }, + { + "epoch": 0.4812211728530639, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1029443740844727, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.6976075172424316, + "num_tokens": 113413851.0, + "step": 4382 + }, + { + "epoch": 0.4813309905556776, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1258459091186523, + "learning_rate": 1e-06, + "loss": 1.0531, + "mean_token_accuracy": 0.6815375089645386, + "num_tokens": 113444131.0, + "step": 4383 + }, + { + "epoch": 0.4814408082582912, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.024664878845215, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6884284019470215, + "num_tokens": 113473839.0, + "step": 4384 + }, + { + "epoch": 0.4815506259609049, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.354219436645508, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6928136944770813, + "num_tokens": 113496652.0, + "step": 4385 + }, + { + "epoch": 0.48166044366351857, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.453575611114502, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7003028392791748, + "num_tokens": 113517813.0, + "step": 4386 + }, + { + "epoch": 0.4817702613661322, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1736223697662354, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6904327273368835, + "num_tokens": 113544589.0, + "step": 4387 + }, + { + "epoch": 0.48188007906874586, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.0340757369995117, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.697080135345459, + "num_tokens": 113575425.0, + "step": 4388 + }, + { + "epoch": 0.48198989677135956, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1285622119903564, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7146202325820923, + "num_tokens": 113600542.0, + "step": 4389 + }, + { + "epoch": 0.4820997144739732, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1600213050842285, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7087385058403015, + "num_tokens": 113626938.0, + "step": 4390 + }, + { + "epoch": 0.48220953217658685, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.2123992443084717, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6944718956947327, + "num_tokens": 113653288.0, + "step": 4391 + }, + { + "epoch": 0.48231934987920055, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.154094696044922, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7008166313171387, + "num_tokens": 113680097.0, + "step": 4392 + }, + { + "epoch": 0.4824291675818142, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.3990960121154785, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6935055255889893, + "num_tokens": 113703073.0, + "step": 4393 + }, + { + "epoch": 0.48253898528442785, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.345928430557251, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7044938802719116, + "num_tokens": 113727831.0, + "step": 4394 + }, + { + "epoch": 0.4826488029870415, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.3882956504821777, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.7035320997238159, + "num_tokens": 113750652.0, + "step": 4395 + }, + { + "epoch": 0.4827586206896552, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.1821208000183105, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7001407146453857, + "num_tokens": 113777924.0, + "step": 4396 + }, + { + "epoch": 0.48286843839226884, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.2962989807128906, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7154938578605652, + "num_tokens": 113801616.0, + "step": 4397 + }, + { + "epoch": 0.4829782560948825, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.2441322803497314, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.6858440637588501, + "num_tokens": 113827240.0, + "step": 4398 + }, + { + "epoch": 0.48308807379749613, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.439178466796875, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.733816921710968, + "num_tokens": 113847276.0, + "step": 4399 + }, + { + "epoch": 0.48319789150010983, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.0418264865875244, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.7033307552337646, + "num_tokens": 113876130.0, + "step": 4400 + }, + { + "epoch": 0.4833077092027235, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.307469129562378, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.705535888671875, + "num_tokens": 113900524.0, + "step": 4401 + }, + { + "epoch": 0.4834175269053371, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.504237174987793, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7149807810783386, + "num_tokens": 113920621.0, + "step": 4402 + }, + { + "epoch": 0.4835273446079508, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.3005383014678955, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7224693894386292, + "num_tokens": 113942610.0, + "step": 4403 + }, + { + "epoch": 0.4836371623105645, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.114100456237793, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6886522769927979, + "num_tokens": 113968029.0, + "step": 4404 + }, + { + "epoch": 0.4837469800131781, + "ewc_loss": 3.933906555175781e-05, + "grad_norm": 2.3196542263031006, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7127838134765625, + "num_tokens": 113992927.0, + "step": 4405 + }, + { + "epoch": 0.48385679771579176, + "ewc_loss": 3.9577484130859375e-05, + "grad_norm": 2.2250685691833496, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.6962034702301025, + "num_tokens": 114018573.0, + "step": 4406 + }, + { + "epoch": 0.48396661541840547, + "ewc_loss": 3.9577484130859375e-05, + "grad_norm": 2.103379487991333, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6845787763595581, + "num_tokens": 114048655.0, + "step": 4407 + }, + { + "epoch": 0.4840764331210191, + "ewc_loss": 3.9577484130859375e-05, + "grad_norm": 2.337632179260254, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6960739493370056, + "num_tokens": 114071771.0, + "step": 4408 + }, + { + "epoch": 0.48418625082363276, + "ewc_loss": 3.9577484130859375e-05, + "grad_norm": 2.4301297664642334, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.7052165865898132, + "num_tokens": 114094334.0, + "step": 4409 + }, + { + "epoch": 0.4842960685262464, + "ewc_loss": 3.9577484130859375e-05, + "grad_norm": 2.2956478595733643, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7105973958969116, + "num_tokens": 114118168.0, + "step": 4410 + }, + { + "epoch": 0.4844058862288601, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.5713512897491455, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6959916353225708, + "num_tokens": 114138611.0, + "step": 4411 + }, + { + "epoch": 0.48451570393147375, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.2994539737701416, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7202732563018799, + "num_tokens": 114160678.0, + "step": 4412 + }, + { + "epoch": 0.4846255216340874, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.475484609603882, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.731757402420044, + "num_tokens": 114180098.0, + "step": 4413 + }, + { + "epoch": 0.4847353393367011, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.3791799545288086, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7220560312271118, + "num_tokens": 114203210.0, + "step": 4414 + }, + { + "epoch": 0.48484515703931474, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 1.9898303747177124, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.7048039436340332, + "num_tokens": 114232636.0, + "step": 4415 + }, + { + "epoch": 0.4849549747419284, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.2840235233306885, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.698423445224762, + "num_tokens": 114256563.0, + "step": 4416 + }, + { + "epoch": 0.48506479244454204, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.369192123413086, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.6990940570831299, + "num_tokens": 114282338.0, + "step": 4417 + }, + { + "epoch": 0.48517461014715574, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.2379403114318848, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6918329000473022, + "num_tokens": 114306573.0, + "step": 4418 + }, + { + "epoch": 0.4852844278497694, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.1680612564086914, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7129878997802734, + "num_tokens": 114330804.0, + "step": 4419 + }, + { + "epoch": 0.48539424555238303, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 1.989379644393921, + "learning_rate": 1e-06, + "loss": 1.0369, + "mean_token_accuracy": 0.6825210452079773, + "num_tokens": 114362071.0, + "step": 4420 + }, + { + "epoch": 0.48550406325499673, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.32089900970459, + "learning_rate": 1e-06, + "loss": 1.039, + "mean_token_accuracy": 0.6849964261054993, + "num_tokens": 114386985.0, + "step": 4421 + }, + { + "epoch": 0.4856138809576104, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.4207029342651367, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7129006385803223, + "num_tokens": 114408362.0, + "step": 4422 + }, + { + "epoch": 0.485723698660224, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.0489554405212402, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6954182386398315, + "num_tokens": 114437239.0, + "step": 4423 + }, + { + "epoch": 0.48583351636283767, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.056889057159424, + "learning_rate": 1e-06, + "loss": 1.0583, + "mean_token_accuracy": 0.678708553314209, + "num_tokens": 114466495.0, + "step": 4424 + }, + { + "epoch": 0.48594333406545137, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.546051263809204, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7147307395935059, + "num_tokens": 114486989.0, + "step": 4425 + }, + { + "epoch": 0.486053151768065, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.3529629707336426, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.698971152305603, + "num_tokens": 114509544.0, + "step": 4426 + }, + { + "epoch": 0.48616296947067866, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.3734848499298096, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6899875402450562, + "num_tokens": 114533663.0, + "step": 4427 + }, + { + "epoch": 0.4862727871732923, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.0625107288360596, + "learning_rate": 1e-06, + "loss": 1.0503, + "mean_token_accuracy": 0.682120144367218, + "num_tokens": 114562909.0, + "step": 4428 + }, + { + "epoch": 0.486382604875906, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.3379783630371094, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6857134103775024, + "num_tokens": 114585352.0, + "step": 4429 + }, + { + "epoch": 0.48649242257851966, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.390239715576172, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7184963226318359, + "num_tokens": 114607829.0, + "step": 4430 + }, + { + "epoch": 0.4866022402811333, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.041560649871826, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.719531774520874, + "num_tokens": 114635523.0, + "step": 4431 + }, + { + "epoch": 0.486712057983747, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.1249938011169434, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.718904972076416, + "num_tokens": 114660260.0, + "step": 4432 + }, + { + "epoch": 0.48682187568636065, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.1119234561920166, + "learning_rate": 1e-06, + "loss": 1.0485, + "mean_token_accuracy": 0.6835269331932068, + "num_tokens": 114688684.0, + "step": 4433 + }, + { + "epoch": 0.4869316933889743, + "ewc_loss": 3.981590270996094e-05, + "grad_norm": 2.372528314590454, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7176348567008972, + "num_tokens": 114711677.0, + "step": 4434 + }, + { + "epoch": 0.48704151109158794, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2862136363983154, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.7005280256271362, + "num_tokens": 114735640.0, + "step": 4435 + }, + { + "epoch": 0.48715132879420164, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.044299602508545, + "learning_rate": 1e-06, + "loss": 1.0428, + "mean_token_accuracy": 0.6875378489494324, + "num_tokens": 114765799.0, + "step": 4436 + }, + { + "epoch": 0.4872611464968153, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 1.9615929126739502, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6880947351455688, + "num_tokens": 114795705.0, + "step": 4437 + }, + { + "epoch": 0.48737096419942894, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.3368911743164062, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6934720873832703, + "num_tokens": 114820558.0, + "step": 4438 + }, + { + "epoch": 0.48748078190204264, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.1209375858306885, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.700014591217041, + "num_tokens": 114847518.0, + "step": 4439 + }, + { + "epoch": 0.4875905996046563, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2187626361846924, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6930832266807556, + "num_tokens": 114875666.0, + "step": 4440 + }, + { + "epoch": 0.48770041730726993, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.129422903060913, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.6987624764442444, + "num_tokens": 114906232.0, + "step": 4441 + }, + { + "epoch": 0.4878102350098836, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.377755641937256, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7051135301589966, + "num_tokens": 114926526.0, + "step": 4442 + }, + { + "epoch": 0.4879200527124973, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2384374141693115, + "learning_rate": 1e-06, + "loss": 1.0712, + "mean_token_accuracy": 0.6872518062591553, + "num_tokens": 114954475.0, + "step": 4443 + }, + { + "epoch": 0.4880298704151109, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.345550775527954, + "learning_rate": 1e-06, + "loss": 1.0417, + "mean_token_accuracy": 0.6913644075393677, + "num_tokens": 114981638.0, + "step": 4444 + }, + { + "epoch": 0.48813968811772457, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.23286509513855, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.719322681427002, + "num_tokens": 115006462.0, + "step": 4445 + }, + { + "epoch": 0.4882495058203382, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.229778289794922, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7175868153572083, + "num_tokens": 115031636.0, + "step": 4446 + }, + { + "epoch": 0.4883593235229519, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2050235271453857, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7092037796974182, + "num_tokens": 115056335.0, + "step": 4447 + }, + { + "epoch": 0.48846914122556556, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.342932939529419, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6943163871765137, + "num_tokens": 115078790.0, + "step": 4448 + }, + { + "epoch": 0.4885789589281792, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.4164113998413086, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.717860221862793, + "num_tokens": 115101699.0, + "step": 4449 + }, + { + "epoch": 0.4886887766307929, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2628893852233887, + "learning_rate": 1e-06, + "loss": 1.0867, + "mean_token_accuracy": 0.6760330200195312, + "num_tokens": 115128503.0, + "step": 4450 + }, + { + "epoch": 0.48879859433340656, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 1.883170485496521, + "learning_rate": 1e-06, + "loss": 1.079, + "mean_token_accuracy": 0.6842539310455322, + "num_tokens": 115164634.0, + "step": 4451 + }, + { + "epoch": 0.4889084120360202, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.391329288482666, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6903545260429382, + "num_tokens": 115187246.0, + "step": 4452 + }, + { + "epoch": 0.48901822973863385, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.1250603199005127, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7056857943534851, + "num_tokens": 115213586.0, + "step": 4453 + }, + { + "epoch": 0.48912804744124755, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.447575330734253, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7248221039772034, + "num_tokens": 115235227.0, + "step": 4454 + }, + { + "epoch": 0.4892378651438612, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.0328571796417236, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7063648104667664, + "num_tokens": 115263710.0, + "step": 4455 + }, + { + "epoch": 0.48934768284647484, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.279287576675415, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7125059962272644, + "num_tokens": 115289195.0, + "step": 4456 + }, + { + "epoch": 0.4894575005490885, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.208739995956421, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7013243436813354, + "num_tokens": 115313974.0, + "step": 4457 + }, + { + "epoch": 0.4895673182517022, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.25260066986084, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6988768577575684, + "num_tokens": 115339114.0, + "step": 4458 + }, + { + "epoch": 0.48967713595431583, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.228881597518921, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6957542896270752, + "num_tokens": 115365289.0, + "step": 4459 + }, + { + "epoch": 0.4897869536569295, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.35056209564209, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7071696519851685, + "num_tokens": 115387444.0, + "step": 4460 + }, + { + "epoch": 0.4898967713595432, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.3276686668395996, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.6974954605102539, + "num_tokens": 115409230.0, + "step": 4461 + }, + { + "epoch": 0.4900065890621568, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.4330034255981445, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7180126309394836, + "num_tokens": 115430694.0, + "step": 4462 + }, + { + "epoch": 0.4901164067647705, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.2377865314483643, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6858983635902405, + "num_tokens": 115456685.0, + "step": 4463 + }, + { + "epoch": 0.4902262244673841, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.1379430294036865, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.6917650103569031, + "num_tokens": 115483755.0, + "step": 4464 + }, + { + "epoch": 0.4903360421699978, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.5406036376953125, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7227820158004761, + "num_tokens": 115502535.0, + "step": 4465 + }, + { + "epoch": 0.49044585987261147, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.2831592559814453, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.7019261717796326, + "num_tokens": 115526350.0, + "step": 4466 + }, + { + "epoch": 0.4905556775752251, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.451066732406616, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7141616344451904, + "num_tokens": 115548180.0, + "step": 4467 + }, + { + "epoch": 0.4906654952778388, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.073549270629883, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6955563426017761, + "num_tokens": 115575251.0, + "step": 4468 + }, + { + "epoch": 0.49077531298045246, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.380004405975342, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.728364109992981, + "num_tokens": 115595870.0, + "step": 4469 + }, + { + "epoch": 0.4908851306830661, + "ewc_loss": 4.00543212890625e-05, + "grad_norm": 2.1707653999328613, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6899206638336182, + "num_tokens": 115622500.0, + "step": 4470 + }, + { + "epoch": 0.49099494838567975, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.475900411605835, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6901373863220215, + "num_tokens": 115644812.0, + "step": 4471 + }, + { + "epoch": 0.49110476608829345, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.3213772773742676, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7112643718719482, + "num_tokens": 115668225.0, + "step": 4472 + }, + { + "epoch": 0.4912145837909071, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.4562580585479736, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.722892165184021, + "num_tokens": 115690591.0, + "step": 4473 + }, + { + "epoch": 0.49132440149352075, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.326512336730957, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6853674650192261, + "num_tokens": 115715069.0, + "step": 4474 + }, + { + "epoch": 0.4914342191961344, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.421488046646118, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.702133297920227, + "num_tokens": 115737035.0, + "step": 4475 + }, + { + "epoch": 0.4915440368987481, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.139277935028076, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.7066441774368286, + "num_tokens": 115762455.0, + "step": 4476 + }, + { + "epoch": 0.49165385460136174, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.289734363555908, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7072608470916748, + "num_tokens": 115786117.0, + "step": 4477 + }, + { + "epoch": 0.4917636723039754, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.329533815383911, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7272304892539978, + "num_tokens": 115809341.0, + "step": 4478 + }, + { + "epoch": 0.4918734900065891, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.309979200363159, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7065721750259399, + "num_tokens": 115832269.0, + "step": 4479 + }, + { + "epoch": 0.49198330770920273, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.1974198818206787, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7025806307792664, + "num_tokens": 115859033.0, + "step": 4480 + }, + { + "epoch": 0.4920931254118164, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.3614413738250732, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.706609308719635, + "num_tokens": 115881543.0, + "step": 4481 + }, + { + "epoch": 0.49220294311443, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.198375701904297, + "learning_rate": 1e-06, + "loss": 1.083, + "mean_token_accuracy": 0.6833025217056274, + "num_tokens": 115908358.0, + "step": 4482 + }, + { + "epoch": 0.4923127608170437, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.196397066116333, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.704775333404541, + "num_tokens": 115934190.0, + "step": 4483 + }, + { + "epoch": 0.4924225785196574, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.241192102432251, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6824656128883362, + "num_tokens": 115958843.0, + "step": 4484 + }, + { + "epoch": 0.492532396222271, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.1509251594543457, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6840296983718872, + "num_tokens": 115984943.0, + "step": 4485 + }, + { + "epoch": 0.49264221392488466, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.040365695953369, + "learning_rate": 1e-06, + "loss": 1.1102, + "mean_token_accuracy": 0.679373025894165, + "num_tokens": 116015263.0, + "step": 4486 + }, + { + "epoch": 0.49275203162749837, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.149895429611206, + "learning_rate": 1e-06, + "loss": 1.0543, + "mean_token_accuracy": 0.6908923983573914, + "num_tokens": 116042471.0, + "step": 4487 + }, + { + "epoch": 0.492861849330112, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.3109290599823, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.7040151357650757, + "num_tokens": 116067499.0, + "step": 4488 + }, + { + "epoch": 0.49297166703272566, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.3618874549865723, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7190531492233276, + "num_tokens": 116088440.0, + "step": 4489 + }, + { + "epoch": 0.49308148473533936, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.404721736907959, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6883746385574341, + "num_tokens": 116111112.0, + "step": 4490 + }, + { + "epoch": 0.493191302437953, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.258584976196289, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7013933062553406, + "num_tokens": 116135455.0, + "step": 4491 + }, + { + "epoch": 0.49330112014056665, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.282970666885376, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.689152717590332, + "num_tokens": 116160890.0, + "step": 4492 + }, + { + "epoch": 0.4934109378431803, + "ewc_loss": 4.029273986816406e-05, + "grad_norm": 2.1352648735046387, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.713151216506958, + "num_tokens": 116187808.0, + "step": 4493 + }, + { + "epoch": 0.493520755545794, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.057504653930664, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.688165009021759, + "num_tokens": 116217838.0, + "step": 4494 + }, + { + "epoch": 0.49363057324840764, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.068870782852173, + "learning_rate": 1e-06, + "loss": 1.0671, + "mean_token_accuracy": 0.6884291172027588, + "num_tokens": 116246981.0, + "step": 4495 + }, + { + "epoch": 0.4937403909510213, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.0836105346679688, + "learning_rate": 1e-06, + "loss": 1.074, + "mean_token_accuracy": 0.6784453392028809, + "num_tokens": 116276532.0, + "step": 4496 + }, + { + "epoch": 0.493850208653635, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.1566827297210693, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.709072470664978, + "num_tokens": 116302544.0, + "step": 4497 + }, + { + "epoch": 0.49396002635624864, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.3861682415008545, + "learning_rate": 1e-06, + "loss": 1.1038, + "mean_token_accuracy": 0.673454225063324, + "num_tokens": 116326255.0, + "step": 4498 + }, + { + "epoch": 0.4940698440588623, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.2212862968444824, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7041939496994019, + "num_tokens": 116351717.0, + "step": 4499 + }, + { + "epoch": 0.49417966176147593, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.091111183166504, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.6895101070404053, + "num_tokens": 116381049.0, + "step": 4500 + }, + { + "epoch": 0.49428947946408963, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.315553903579712, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7141073346138, + "num_tokens": 116404558.0, + "step": 4501 + }, + { + "epoch": 0.4943992971667033, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.257138729095459, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.6962716579437256, + "num_tokens": 116430821.0, + "step": 4502 + }, + { + "epoch": 0.4945091148693169, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.188737630844116, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7164139747619629, + "num_tokens": 116456843.0, + "step": 4503 + }, + { + "epoch": 0.49461893257193057, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.001615524291992, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7016437649726868, + "num_tokens": 116488522.0, + "step": 4504 + }, + { + "epoch": 0.49472875027454427, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 1.95966374874115, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6903389096260071, + "num_tokens": 116518860.0, + "step": 4505 + }, + { + "epoch": 0.4948385679771579, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.29226016998291, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7134048342704773, + "num_tokens": 116545374.0, + "step": 4506 + }, + { + "epoch": 0.49494838567977156, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.2139408588409424, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7042593359947205, + "num_tokens": 116572772.0, + "step": 4507 + }, + { + "epoch": 0.49505820338238526, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.0272157192230225, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.696663498878479, + "num_tokens": 116601968.0, + "step": 4508 + }, + { + "epoch": 0.4951680210849989, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.429306745529175, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6996279954910278, + "num_tokens": 116624913.0, + "step": 4509 + }, + { + "epoch": 0.49527783878761256, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.3974881172180176, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6992456316947937, + "num_tokens": 116646611.0, + "step": 4510 + }, + { + "epoch": 0.4953876564902262, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.3920488357543945, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.6956756114959717, + "num_tokens": 116669267.0, + "step": 4511 + }, + { + "epoch": 0.4954974741928399, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.2752034664154053, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7084368467330933, + "num_tokens": 116694288.0, + "step": 4512 + }, + { + "epoch": 0.49560729189545355, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.2194337844848633, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7104235887527466, + "num_tokens": 116719128.0, + "step": 4513 + }, + { + "epoch": 0.4957171095980672, + "ewc_loss": 4.0531158447265625e-05, + "grad_norm": 2.318418502807617, + "learning_rate": 1e-06, + "loss": 1.0787, + "mean_token_accuracy": 0.686150312423706, + "num_tokens": 116745096.0, + "step": 4514 + }, + { + "epoch": 0.4958269273006809, + "ewc_loss": 4.076957702636719e-05, + "grad_norm": 2.6191461086273193, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7138140201568604, + "num_tokens": 116763826.0, + "step": 4515 + }, + { + "epoch": 0.49593674500329454, + "ewc_loss": 4.076957702636719e-05, + "grad_norm": 2.1113109588623047, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7023962736129761, + "num_tokens": 116795641.0, + "step": 4516 + }, + { + "epoch": 0.4960465627059082, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.1253998279571533, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.6905978918075562, + "num_tokens": 116822911.0, + "step": 4517 + }, + { + "epoch": 0.49615638040852184, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.3264358043670654, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7019485831260681, + "num_tokens": 116846246.0, + "step": 4518 + }, + { + "epoch": 0.49626619811113554, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.0205366611480713, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.701150119304657, + "num_tokens": 116874624.0, + "step": 4519 + }, + { + "epoch": 0.4963760158137492, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.0992376804351807, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7152999639511108, + "num_tokens": 116899231.0, + "step": 4520 + }, + { + "epoch": 0.49648583351636283, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.350048542022705, + "learning_rate": 1e-06, + "loss": 1.0868, + "mean_token_accuracy": 0.6726410388946533, + "num_tokens": 116924433.0, + "step": 4521 + }, + { + "epoch": 0.4965956512189765, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.1241188049316406, + "learning_rate": 1e-06, + "loss": 1.0629, + "mean_token_accuracy": 0.685797393321991, + "num_tokens": 116954796.0, + "step": 4522 + }, + { + "epoch": 0.4967054689215902, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 1.95527982711792, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7050691843032837, + "num_tokens": 116984925.0, + "step": 4523 + }, + { + "epoch": 0.4968152866242038, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.1015241146087646, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.698263943195343, + "num_tokens": 117012418.0, + "step": 4524 + }, + { + "epoch": 0.49692510432681747, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.0246591567993164, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7144478559494019, + "num_tokens": 117040062.0, + "step": 4525 + }, + { + "epoch": 0.49703492202943117, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 1.9723049402236938, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6883184909820557, + "num_tokens": 117069854.0, + "step": 4526 + }, + { + "epoch": 0.4971447397320448, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.0171430110931396, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7004155516624451, + "num_tokens": 117098015.0, + "step": 4527 + }, + { + "epoch": 0.49725455743465846, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.19568133354187, + "learning_rate": 1e-06, + "loss": 1.0698, + "mean_token_accuracy": 0.6817071437835693, + "num_tokens": 117127797.0, + "step": 4528 + }, + { + "epoch": 0.4973643751372721, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.3946216106414795, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.6963861584663391, + "num_tokens": 117147759.0, + "step": 4529 + }, + { + "epoch": 0.4974741928398858, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2505316734313965, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6949387788772583, + "num_tokens": 117171512.0, + "step": 4530 + }, + { + "epoch": 0.49758401054249946, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2714481353759766, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6915163397789001, + "num_tokens": 117196904.0, + "step": 4531 + }, + { + "epoch": 0.4976938282451131, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.498957872390747, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7059726715087891, + "num_tokens": 117218115.0, + "step": 4532 + }, + { + "epoch": 0.49780364594772675, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.085529327392578, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7108174562454224, + "num_tokens": 117244201.0, + "step": 4533 + }, + { + "epoch": 0.49791346365034045, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.189913034439087, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7351136207580566, + "num_tokens": 117267186.0, + "step": 4534 + }, + { + "epoch": 0.4980232813529541, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 1.9958261251449585, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6947590112686157, + "num_tokens": 117297522.0, + "step": 4535 + }, + { + "epoch": 0.49813309905556774, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.3141088485717773, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6934026479721069, + "num_tokens": 117323164.0, + "step": 4536 + }, + { + "epoch": 0.49824291675818144, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.299147605895996, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.7010619640350342, + "num_tokens": 117347079.0, + "step": 4537 + }, + { + "epoch": 0.4983527344607951, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2660489082336426, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7063635587692261, + "num_tokens": 117373678.0, + "step": 4538 + }, + { + "epoch": 0.49846255216340873, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.184152364730835, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6913154721260071, + "num_tokens": 117400721.0, + "step": 4539 + }, + { + "epoch": 0.4985723698660224, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2422120571136475, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7114068269729614, + "num_tokens": 117425568.0, + "step": 4540 + }, + { + "epoch": 0.4986821875686361, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.296929121017456, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7000089883804321, + "num_tokens": 117448642.0, + "step": 4541 + }, + { + "epoch": 0.4987920052712497, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.4162709712982178, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.701631486415863, + "num_tokens": 117471867.0, + "step": 4542 + }, + { + "epoch": 0.4989018229738634, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 1.9049495458602905, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.7001504302024841, + "num_tokens": 117503893.0, + "step": 4543 + }, + { + "epoch": 0.4990116406764771, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.124905824661255, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6949414014816284, + "num_tokens": 117532300.0, + "step": 4544 + }, + { + "epoch": 0.4991214583790907, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2411794662475586, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7117039561271667, + "num_tokens": 117557057.0, + "step": 4545 + }, + { + "epoch": 0.49923127608170437, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.070599317550659, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.694054365158081, + "num_tokens": 117586763.0, + "step": 4546 + }, + { + "epoch": 0.499341093784318, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.201448440551758, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.7062721252441406, + "num_tokens": 117611783.0, + "step": 4547 + }, + { + "epoch": 0.4994509114869317, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.120976686477661, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7271204590797424, + "num_tokens": 117638085.0, + "step": 4548 + }, + { + "epoch": 0.49956072918954536, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.017638683319092, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6968317031860352, + "num_tokens": 117669103.0, + "step": 4549 + }, + { + "epoch": 0.499670546892159, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.3282630443573, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6889708042144775, + "num_tokens": 117693830.0, + "step": 4550 + }, + { + "epoch": 0.49978036459477265, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.521491050720215, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6920267343521118, + "num_tokens": 117714230.0, + "step": 4551 + }, + { + "epoch": 0.49989018229738635, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 1.913295030593872, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6937824487686157, + "num_tokens": 117747833.0, + "step": 4552 + }, + { + "epoch": 0.5, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2848010063171387, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6860190033912659, + "num_tokens": 117773052.0, + "step": 4553 + }, + { + "epoch": 0.5001098177026136, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.1321330070495605, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7061574459075928, + "num_tokens": 117798234.0, + "step": 4554 + }, + { + "epoch": 0.5002196354052273, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.5060973167419434, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7099496126174927, + "num_tokens": 117820361.0, + "step": 4555 + }, + { + "epoch": 0.5003294531078409, + "ewc_loss": 4.100799560546875e-05, + "grad_norm": 2.2467167377471924, + "learning_rate": 1e-06, + "loss": 1.0317, + "mean_token_accuracy": 0.6882634162902832, + "num_tokens": 117846772.0, + "step": 4556 + }, + { + "epoch": 0.5004392708104547, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.0120420455932617, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.6901893615722656, + "num_tokens": 117877864.0, + "step": 4557 + }, + { + "epoch": 0.5005490885130683, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.168572425842285, + "learning_rate": 1e-06, + "loss": 1.0543, + "mean_token_accuracy": 0.6818978786468506, + "num_tokens": 117905908.0, + "step": 4558 + }, + { + "epoch": 0.500658906215682, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.3803927898406982, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6996749639511108, + "num_tokens": 117931453.0, + "step": 4559 + }, + { + "epoch": 0.5007687239182956, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.4565112590789795, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7077621221542358, + "num_tokens": 117953672.0, + "step": 4560 + }, + { + "epoch": 0.5008785416209093, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.1981570720672607, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6865916848182678, + "num_tokens": 117979254.0, + "step": 4561 + }, + { + "epoch": 0.5009883593235229, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.116550922393799, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7039799094200134, + "num_tokens": 118007249.0, + "step": 4562 + }, + { + "epoch": 0.5010981770261366, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.1028425693511963, + "learning_rate": 1e-06, + "loss": 1.063, + "mean_token_accuracy": 0.6826986074447632, + "num_tokens": 118038228.0, + "step": 4563 + }, + { + "epoch": 0.5012079947287503, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.1276907920837402, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.7030066847801208, + "num_tokens": 118064982.0, + "step": 4564 + }, + { + "epoch": 0.501317812431364, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.2209982872009277, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7048249840736389, + "num_tokens": 118088105.0, + "step": 4565 + }, + { + "epoch": 0.5014276301339776, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 1.9072798490524292, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6965187788009644, + "num_tokens": 118123176.0, + "step": 4566 + }, + { + "epoch": 0.5015374478365913, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.126297950744629, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7012367248535156, + "num_tokens": 118150578.0, + "step": 4567 + }, + { + "epoch": 0.5016472655392049, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.1132845878601074, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7233842611312866, + "num_tokens": 118177358.0, + "step": 4568 + }, + { + "epoch": 0.5017570832418186, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.125352144241333, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.7006345391273499, + "num_tokens": 118207315.0, + "step": 4569 + }, + { + "epoch": 0.5018669009444322, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.2280333042144775, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7002076506614685, + "num_tokens": 118231938.0, + "step": 4570 + }, + { + "epoch": 0.5019767186470458, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.0233147144317627, + "learning_rate": 1e-06, + "loss": 1.1007, + "mean_token_accuracy": 0.6698368787765503, + "num_tokens": 118264767.0, + "step": 4571 + }, + { + "epoch": 0.5020865363496596, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.3485705852508545, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6876600980758667, + "num_tokens": 118287546.0, + "step": 4572 + }, + { + "epoch": 0.5021963540522733, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.3085598945617676, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7110309600830078, + "num_tokens": 118310519.0, + "step": 4573 + }, + { + "epoch": 0.5023061717548869, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.138061046600342, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7028056979179382, + "num_tokens": 118335584.0, + "step": 4574 + }, + { + "epoch": 0.5024159894575005, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.4163029193878174, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7167425751686096, + "num_tokens": 118355882.0, + "step": 4575 + }, + { + "epoch": 0.5025258071601142, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.2806806564331055, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6982849836349487, + "num_tokens": 118379758.0, + "step": 4576 + }, + { + "epoch": 0.5026356248627278, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.066089630126953, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6928063631057739, + "num_tokens": 118409998.0, + "step": 4577 + }, + { + "epoch": 0.5027454425653415, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.0488221645355225, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7035326361656189, + "num_tokens": 118437895.0, + "step": 4578 + }, + { + "epoch": 0.5028552602679552, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.0059850215911865, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6963838338851929, + "num_tokens": 118466254.0, + "step": 4579 + }, + { + "epoch": 0.5029650779705689, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.289576292037964, + "learning_rate": 1e-06, + "loss": 1.0675, + "mean_token_accuracy": 0.6880753040313721, + "num_tokens": 118491266.0, + "step": 4580 + }, + { + "epoch": 0.5030748956731825, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.441164255142212, + "learning_rate": 1e-06, + "loss": 1.0546, + "mean_token_accuracy": 0.6879016160964966, + "num_tokens": 118513453.0, + "step": 4581 + }, + { + "epoch": 0.5031847133757962, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.3671927452087402, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6873496770858765, + "num_tokens": 118536240.0, + "step": 4582 + }, + { + "epoch": 0.5032945310784098, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0385494232177734, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6964448094367981, + "num_tokens": 118566445.0, + "step": 4583 + }, + { + "epoch": 0.5034043487810235, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.240508794784546, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6935775279998779, + "num_tokens": 118591213.0, + "step": 4584 + }, + { + "epoch": 0.5035141664836371, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.207387924194336, + "learning_rate": 1e-06, + "loss": 1.126, + "mean_token_accuracy": 0.6707099676132202, + "num_tokens": 118621747.0, + "step": 4585 + }, + { + "epoch": 0.5036239841862509, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 1.7398462295532227, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.695659875869751, + "num_tokens": 118660496.0, + "step": 4586 + }, + { + "epoch": 0.5037338018888645, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.3911843299865723, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.713804304599762, + "num_tokens": 118682445.0, + "step": 4587 + }, + { + "epoch": 0.5038436195914782, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.100071907043457, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.702576220035553, + "num_tokens": 118709958.0, + "step": 4588 + }, + { + "epoch": 0.5039534372940918, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.175191879272461, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.7073023319244385, + "num_tokens": 118736291.0, + "step": 4589 + }, + { + "epoch": 0.5040632549967055, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.097288131713867, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.6888691782951355, + "num_tokens": 118763755.0, + "step": 4590 + }, + { + "epoch": 0.5041730726993191, + "ewc_loss": 4.124641418457031e-05, + "grad_norm": 2.35609769821167, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.711129903793335, + "num_tokens": 118785592.0, + "step": 4591 + }, + { + "epoch": 0.5042828904019327, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.9210671186447144, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6990358233451843, + "num_tokens": 118815809.0, + "step": 4592 + }, + { + "epoch": 0.5043927081045465, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.962632417678833, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7025598287582397, + "num_tokens": 118844118.0, + "step": 4593 + }, + { + "epoch": 0.5045025258071602, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.8707183599472046, + "learning_rate": 1e-06, + "loss": 1.1367, + "mean_token_accuracy": 0.6605850458145142, + "num_tokens": 118878278.0, + "step": 4594 + }, + { + "epoch": 0.5046123435097738, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.1080265045166016, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6941395998001099, + "num_tokens": 118904888.0, + "step": 4595 + }, + { + "epoch": 0.5047221612123874, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0208137035369873, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.700070858001709, + "num_tokens": 118936516.0, + "step": 4596 + }, + { + "epoch": 0.5048319789150011, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.9680885076522827, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.6880228519439697, + "num_tokens": 118965982.0, + "step": 4597 + }, + { + "epoch": 0.5049417966176147, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.259918689727783, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7151049375534058, + "num_tokens": 118989660.0, + "step": 4598 + }, + { + "epoch": 0.5050516143202284, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.2854011058807373, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7132841348648071, + "num_tokens": 119014512.0, + "step": 4599 + }, + { + "epoch": 0.505161432022842, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.9834563732147217, + "learning_rate": 1e-06, + "loss": 1.0614, + "mean_token_accuracy": 0.680692195892334, + "num_tokens": 119046114.0, + "step": 4600 + }, + { + "epoch": 0.5052712497254558, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.5115532875061035, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.6995795369148254, + "num_tokens": 119069967.0, + "step": 4601 + }, + { + "epoch": 0.5053810674280694, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0565524101257324, + "learning_rate": 1e-06, + "loss": 1.0783, + "mean_token_accuracy": 0.6799873113632202, + "num_tokens": 119102712.0, + "step": 4602 + }, + { + "epoch": 0.5054908851306831, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.137718915939331, + "learning_rate": 1e-06, + "loss": 0.9908, + "mean_token_accuracy": 0.6991839408874512, + "num_tokens": 119131270.0, + "step": 4603 + }, + { + "epoch": 0.5056007028332967, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.925203323364258, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7173054814338684, + "num_tokens": 119147580.0, + "step": 4604 + }, + { + "epoch": 0.5057105205359104, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.085085153579712, + "learning_rate": 1e-06, + "loss": 1.0644, + "mean_token_accuracy": 0.6800962686538696, + "num_tokens": 119176404.0, + "step": 4605 + }, + { + "epoch": 0.505820338238524, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.175828218460083, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6988120079040527, + "num_tokens": 119203393.0, + "step": 4606 + }, + { + "epoch": 0.5059301559411377, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.113598108291626, + "learning_rate": 1e-06, + "loss": 1.0926, + "mean_token_accuracy": 0.6737560033798218, + "num_tokens": 119231833.0, + "step": 4607 + }, + { + "epoch": 0.5060399736437514, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.3643743991851807, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.699424147605896, + "num_tokens": 119256386.0, + "step": 4608 + }, + { + "epoch": 0.5061497913463651, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.071345090866089, + "learning_rate": 1e-06, + "loss": 1.0359, + "mean_token_accuracy": 0.6827893257141113, + "num_tokens": 119285024.0, + "step": 4609 + }, + { + "epoch": 0.5062596090489787, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.2934718132019043, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.721655011177063, + "num_tokens": 119308733.0, + "step": 4610 + }, + { + "epoch": 0.5063694267515924, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.322538137435913, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6949371099472046, + "num_tokens": 119332389.0, + "step": 4611 + }, + { + "epoch": 0.506479244454206, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.1503894329071045, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6927757263183594, + "num_tokens": 119360140.0, + "step": 4612 + }, + { + "epoch": 0.5065890621568196, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.466303586959839, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.7186545133590698, + "num_tokens": 119381098.0, + "step": 4613 + }, + { + "epoch": 0.5066988798594333, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.176038980484009, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7294630408287048, + "num_tokens": 119406309.0, + "step": 4614 + }, + { + "epoch": 0.506808697562047, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.3227148056030273, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7246053218841553, + "num_tokens": 119429642.0, + "step": 4615 + }, + { + "epoch": 0.5069185152646607, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 1.9303137063980103, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6919154524803162, + "num_tokens": 119461475.0, + "step": 4616 + }, + { + "epoch": 0.5070283329672743, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.302255868911743, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7117187976837158, + "num_tokens": 119485600.0, + "step": 4617 + }, + { + "epoch": 0.507138150669888, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.207083225250244, + "learning_rate": 1e-06, + "loss": 1.0834, + "mean_token_accuracy": 0.6834497451782227, + "num_tokens": 119511824.0, + "step": 4618 + }, + { + "epoch": 0.5072479683725016, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.447040319442749, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7089660167694092, + "num_tokens": 119533133.0, + "step": 4619 + }, + { + "epoch": 0.5073577860751153, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.280388116836548, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7017815113067627, + "num_tokens": 119557080.0, + "step": 4620 + }, + { + "epoch": 0.5074676037777289, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.4800426959991455, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6906068325042725, + "num_tokens": 119577569.0, + "step": 4621 + }, + { + "epoch": 0.5075774214803427, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.205674171447754, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6926932334899902, + "num_tokens": 119605645.0, + "step": 4622 + }, + { + "epoch": 0.5076872391829563, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.391289710998535, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.706859827041626, + "num_tokens": 119626742.0, + "step": 4623 + }, + { + "epoch": 0.50779705688557, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.388815402984619, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.6996084451675415, + "num_tokens": 119648331.0, + "step": 4624 + }, + { + "epoch": 0.5079068745881836, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.307377815246582, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7078964710235596, + "num_tokens": 119671073.0, + "step": 4625 + }, + { + "epoch": 0.5080166922907973, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.1966941356658936, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7065038681030273, + "num_tokens": 119697145.0, + "step": 4626 + }, + { + "epoch": 0.5081265099934109, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.432023286819458, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.693347692489624, + "num_tokens": 119719279.0, + "step": 4627 + }, + { + "epoch": 0.5082363276960246, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0604352951049805, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7070542573928833, + "num_tokens": 119746924.0, + "step": 4628 + }, + { + "epoch": 0.5083461453986382, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.4681036472320557, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.6818152666091919, + "num_tokens": 119769037.0, + "step": 4629 + }, + { + "epoch": 0.508455963101252, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.010035514831543, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.6871671676635742, + "num_tokens": 119797640.0, + "step": 4630 + }, + { + "epoch": 0.5085657808038656, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.152549982070923, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7154425978660583, + "num_tokens": 119821460.0, + "step": 4631 + }, + { + "epoch": 0.5086755985064793, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0926613807678223, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6841773986816406, + "num_tokens": 119850542.0, + "step": 4632 + }, + { + "epoch": 0.5087854162090929, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.046480655670166, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7071764469146729, + "num_tokens": 119877816.0, + "step": 4633 + }, + { + "epoch": 0.5088952339117065, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0732150077819824, + "learning_rate": 1e-06, + "loss": 1.0498, + "mean_token_accuracy": 0.6831796765327454, + "num_tokens": 119907607.0, + "step": 4634 + }, + { + "epoch": 0.5090050516143202, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.1662216186523438, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6977258920669556, + "num_tokens": 119933810.0, + "step": 4635 + }, + { + "epoch": 0.5091148693169338, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.258211374282837, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6791145205497742, + "num_tokens": 119960256.0, + "step": 4636 + }, + { + "epoch": 0.5092246870195476, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0555927753448486, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7012448310852051, + "num_tokens": 119988437.0, + "step": 4637 + }, + { + "epoch": 0.5093345047221612, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.2429773807525635, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7031117081642151, + "num_tokens": 120012553.0, + "step": 4638 + }, + { + "epoch": 0.5094443224247749, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.214064836502075, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.6989986896514893, + "num_tokens": 120037367.0, + "step": 4639 + }, + { + "epoch": 0.5095541401273885, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.5433859825134277, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7201346158981323, + "num_tokens": 120056816.0, + "step": 4640 + }, + { + "epoch": 0.5096639578300022, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 1.9326499700546265, + "learning_rate": 1e-06, + "loss": 1.061, + "mean_token_accuracy": 0.6825082898139954, + "num_tokens": 120090690.0, + "step": 4641 + }, + { + "epoch": 0.5097737755326158, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2766330242156982, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7057361602783203, + "num_tokens": 120114819.0, + "step": 4642 + }, + { + "epoch": 0.5098835932352295, + "ewc_loss": 4.172325134277344e-05, + "grad_norm": 2.0334949493408203, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7065478563308716, + "num_tokens": 120144814.0, + "step": 4643 + }, + { + "epoch": 0.5099934109378432, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2392306327819824, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7219640016555786, + "num_tokens": 120168876.0, + "step": 4644 + }, + { + "epoch": 0.5101032286404569, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2212815284729004, + "learning_rate": 1e-06, + "loss": 1.0608, + "mean_token_accuracy": 0.6817471981048584, + "num_tokens": 120196920.0, + "step": 4645 + }, + { + "epoch": 0.5102130463430705, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.3040146827697754, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6964340209960938, + "num_tokens": 120222102.0, + "step": 4646 + }, + { + "epoch": 0.5103228640456842, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.210587501525879, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7029541730880737, + "num_tokens": 120247712.0, + "step": 4647 + }, + { + "epoch": 0.5104326817482978, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 1.9280859231948853, + "learning_rate": 1e-06, + "loss": 1.0512, + "mean_token_accuracy": 0.6898708343505859, + "num_tokens": 120281315.0, + "step": 4648 + }, + { + "epoch": 0.5105424994509115, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.1312613487243652, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.7033575773239136, + "num_tokens": 120309377.0, + "step": 4649 + }, + { + "epoch": 0.5106523171535251, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.084333658218384, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6882239580154419, + "num_tokens": 120337949.0, + "step": 4650 + }, + { + "epoch": 0.5107621348561389, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.184626340866089, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.706828773021698, + "num_tokens": 120362736.0, + "step": 4651 + }, + { + "epoch": 0.5108719525587525, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.1158406734466553, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6914785504341125, + "num_tokens": 120394003.0, + "step": 4652 + }, + { + "epoch": 0.5109817702613662, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.365384817123413, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6920284628868103, + "num_tokens": 120417078.0, + "step": 4653 + }, + { + "epoch": 0.5110915879639798, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2439699172973633, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7236389517784119, + "num_tokens": 120440605.0, + "step": 4654 + }, + { + "epoch": 0.5112014056665934, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.3828916549682617, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7058811783790588, + "num_tokens": 120463428.0, + "step": 4655 + }, + { + "epoch": 0.5113112233692071, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.1086628437042236, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7020626068115234, + "num_tokens": 120493073.0, + "step": 4656 + }, + { + "epoch": 0.5114210410718207, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2086844444274902, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7000129222869873, + "num_tokens": 120518143.0, + "step": 4657 + }, + { + "epoch": 0.5115308587744345, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.284093141555786, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.704883337020874, + "num_tokens": 120542359.0, + "step": 4658 + }, + { + "epoch": 0.5116406764770481, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.154630184173584, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7036228179931641, + "num_tokens": 120571395.0, + "step": 4659 + }, + { + "epoch": 0.5117504941796618, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 1.8686408996582031, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.6825406551361084, + "num_tokens": 120607912.0, + "step": 4660 + }, + { + "epoch": 0.5118603118822754, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.0012383460998535, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6998876333236694, + "num_tokens": 120639223.0, + "step": 4661 + }, + { + "epoch": 0.5119701295848891, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.554783344268799, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7087900638580322, + "num_tokens": 120659595.0, + "step": 4662 + }, + { + "epoch": 0.5120799472875027, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 1.875909447669983, + "learning_rate": 1e-06, + "loss": 1.0508, + "mean_token_accuracy": 0.687829852104187, + "num_tokens": 120694465.0, + "step": 4663 + }, + { + "epoch": 0.5121897649901164, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.193587064743042, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6981556415557861, + "num_tokens": 120722417.0, + "step": 4664 + }, + { + "epoch": 0.51229958269273, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.1255626678466797, + "learning_rate": 1e-06, + "loss": 1.0748, + "mean_token_accuracy": 0.6820540428161621, + "num_tokens": 120750601.0, + "step": 4665 + }, + { + "epoch": 0.5124094003953438, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.0060391426086426, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6776336431503296, + "num_tokens": 120781378.0, + "step": 4666 + }, + { + "epoch": 0.5125192180979574, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2152347564697266, + "learning_rate": 1e-06, + "loss": 1.0273, + "mean_token_accuracy": 0.6927154064178467, + "num_tokens": 120807616.0, + "step": 4667 + }, + { + "epoch": 0.5126290358005711, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.282824993133545, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6944192051887512, + "num_tokens": 120832422.0, + "step": 4668 + }, + { + "epoch": 0.5127388535031847, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.1492130756378174, + "learning_rate": 1e-06, + "loss": 1.0698, + "mean_token_accuracy": 0.6767889261245728, + "num_tokens": 120859714.0, + "step": 4669 + }, + { + "epoch": 0.5128486712057984, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.315892457962036, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7140821814537048, + "num_tokens": 120883835.0, + "step": 4670 + }, + { + "epoch": 0.512958488908412, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2297699451446533, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.696424663066864, + "num_tokens": 120907148.0, + "step": 4671 + }, + { + "epoch": 0.5130683066110256, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.0228025913238525, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7020672559738159, + "num_tokens": 120937197.0, + "step": 4672 + }, + { + "epoch": 0.5131781243136394, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.3100342750549316, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7114489078521729, + "num_tokens": 120960227.0, + "step": 4673 + }, + { + "epoch": 0.513287942016253, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.077174425125122, + "learning_rate": 1e-06, + "loss": 1.0664, + "mean_token_accuracy": 0.6977744698524475, + "num_tokens": 120987671.0, + "step": 4674 + }, + { + "epoch": 0.5133977597188667, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.176553249359131, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7177020311355591, + "num_tokens": 121014518.0, + "step": 4675 + }, + { + "epoch": 0.5135075774214803, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.078688383102417, + "learning_rate": 1e-06, + "loss": 1.0213, + "mean_token_accuracy": 0.6908296346664429, + "num_tokens": 121046518.0, + "step": 4676 + }, + { + "epoch": 0.513617395124094, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.302032232284546, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7066012620925903, + "num_tokens": 121070381.0, + "step": 4677 + }, + { + "epoch": 0.5137272128267076, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2512195110321045, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7074713706970215, + "num_tokens": 121093702.0, + "step": 4678 + }, + { + "epoch": 0.5138370305293213, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.6547553539276123, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.6956616044044495, + "num_tokens": 121114919.0, + "step": 4679 + }, + { + "epoch": 0.513946848231935, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 2.2891194820404053, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.699439287185669, + "num_tokens": 121137886.0, + "step": 4680 + }, + { + "epoch": 0.5140566659345487, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.160665273666382, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.701829195022583, + "num_tokens": 121164767.0, + "step": 4681 + }, + { + "epoch": 0.5141664836371623, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.041827917098999, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6910654306411743, + "num_tokens": 121194080.0, + "step": 4682 + }, + { + "epoch": 0.514276301339776, + "ewc_loss": 4.220008850097656e-05, + "grad_norm": 1.9838603734970093, + "learning_rate": 1e-06, + "loss": 1.1109, + "mean_token_accuracy": 0.6683027148246765, + "num_tokens": 121226531.0, + "step": 4683 + }, + { + "epoch": 0.5143861190423896, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.2805111408233643, + "learning_rate": 1e-06, + "loss": 1.099, + "mean_token_accuracy": 0.6755425930023193, + "num_tokens": 121254072.0, + "step": 4684 + }, + { + "epoch": 0.5144959367450033, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.276437282562256, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6998783349990845, + "num_tokens": 121279731.0, + "step": 4685 + }, + { + "epoch": 0.5146057544476169, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 1.8708802461624146, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7178969383239746, + "num_tokens": 121312266.0, + "step": 4686 + }, + { + "epoch": 0.5147155721502307, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.2109596729278564, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.6984569430351257, + "num_tokens": 121337329.0, + "step": 4687 + }, + { + "epoch": 0.5148253898528443, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 1.9414963722229004, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6927576661109924, + "num_tokens": 121369123.0, + "step": 4688 + }, + { + "epoch": 0.514935207555458, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.300489902496338, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7065207958221436, + "num_tokens": 121391602.0, + "step": 4689 + }, + { + "epoch": 0.5150450252580716, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.1152963638305664, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6867765784263611, + "num_tokens": 121418720.0, + "step": 4690 + }, + { + "epoch": 0.5151548429606853, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.122746229171753, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7197728157043457, + "num_tokens": 121444116.0, + "step": 4691 + }, + { + "epoch": 0.5152646606632989, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 1.982300877571106, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.7000665068626404, + "num_tokens": 121474478.0, + "step": 4692 + }, + { + "epoch": 0.5153744783659125, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.0614497661590576, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.696864902973175, + "num_tokens": 121502744.0, + "step": 4693 + }, + { + "epoch": 0.5154842960685262, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.087616205215454, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7219797968864441, + "num_tokens": 121530615.0, + "step": 4694 + }, + { + "epoch": 0.51559411377114, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.0506432056427, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7004724740982056, + "num_tokens": 121559108.0, + "step": 4695 + }, + { + "epoch": 0.5157039314737536, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.058943271636963, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.7077521085739136, + "num_tokens": 121588392.0, + "step": 4696 + }, + { + "epoch": 0.5158137491763672, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.01727032661438, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7091406583786011, + "num_tokens": 121617083.0, + "step": 4697 + }, + { + "epoch": 0.5159235668789809, + "ewc_loss": 4.2438507080078125e-05, + "grad_norm": 2.2720158100128174, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7138559818267822, + "num_tokens": 121641301.0, + "step": 4698 + }, + { + "epoch": 0.5160333845815945, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.0218822956085205, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7046212553977966, + "num_tokens": 121671011.0, + "step": 4699 + }, + { + "epoch": 0.5161432022842082, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.1518807411193848, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7143316864967346, + "num_tokens": 121697278.0, + "step": 4700 + }, + { + "epoch": 0.5162530199868218, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.1469430923461914, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.703027069568634, + "num_tokens": 121724693.0, + "step": 4701 + }, + { + "epoch": 0.5163628376894356, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.0911097526550293, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6973569393157959, + "num_tokens": 121753616.0, + "step": 4702 + }, + { + "epoch": 0.5164726553920492, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.0777668952941895, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.705203652381897, + "num_tokens": 121781118.0, + "step": 4703 + }, + { + "epoch": 0.5165824730946629, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.0884785652160645, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.720208466053009, + "num_tokens": 121807666.0, + "step": 4704 + }, + { + "epoch": 0.5166922907972765, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.5066978931427, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7319401502609253, + "num_tokens": 121826186.0, + "step": 4705 + }, + { + "epoch": 0.5168021084998902, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.18030047416687, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7108917236328125, + "num_tokens": 121853749.0, + "step": 4706 + }, + { + "epoch": 0.5169119262025038, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.274489164352417, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.698244571685791, + "num_tokens": 121879214.0, + "step": 4707 + }, + { + "epoch": 0.5170217439051175, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.191425323486328, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7040987610816956, + "num_tokens": 121905782.0, + "step": 4708 + }, + { + "epoch": 0.5171315616077312, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.073699951171875, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6903791427612305, + "num_tokens": 121935314.0, + "step": 4709 + }, + { + "epoch": 0.5172413793103449, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 1.9257935285568237, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7143470048904419, + "num_tokens": 121965963.0, + "step": 4710 + }, + { + "epoch": 0.5173511970129585, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 1.9724323749542236, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7071048021316528, + "num_tokens": 121994548.0, + "step": 4711 + }, + { + "epoch": 0.5174610147155722, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.09211802482605, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7163468599319458, + "num_tokens": 122021964.0, + "step": 4712 + }, + { + "epoch": 0.5175708324181858, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.5577659606933594, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7148623466491699, + "num_tokens": 122043758.0, + "step": 4713 + }, + { + "epoch": 0.5176806501207994, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.2566230297088623, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7128762006759644, + "num_tokens": 122068302.0, + "step": 4714 + }, + { + "epoch": 0.5177904678234131, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.2256922721862793, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7044693827629089, + "num_tokens": 122095220.0, + "step": 4715 + }, + { + "epoch": 0.5179002855260268, + "ewc_loss": 4.291534423828125e-05, + "grad_norm": 2.367271661758423, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7021632790565491, + "num_tokens": 122117264.0, + "step": 4716 + }, + { + "epoch": 0.5180101032286405, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.4628376960754395, + "learning_rate": 1e-06, + "loss": 1.113, + "mean_token_accuracy": 0.6607470512390137, + "num_tokens": 122140503.0, + "step": 4717 + }, + { + "epoch": 0.5181199209312541, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.293691873550415, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.6985887289047241, + "num_tokens": 122162514.0, + "step": 4718 + }, + { + "epoch": 0.5182297386338678, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.356818914413452, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6963430643081665, + "num_tokens": 122184999.0, + "step": 4719 + }, + { + "epoch": 0.5183395563364814, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.143918991088867, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7113760113716125, + "num_tokens": 122213918.0, + "step": 4720 + }, + { + "epoch": 0.5184493740390951, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.3494162559509277, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7070566415786743, + "num_tokens": 122237734.0, + "step": 4721 + }, + { + "epoch": 0.5185591917417087, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.30853533744812, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7006034851074219, + "num_tokens": 122260160.0, + "step": 4722 + }, + { + "epoch": 0.5186690094443224, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.089362621307373, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7200895547866821, + "num_tokens": 122286833.0, + "step": 4723 + }, + { + "epoch": 0.5187788271469361, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.1950464248657227, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7346713542938232, + "num_tokens": 122310433.0, + "step": 4724 + }, + { + "epoch": 0.5188886448495498, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.2744014263153076, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6983159780502319, + "num_tokens": 122334006.0, + "step": 4725 + }, + { + "epoch": 0.5189984625521634, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.13686466217041, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.706412672996521, + "num_tokens": 122361228.0, + "step": 4726 + }, + { + "epoch": 0.5191082802547771, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.208296298980713, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7217833399772644, + "num_tokens": 122386834.0, + "step": 4727 + }, + { + "epoch": 0.5192180979573907, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.2965400218963623, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.7003830075263977, + "num_tokens": 122411805.0, + "step": 4728 + }, + { + "epoch": 0.5193279156600044, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 1.9573696851730347, + "learning_rate": 1e-06, + "loss": 1.1064, + "mean_token_accuracy": 0.677676260471344, + "num_tokens": 122446482.0, + "step": 4729 + }, + { + "epoch": 0.519437733362618, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.1839730739593506, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7068856954574585, + "num_tokens": 122473380.0, + "step": 4730 + }, + { + "epoch": 0.5195475510652318, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.357349157333374, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7202277183532715, + "num_tokens": 122497542.0, + "step": 4731 + }, + { + "epoch": 0.5196573687678454, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.0384230613708496, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6908014416694641, + "num_tokens": 122526094.0, + "step": 4732 + }, + { + "epoch": 0.519767186470459, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.041274070739746, + "learning_rate": 1e-06, + "loss": 1.0425, + "mean_token_accuracy": 0.6910311579704285, + "num_tokens": 122555978.0, + "step": 4733 + }, + { + "epoch": 0.5198770041730727, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.086479663848877, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.704316258430481, + "num_tokens": 122584360.0, + "step": 4734 + }, + { + "epoch": 0.5199868218756863, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.3483643531799316, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.69582200050354, + "num_tokens": 122606846.0, + "step": 4735 + }, + { + "epoch": 0.5200966395783, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.57814884185791, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7173050045967102, + "num_tokens": 122626112.0, + "step": 4736 + }, + { + "epoch": 0.5202064572809136, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.2907934188842773, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6930645704269409, + "num_tokens": 122650732.0, + "step": 4737 + }, + { + "epoch": 0.5203162749835274, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.285003900527954, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7114592790603638, + "num_tokens": 122674126.0, + "step": 4738 + }, + { + "epoch": 0.520426092686141, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.4131717681884766, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7095590233802795, + "num_tokens": 122697488.0, + "step": 4739 + }, + { + "epoch": 0.5205359103887547, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.3607430458068848, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.7011323571205139, + "num_tokens": 122722311.0, + "step": 4740 + }, + { + "epoch": 0.5206457280913683, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.2245442867279053, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7088432312011719, + "num_tokens": 122748210.0, + "step": 4741 + }, + { + "epoch": 0.520755545793982, + "ewc_loss": 4.315376281738281e-05, + "grad_norm": 2.3196969032287598, + "learning_rate": 1e-06, + "loss": 1.0637, + "mean_token_accuracy": 0.6873059272766113, + "num_tokens": 122772245.0, + "step": 4742 + }, + { + "epoch": 0.5208653634965956, + "ewc_loss": 4.3392181396484375e-05, + "grad_norm": 2.146186590194702, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6998252272605896, + "num_tokens": 122798903.0, + "step": 4743 + }, + { + "epoch": 0.5209751811992093, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.707961320877075, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.6987311840057373, + "num_tokens": 122818671.0, + "step": 4744 + }, + { + "epoch": 0.521084998901823, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.112220525741577, + "learning_rate": 1e-06, + "loss": 1.1372, + "mean_token_accuracy": 0.6675840020179749, + "num_tokens": 122849140.0, + "step": 4745 + }, + { + "epoch": 0.5211948166044367, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.067802667617798, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.69374680519104, + "num_tokens": 122879171.0, + "step": 4746 + }, + { + "epoch": 0.5213046343070503, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.139972448348999, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6967774629592896, + "num_tokens": 122908443.0, + "step": 4747 + }, + { + "epoch": 0.521414452009664, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.39894437789917, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7179943919181824, + "num_tokens": 122928803.0, + "step": 4748 + }, + { + "epoch": 0.5215242697122776, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.2126994132995605, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7018377780914307, + "num_tokens": 122954480.0, + "step": 4749 + }, + { + "epoch": 0.5216340874148913, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.1289596557617188, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7115045189857483, + "num_tokens": 122982025.0, + "step": 4750 + }, + { + "epoch": 0.5217439051175049, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.281670570373535, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7061696648597717, + "num_tokens": 123009122.0, + "step": 4751 + }, + { + "epoch": 0.5218537228201185, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.252319097518921, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7259699106216431, + "num_tokens": 123034418.0, + "step": 4752 + }, + { + "epoch": 0.5219635405227323, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.421076774597168, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7045801281929016, + "num_tokens": 123055576.0, + "step": 4753 + }, + { + "epoch": 0.522073358225346, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.6405997276306152, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7037702798843384, + "num_tokens": 123074540.0, + "step": 4754 + }, + { + "epoch": 0.5221831759279596, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.467102527618408, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7061846256256104, + "num_tokens": 123097449.0, + "step": 4755 + }, + { + "epoch": 0.5222929936305732, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.198798179626465, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6914783716201782, + "num_tokens": 123125290.0, + "step": 4756 + }, + { + "epoch": 0.5224028113331869, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 1.9797478914260864, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.6986657381057739, + "num_tokens": 123155920.0, + "step": 4757 + }, + { + "epoch": 0.5225126290358005, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.3443245887756348, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7182776927947998, + "num_tokens": 123176959.0, + "step": 4758 + }, + { + "epoch": 0.5226224467384142, + "ewc_loss": 4.363059997558594e-05, + "grad_norm": 2.197237491607666, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6908215284347534, + "num_tokens": 123202287.0, + "step": 4759 + }, + { + "epoch": 0.5227322644410279, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.3460750579833984, + "learning_rate": 1e-06, + "loss": 1.0687, + "mean_token_accuracy": 0.6883915662765503, + "num_tokens": 123225282.0, + "step": 4760 + }, + { + "epoch": 0.5228420821436416, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2987470626831055, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6807085275650024, + "num_tokens": 123251470.0, + "step": 4761 + }, + { + "epoch": 0.5229518998462552, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.026188611984253, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.6877865791320801, + "num_tokens": 123280914.0, + "step": 4762 + }, + { + "epoch": 0.5230617175488689, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1720774173736572, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7107458114624023, + "num_tokens": 123307373.0, + "step": 4763 + }, + { + "epoch": 0.5231715352514825, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 1.9405808448791504, + "learning_rate": 1e-06, + "loss": 1.063, + "mean_token_accuracy": 0.6802939176559448, + "num_tokens": 123341593.0, + "step": 4764 + }, + { + "epoch": 0.5232813529540962, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.222755193710327, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.711882472038269, + "num_tokens": 123367583.0, + "step": 4765 + }, + { + "epoch": 0.5233911706567098, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 1.844638466835022, + "learning_rate": 1e-06, + "loss": 1.0525, + "mean_token_accuracy": 0.6885955333709717, + "num_tokens": 123404988.0, + "step": 4766 + }, + { + "epoch": 0.5235009883593236, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1382992267608643, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7135474681854248, + "num_tokens": 123429995.0, + "step": 4767 + }, + { + "epoch": 0.5236108060619372, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.0899057388305664, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6945523619651794, + "num_tokens": 123458296.0, + "step": 4768 + }, + { + "epoch": 0.5237206237645509, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.13189959526062, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.7015975117683411, + "num_tokens": 123484696.0, + "step": 4769 + }, + { + "epoch": 0.5238304414671645, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2190256118774414, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6983578205108643, + "num_tokens": 123511954.0, + "step": 4770 + }, + { + "epoch": 0.5239402591697782, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.5253303050994873, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7169229984283447, + "num_tokens": 123532302.0, + "step": 4771 + }, + { + "epoch": 0.5240500768723918, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.324275255203247, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.7087024450302124, + "num_tokens": 123557394.0, + "step": 4772 + }, + { + "epoch": 0.5241598945750054, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.3811073303222656, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6913769245147705, + "num_tokens": 123581035.0, + "step": 4773 + }, + { + "epoch": 0.5242697122776192, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1146926879882812, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.7020636796951294, + "num_tokens": 123609927.0, + "step": 4774 + }, + { + "epoch": 0.5243795299802329, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2793262004852295, + "learning_rate": 1e-06, + "loss": 1.0539, + "mean_token_accuracy": 0.6855785846710205, + "num_tokens": 123636801.0, + "step": 4775 + }, + { + "epoch": 0.5244893476828465, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1253581047058105, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.723822832107544, + "num_tokens": 123662412.0, + "step": 4776 + }, + { + "epoch": 0.5245991653854601, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2165684700012207, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.690340518951416, + "num_tokens": 123688200.0, + "step": 4777 + }, + { + "epoch": 0.5247089830880738, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.633122205734253, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7030428647994995, + "num_tokens": 123707255.0, + "step": 4778 + }, + { + "epoch": 0.5248188007906874, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.374875783920288, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7103123068809509, + "num_tokens": 123730727.0, + "step": 4779 + }, + { + "epoch": 0.5249286184933011, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.4006545543670654, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7432233095169067, + "num_tokens": 123750769.0, + "step": 4780 + }, + { + "epoch": 0.5250384361959147, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.18144154548645, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7188555002212524, + "num_tokens": 123774443.0, + "step": 4781 + }, + { + "epoch": 0.5251482538985285, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.468803882598877, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7023547887802124, + "num_tokens": 123797174.0, + "step": 4782 + }, + { + "epoch": 0.5252580716011421, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.240126132965088, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.6968764066696167, + "num_tokens": 123823711.0, + "step": 4783 + }, + { + "epoch": 0.5253678893037558, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.3656842708587646, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6905674338340759, + "num_tokens": 123847343.0, + "step": 4784 + }, + { + "epoch": 0.5254777070063694, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2451069355010986, + "learning_rate": 1e-06, + "loss": 1.0445, + "mean_token_accuracy": 0.6853071451187134, + "num_tokens": 123872690.0, + "step": 4785 + }, + { + "epoch": 0.5255875247089831, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.2908735275268555, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.7038078904151917, + "num_tokens": 123897410.0, + "step": 4786 + }, + { + "epoch": 0.5256973424115967, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.3490564823150635, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6933466196060181, + "num_tokens": 123921194.0, + "step": 4787 + }, + { + "epoch": 0.5258071601142104, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1584110260009766, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7094135284423828, + "num_tokens": 123946528.0, + "step": 4788 + }, + { + "epoch": 0.5259169778168241, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.0873324871063232, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7042921185493469, + "num_tokens": 123974449.0, + "step": 4789 + }, + { + "epoch": 0.5260267955194378, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.4339351654052734, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7194806337356567, + "num_tokens": 123994838.0, + "step": 4790 + }, + { + "epoch": 0.5261366132220514, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.142285108566284, + "learning_rate": 1e-06, + "loss": 1.1213, + "mean_token_accuracy": 0.6652201414108276, + "num_tokens": 124025635.0, + "step": 4791 + }, + { + "epoch": 0.526246430924665, + "ewc_loss": 4.38690185546875e-05, + "grad_norm": 2.1969425678253174, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.7012902498245239, + "num_tokens": 124052257.0, + "step": 4792 + }, + { + "epoch": 0.5263562486272787, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.3101465702056885, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7065824866294861, + "num_tokens": 124076722.0, + "step": 4793 + }, + { + "epoch": 0.5264660663298923, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.6831016540527344, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.691315770149231, + "num_tokens": 124095514.0, + "step": 4794 + }, + { + "epoch": 0.526575884032506, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.1635854244232178, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6754776239395142, + "num_tokens": 124123063.0, + "step": 4795 + }, + { + "epoch": 0.5266857017351197, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2447919845581055, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6891334652900696, + "num_tokens": 124146795.0, + "step": 4796 + }, + { + "epoch": 0.5267955194377334, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2777323722839355, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7087419629096985, + "num_tokens": 124170669.0, + "step": 4797 + }, + { + "epoch": 0.526905337140347, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.174065113067627, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.690306544303894, + "num_tokens": 124195844.0, + "step": 4798 + }, + { + "epoch": 0.5270151548429607, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.1751387119293213, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6857043504714966, + "num_tokens": 124223053.0, + "step": 4799 + }, + { + "epoch": 0.5271249725455743, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.151409149169922, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6978546977043152, + "num_tokens": 124248218.0, + "step": 4800 + }, + { + "epoch": 0.527234790248188, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.146904468536377, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.7185846567153931, + "num_tokens": 124273859.0, + "step": 4801 + }, + { + "epoch": 0.5273446079508016, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.513221263885498, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.715942919254303, + "num_tokens": 124294213.0, + "step": 4802 + }, + { + "epoch": 0.5274544256534154, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.085453748703003, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.6998458504676819, + "num_tokens": 124319793.0, + "step": 4803 + }, + { + "epoch": 0.527564243356029, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2013797760009766, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7076995372772217, + "num_tokens": 124346866.0, + "step": 4804 + }, + { + "epoch": 0.5276740610586427, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.416071653366089, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.705844521522522, + "num_tokens": 124368693.0, + "step": 4805 + }, + { + "epoch": 0.5277838787612563, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.5225377082824707, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6853718757629395, + "num_tokens": 124388582.0, + "step": 4806 + }, + { + "epoch": 0.52789369646387, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.169337272644043, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.71547532081604, + "num_tokens": 124413481.0, + "step": 4807 + }, + { + "epoch": 0.5280035141664836, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.276339054107666, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7112227082252502, + "num_tokens": 124436784.0, + "step": 4808 + }, + { + "epoch": 0.5281133318690973, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2653069496154785, + "learning_rate": 1e-06, + "loss": 1.0365, + "mean_token_accuracy": 0.6884147524833679, + "num_tokens": 124460433.0, + "step": 4809 + }, + { + "epoch": 0.528223149571711, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2897958755493164, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6898234486579895, + "num_tokens": 124484380.0, + "step": 4810 + }, + { + "epoch": 0.5283329672743247, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2113194465637207, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7133982181549072, + "num_tokens": 124509622.0, + "step": 4811 + }, + { + "epoch": 0.5284427849769383, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 1.9591481685638428, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6898829936981201, + "num_tokens": 124540590.0, + "step": 4812 + }, + { + "epoch": 0.528552602679552, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.5840914249420166, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7085109353065491, + "num_tokens": 124562283.0, + "step": 4813 + }, + { + "epoch": 0.5286624203821656, + "ewc_loss": 4.458427429199219e-05, + "grad_norm": 2.070664167404175, + "learning_rate": 1e-06, + "loss": 1.068, + "mean_token_accuracy": 0.682929277420044, + "num_tokens": 124594488.0, + "step": 4814 + }, + { + "epoch": 0.5287722380847792, + "ewc_loss": 4.458427429199219e-05, + "grad_norm": 2.268096923828125, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7126675248146057, + "num_tokens": 124621489.0, + "step": 4815 + }, + { + "epoch": 0.5288820557873929, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 1.973243236541748, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.740399181842804, + "num_tokens": 124649436.0, + "step": 4816 + }, + { + "epoch": 0.5289918734900065, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.408613443374634, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7131386399269104, + "num_tokens": 124671012.0, + "step": 4817 + }, + { + "epoch": 0.5291016911926203, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2733402252197266, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6955164670944214, + "num_tokens": 124694958.0, + "step": 4818 + }, + { + "epoch": 0.5292115088952339, + "ewc_loss": 4.458427429199219e-05, + "grad_norm": 2.4981420040130615, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7056183815002441, + "num_tokens": 124715344.0, + "step": 4819 + }, + { + "epoch": 0.5293213265978476, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1583056449890137, + "learning_rate": 1e-06, + "loss": 1.1348, + "mean_token_accuracy": 0.6663763523101807, + "num_tokens": 124743751.0, + "step": 4820 + }, + { + "epoch": 0.5294311443004612, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2668092250823975, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7252907752990723, + "num_tokens": 124769264.0, + "step": 4821 + }, + { + "epoch": 0.5295409620030749, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2138712406158447, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6948398351669312, + "num_tokens": 124796623.0, + "step": 4822 + }, + { + "epoch": 0.5296507797056885, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.117276668548584, + "learning_rate": 1e-06, + "loss": 1.0928, + "mean_token_accuracy": 0.672522783279419, + "num_tokens": 124825236.0, + "step": 4823 + }, + { + "epoch": 0.5297605974083022, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2420687675476074, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6946814656257629, + "num_tokens": 124849570.0, + "step": 4824 + }, + { + "epoch": 0.5298704151109159, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.157378911972046, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6878647208213806, + "num_tokens": 124876140.0, + "step": 4825 + }, + { + "epoch": 0.5299802328135296, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.0912928581237793, + "learning_rate": 1e-06, + "loss": 1.036, + "mean_token_accuracy": 0.6908574104309082, + "num_tokens": 124902546.0, + "step": 4826 + }, + { + "epoch": 0.5300900505161432, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.2769627571105957, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7040148377418518, + "num_tokens": 124926641.0, + "step": 4827 + }, + { + "epoch": 0.5301998682187569, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.0328338146209717, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.70071941614151, + "num_tokens": 124955451.0, + "step": 4828 + }, + { + "epoch": 0.5303096859213705, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.0481271743774414, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7138681411743164, + "num_tokens": 124983402.0, + "step": 4829 + }, + { + "epoch": 0.5304195036239842, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.224855899810791, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.7016208171844482, + "num_tokens": 125008433.0, + "step": 4830 + }, + { + "epoch": 0.5305293213265978, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.0459163188934326, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.6989238858222961, + "num_tokens": 125036151.0, + "step": 4831 + }, + { + "epoch": 0.5306391390292116, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.3861968517303467, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7214826345443726, + "num_tokens": 125058766.0, + "step": 4832 + }, + { + "epoch": 0.5307489567318252, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.140878200531006, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7129274606704712, + "num_tokens": 125084381.0, + "step": 4833 + }, + { + "epoch": 0.5308587744344389, + "ewc_loss": 4.4345855712890625e-05, + "grad_norm": 2.1851494312286377, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6873704195022583, + "num_tokens": 125108921.0, + "step": 4834 + }, + { + "epoch": 0.5309685921370525, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.0025436878204346, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7030213475227356, + "num_tokens": 125140054.0, + "step": 4835 + }, + { + "epoch": 0.5310784098396661, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1075971126556396, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6871902942657471, + "num_tokens": 125167042.0, + "step": 4836 + }, + { + "epoch": 0.5311882275422798, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.0092990398406982, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7151095271110535, + "num_tokens": 125195924.0, + "step": 4837 + }, + { + "epoch": 0.5312980452448934, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.2253503799438477, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.6993634104728699, + "num_tokens": 125221340.0, + "step": 4838 + }, + { + "epoch": 0.5314078629475072, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.3149280548095703, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.6977315545082092, + "num_tokens": 125243894.0, + "step": 4839 + }, + { + "epoch": 0.5315176806501208, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.2824461460113525, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7120614051818848, + "num_tokens": 125268549.0, + "step": 4840 + }, + { + "epoch": 0.5316274983527345, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.175250291824341, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6993810534477234, + "num_tokens": 125295550.0, + "step": 4841 + }, + { + "epoch": 0.5317373160553481, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.211150884628296, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.701622724533081, + "num_tokens": 125320744.0, + "step": 4842 + }, + { + "epoch": 0.5318471337579618, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.183522939682007, + "learning_rate": 1e-06, + "loss": 1.0535, + "mean_token_accuracy": 0.686982274055481, + "num_tokens": 125348066.0, + "step": 4843 + }, + { + "epoch": 0.5319569514605754, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 1.9119093418121338, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6955580711364746, + "num_tokens": 125378429.0, + "step": 4844 + }, + { + "epoch": 0.5320667691631891, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1683762073516846, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.6952105760574341, + "num_tokens": 125403980.0, + "step": 4845 + }, + { + "epoch": 0.5321765868658027, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 1.969749927520752, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6930370330810547, + "num_tokens": 125434953.0, + "step": 4846 + }, + { + "epoch": 0.5322864045684165, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.2171921730041504, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7023815512657166, + "num_tokens": 125461614.0, + "step": 4847 + }, + { + "epoch": 0.5323962222710301, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 1.9827600717544556, + "learning_rate": 1e-06, + "loss": 1.0848, + "mean_token_accuracy": 0.6792038679122925, + "num_tokens": 125494269.0, + "step": 4848 + }, + { + "epoch": 0.5325060399736438, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1563565731048584, + "learning_rate": 1e-06, + "loss": 1.0841, + "mean_token_accuracy": 0.6716914176940918, + "num_tokens": 125523086.0, + "step": 4849 + }, + { + "epoch": 0.5326158576762574, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.324456214904785, + "learning_rate": 1e-06, + "loss": 1.0443, + "mean_token_accuracy": 0.6909180879592896, + "num_tokens": 125548116.0, + "step": 4850 + }, + { + "epoch": 0.532725675378871, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1418235301971436, + "learning_rate": 1e-06, + "loss": 1.0819, + "mean_token_accuracy": 0.6739262342453003, + "num_tokens": 125576092.0, + "step": 4851 + }, + { + "epoch": 0.5328354930814847, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.1146433353424072, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7042815089225769, + "num_tokens": 125606678.0, + "step": 4852 + }, + { + "epoch": 0.5329453107840983, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.353353500366211, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7051597237586975, + "num_tokens": 125628470.0, + "step": 4853 + }, + { + "epoch": 0.5330551284867121, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.476231813430786, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6935411691665649, + "num_tokens": 125649194.0, + "step": 4854 + }, + { + "epoch": 0.5331649461893258, + "ewc_loss": 4.482269287109375e-05, + "grad_norm": 2.0697526931762695, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6884319186210632, + "num_tokens": 125681786.0, + "step": 4855 + }, + { + "epoch": 0.5332747638919394, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1187033653259277, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7062058448791504, + "num_tokens": 125708871.0, + "step": 4856 + }, + { + "epoch": 0.533384581594553, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.109269618988037, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7074003219604492, + "num_tokens": 125736407.0, + "step": 4857 + }, + { + "epoch": 0.5334943992971667, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.0972940921783447, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6967320442199707, + "num_tokens": 125764625.0, + "step": 4858 + }, + { + "epoch": 0.5336042169997803, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.3156745433807373, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7002931237220764, + "num_tokens": 125788811.0, + "step": 4859 + }, + { + "epoch": 0.533714034702394, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.0506956577301025, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.6961473226547241, + "num_tokens": 125816450.0, + "step": 4860 + }, + { + "epoch": 0.5338238524050077, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.2266581058502197, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.702446699142456, + "num_tokens": 125843742.0, + "step": 4861 + }, + { + "epoch": 0.5339336701076214, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.450324296951294, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6878477334976196, + "num_tokens": 125867806.0, + "step": 4862 + }, + { + "epoch": 0.534043487810235, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.0439577102661133, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6929945349693298, + "num_tokens": 125898897.0, + "step": 4863 + }, + { + "epoch": 0.5341533055128487, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.136930465698242, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.715114951133728, + "num_tokens": 125926333.0, + "step": 4864 + }, + { + "epoch": 0.5342631232154623, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1730339527130127, + "learning_rate": 1e-06, + "loss": 1.0517, + "mean_token_accuracy": 0.6822370290756226, + "num_tokens": 125955239.0, + "step": 4865 + }, + { + "epoch": 0.534372940918076, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9984874725341797, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7108604907989502, + "num_tokens": 125984817.0, + "step": 4866 + }, + { + "epoch": 0.5344827586206896, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.4622879028320312, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.706967830657959, + "num_tokens": 126007118.0, + "step": 4867 + }, + { + "epoch": 0.5345925763233034, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1607038974761963, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7042219042778015, + "num_tokens": 126033833.0, + "step": 4868 + }, + { + "epoch": 0.534702394025917, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1772100925445557, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6954182386398315, + "num_tokens": 126058969.0, + "step": 4869 + }, + { + "epoch": 0.5348122117285307, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.03409481048584, + "learning_rate": 1e-06, + "loss": 1.0471, + "mean_token_accuracy": 0.6828432679176331, + "num_tokens": 126087601.0, + "step": 4870 + }, + { + "epoch": 0.5349220294311443, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.503474712371826, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7002437114715576, + "num_tokens": 126108863.0, + "step": 4871 + }, + { + "epoch": 0.535031847133758, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.100245475769043, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7151370048522949, + "num_tokens": 126135520.0, + "step": 4872 + }, + { + "epoch": 0.5351416648363716, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.232487678527832, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7012922763824463, + "num_tokens": 126159142.0, + "step": 4873 + }, + { + "epoch": 0.5352514825389852, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.065122604370117, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7115432620048523, + "num_tokens": 126185981.0, + "step": 4874 + }, + { + "epoch": 0.5353613002415989, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.2381980419158936, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6954928040504456, + "num_tokens": 126211810.0, + "step": 4875 + }, + { + "epoch": 0.5354711179442126, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1957786083221436, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7049199342727661, + "num_tokens": 126236974.0, + "step": 4876 + }, + { + "epoch": 0.5355809356468263, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.303929567337036, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7018535137176514, + "num_tokens": 126261310.0, + "step": 4877 + }, + { + "epoch": 0.5356907533494399, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.6397933959960938, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.69264817237854, + "num_tokens": 126281643.0, + "step": 4878 + }, + { + "epoch": 0.5358005710520536, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.0637543201446533, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.689070999622345, + "num_tokens": 126312906.0, + "step": 4879 + }, + { + "epoch": 0.5359103887546672, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9409468173980713, + "learning_rate": 1e-06, + "loss": 1.1114, + "mean_token_accuracy": 0.6680139303207397, + "num_tokens": 126345375.0, + "step": 4880 + }, + { + "epoch": 0.5360202064572809, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.041764497756958, + "learning_rate": 1e-06, + "loss": 1.0602, + "mean_token_accuracy": 0.6804949641227722, + "num_tokens": 126376856.0, + "step": 4881 + }, + { + "epoch": 0.5361300241598945, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.2402989864349365, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6989611387252808, + "num_tokens": 126401481.0, + "step": 4882 + }, + { + "epoch": 0.5362398418625083, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.6323838233947754, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6996898651123047, + "num_tokens": 126421653.0, + "step": 4883 + }, + { + "epoch": 0.5363496595651219, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.2538583278656006, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.6899542808532715, + "num_tokens": 126446537.0, + "step": 4884 + }, + { + "epoch": 0.5364594772677356, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.220787286758423, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7293376326560974, + "num_tokens": 126472324.0, + "step": 4885 + }, + { + "epoch": 0.5365692949703492, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1455600261688232, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7278873920440674, + "num_tokens": 126498138.0, + "step": 4886 + }, + { + "epoch": 0.5366791126729629, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.2916712760925293, + "learning_rate": 1e-06, + "loss": 1.0629, + "mean_token_accuracy": 0.6772885918617249, + "num_tokens": 126524152.0, + "step": 4887 + }, + { + "epoch": 0.5367889303755765, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1641459465026855, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7022154331207275, + "num_tokens": 126551253.0, + "step": 4888 + }, + { + "epoch": 0.5368987480781902, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.91568124294281, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6908735036849976, + "num_tokens": 126583556.0, + "step": 4889 + }, + { + "epoch": 0.5370085657808039, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.5734591484069824, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7171326875686646, + "num_tokens": 126602628.0, + "step": 4890 + }, + { + "epoch": 0.5371183834834176, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.03318190574646, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7095943689346313, + "num_tokens": 126630945.0, + "step": 4891 + }, + { + "epoch": 0.5372282011860312, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.005723237991333, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.7036113142967224, + "num_tokens": 126660562.0, + "step": 4892 + }, + { + "epoch": 0.5373380188886449, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9592509269714355, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7042230367660522, + "num_tokens": 126691431.0, + "step": 4893 + }, + { + "epoch": 0.5374478365912585, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.0509164333343506, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7052746415138245, + "num_tokens": 126718942.0, + "step": 4894 + }, + { + "epoch": 0.5375576542938721, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.7588984966278076, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7302440404891968, + "num_tokens": 126737774.0, + "step": 4895 + }, + { + "epoch": 0.5376674719964858, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.266073226928711, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6971694231033325, + "num_tokens": 126764676.0, + "step": 4896 + }, + { + "epoch": 0.5377772896990995, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.314710855484009, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6993721723556519, + "num_tokens": 126790046.0, + "step": 4897 + }, + { + "epoch": 0.5378871074017132, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.6264612674713135, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7367689609527588, + "num_tokens": 126808339.0, + "step": 4898 + }, + { + "epoch": 0.5379969251043268, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1981818675994873, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7255030274391174, + "num_tokens": 126832548.0, + "step": 4899 + }, + { + "epoch": 0.5381067428069405, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9579734802246094, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6953357458114624, + "num_tokens": 126864910.0, + "step": 4900 + }, + { + "epoch": 0.5382165605095541, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.131560802459717, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6935091018676758, + "num_tokens": 126893270.0, + "step": 4901 + }, + { + "epoch": 0.5383263782121678, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.092369794845581, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6983718872070312, + "num_tokens": 126921753.0, + "step": 4902 + }, + { + "epoch": 0.5384361959147814, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9474846124649048, + "learning_rate": 1e-06, + "loss": 1.0521, + "mean_token_accuracy": 0.6863795518875122, + "num_tokens": 126955918.0, + "step": 4903 + }, + { + "epoch": 0.5385460136173951, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.3440210819244385, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7104687094688416, + "num_tokens": 126978163.0, + "step": 4904 + }, + { + "epoch": 0.5386558313200088, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.3713409900665283, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.7049317359924316, + "num_tokens": 127001935.0, + "step": 4905 + }, + { + "epoch": 0.5387656490226225, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.1859729290008545, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7052609920501709, + "num_tokens": 127026364.0, + "step": 4906 + }, + { + "epoch": 0.5388754667252361, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.3157262802124023, + "learning_rate": 1e-06, + "loss": 1.0424, + "mean_token_accuracy": 0.6863862872123718, + "num_tokens": 127051022.0, + "step": 4907 + }, + { + "epoch": 0.5389852844278498, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.266629457473755, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6969425678253174, + "num_tokens": 127077302.0, + "step": 4908 + }, + { + "epoch": 0.5390951021304634, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 1.9924569129943848, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6966937780380249, + "num_tokens": 127107594.0, + "step": 4909 + }, + { + "epoch": 0.5392049198330771, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.4113941192626953, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7059034109115601, + "num_tokens": 127129814.0, + "step": 4910 + }, + { + "epoch": 0.5393147375356907, + "ewc_loss": 4.506111145019531e-05, + "grad_norm": 2.5440914630889893, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7155709862709045, + "num_tokens": 127148691.0, + "step": 4911 + }, + { + "epoch": 0.5394245552383045, + "ewc_loss": 4.5299530029296875e-05, + "grad_norm": 2.3160479068756104, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7359152436256409, + "num_tokens": 127168865.0, + "step": 4912 + }, + { + "epoch": 0.5395343729409181, + "ewc_loss": 4.5299530029296875e-05, + "grad_norm": 2.0063204765319824, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6966543197631836, + "num_tokens": 127199139.0, + "step": 4913 + }, + { + "epoch": 0.5396441906435318, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.5010242462158203, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7109977602958679, + "num_tokens": 127219551.0, + "step": 4914 + }, + { + "epoch": 0.5397540083461454, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.1440491676330566, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7072355151176453, + "num_tokens": 127247441.0, + "step": 4915 + }, + { + "epoch": 0.539863826048759, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.901003360748291, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7087859511375427, + "num_tokens": 127280922.0, + "step": 4916 + }, + { + "epoch": 0.5399736437513727, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.179598331451416, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7224689722061157, + "num_tokens": 127303976.0, + "step": 4917 + }, + { + "epoch": 0.5400834614539863, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.1463823318481445, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7241319417953491, + "num_tokens": 127329285.0, + "step": 4918 + }, + { + "epoch": 0.5401932791566001, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.350748300552368, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7193582653999329, + "num_tokens": 127352261.0, + "step": 4919 + }, + { + "epoch": 0.5403030968592137, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.136995315551758, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.702325701713562, + "num_tokens": 127380086.0, + "step": 4920 + }, + { + "epoch": 0.5404129145618274, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.450148582458496, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7170778512954712, + "num_tokens": 127402010.0, + "step": 4921 + }, + { + "epoch": 0.540522732264441, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.2168357372283936, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6904445886611938, + "num_tokens": 127429070.0, + "step": 4922 + }, + { + "epoch": 0.5406325499670547, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.003744602203369, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.711638331413269, + "num_tokens": 127458094.0, + "step": 4923 + }, + { + "epoch": 0.5407423676696683, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.0607383251190186, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7030243873596191, + "num_tokens": 127487961.0, + "step": 4924 + }, + { + "epoch": 0.540852185372282, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.225757122039795, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7025831937789917, + "num_tokens": 127514020.0, + "step": 4925 + }, + { + "epoch": 0.5409620030748957, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.116767644882202, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.691493034362793, + "num_tokens": 127542668.0, + "step": 4926 + }, + { + "epoch": 0.5410718207775094, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.5097320079803467, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7058719396591187, + "num_tokens": 127562630.0, + "step": 4927 + }, + { + "epoch": 0.541181638480123, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.1061105728149414, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7043823003768921, + "num_tokens": 127592931.0, + "step": 4928 + }, + { + "epoch": 0.5412914561827367, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.114574909210205, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6918333172798157, + "num_tokens": 127621487.0, + "step": 4929 + }, + { + "epoch": 0.5414012738853503, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.982399821281433, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7294061183929443, + "num_tokens": 127650196.0, + "step": 4930 + }, + { + "epoch": 0.541511091587964, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.343053102493286, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7171409130096436, + "num_tokens": 127673970.0, + "step": 4931 + }, + { + "epoch": 0.5416209092905776, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.0757102966308594, + "learning_rate": 1e-06, + "loss": 1.0369, + "mean_token_accuracy": 0.6871682405471802, + "num_tokens": 127703486.0, + "step": 4932 + }, + { + "epoch": 0.5417307269931912, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.0636179447174072, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.705010712146759, + "num_tokens": 127731693.0, + "step": 4933 + }, + { + "epoch": 0.541840544695805, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.265148878097534, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7243518233299255, + "num_tokens": 127754534.0, + "step": 4934 + }, + { + "epoch": 0.5419503623984187, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.9626272916793823, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6980694532394409, + "num_tokens": 127785708.0, + "step": 4935 + }, + { + "epoch": 0.5420601801010323, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.3942019939422607, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7258023619651794, + "num_tokens": 127807870.0, + "step": 4936 + }, + { + "epoch": 0.5421699978036459, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.3449325561523438, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6953964829444885, + "num_tokens": 127830538.0, + "step": 4937 + }, + { + "epoch": 0.5422798155062596, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.039443016052246, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.6764646768569946, + "num_tokens": 127861733.0, + "step": 4938 + }, + { + "epoch": 0.5423896332088732, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.2602665424346924, + "learning_rate": 1e-06, + "loss": 1.0459, + "mean_token_accuracy": 0.688332736492157, + "num_tokens": 127886190.0, + "step": 4939 + }, + { + "epoch": 0.5424994509114869, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.1579172611236572, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6971899271011353, + "num_tokens": 127914203.0, + "step": 4940 + }, + { + "epoch": 0.5426092686141006, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.091181516647339, + "learning_rate": 1e-06, + "loss": 1.0606, + "mean_token_accuracy": 0.6846902370452881, + "num_tokens": 127942783.0, + "step": 4941 + }, + { + "epoch": 0.5427190863167143, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.2698957920074463, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6970099210739136, + "num_tokens": 127967032.0, + "step": 4942 + }, + { + "epoch": 0.5428289040193279, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.4573607444763184, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.716038703918457, + "num_tokens": 127988188.0, + "step": 4943 + }, + { + "epoch": 0.5429387217219416, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.4006264209747314, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6905074715614319, + "num_tokens": 128012088.0, + "step": 4944 + }, + { + "epoch": 0.5430485394245552, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.234666585922241, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.698321521282196, + "num_tokens": 128037208.0, + "step": 4945 + }, + { + "epoch": 0.5431583571271689, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.168545722961426, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7323026061058044, + "num_tokens": 128062799.0, + "step": 4946 + }, + { + "epoch": 0.5432681748297825, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.9965485334396362, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7161388397216797, + "num_tokens": 128093529.0, + "step": 4947 + }, + { + "epoch": 0.5433779925323963, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.98640775680542, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6797082424163818, + "num_tokens": 128123690.0, + "step": 4948 + }, + { + "epoch": 0.5434878102350099, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.2855966091156006, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6934511661529541, + "num_tokens": 128148433.0, + "step": 4949 + }, + { + "epoch": 0.5435976279376236, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.2102296352386475, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.6871203184127808, + "num_tokens": 128173520.0, + "step": 4950 + }, + { + "epoch": 0.5437074456402372, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 1.9967375993728638, + "learning_rate": 1e-06, + "loss": 1.039, + "mean_token_accuracy": 0.7017391920089722, + "num_tokens": 128205469.0, + "step": 4951 + }, + { + "epoch": 0.5438172633428509, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.263607978820801, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7290489673614502, + "num_tokens": 128228251.0, + "step": 4952 + }, + { + "epoch": 0.5439270810454645, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.0430548191070557, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7118375301361084, + "num_tokens": 128255630.0, + "step": 4953 + }, + { + "epoch": 0.5440368987480781, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.368687391281128, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7077975273132324, + "num_tokens": 128278625.0, + "step": 4954 + }, + { + "epoch": 0.5441467164506919, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.3854031562805176, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.717140257358551, + "num_tokens": 128299970.0, + "step": 4955 + }, + { + "epoch": 0.5442565341533055, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.1867635250091553, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.711250901222229, + "num_tokens": 128325057.0, + "step": 4956 + }, + { + "epoch": 0.5443663518559192, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.4036386013031006, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.7001302242279053, + "num_tokens": 128348579.0, + "step": 4957 + }, + { + "epoch": 0.5444761695585328, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.162039279937744, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7131496071815491, + "num_tokens": 128374047.0, + "step": 4958 + }, + { + "epoch": 0.5445859872611465, + "ewc_loss": 4.553794860839844e-05, + "grad_norm": 2.608863353729248, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7055679559707642, + "num_tokens": 128394609.0, + "step": 4959 + }, + { + "epoch": 0.5446958049637601, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.1736738681793213, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.699780285358429, + "num_tokens": 128419846.0, + "step": 4960 + }, + { + "epoch": 0.5448056226663738, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 1.8686461448669434, + "learning_rate": 1e-06, + "loss": 1.08, + "mean_token_accuracy": 0.6769251227378845, + "num_tokens": 128453276.0, + "step": 4961 + }, + { + "epoch": 0.5449154403689875, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.247675895690918, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.698022723197937, + "num_tokens": 128476588.0, + "step": 4962 + }, + { + "epoch": 0.5450252580716012, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.145662307739258, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7012749910354614, + "num_tokens": 128503964.0, + "step": 4963 + }, + { + "epoch": 0.5451350757742148, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.0801029205322266, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7173652052879333, + "num_tokens": 128531796.0, + "step": 4964 + }, + { + "epoch": 0.5452448934768285, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.2545766830444336, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7097023725509644, + "num_tokens": 128554987.0, + "step": 4965 + }, + { + "epoch": 0.5453547111794421, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.3856148719787598, + "learning_rate": 1e-06, + "loss": 1.0437, + "mean_token_accuracy": 0.6896292567253113, + "num_tokens": 128580111.0, + "step": 4966 + }, + { + "epoch": 0.5454645288820558, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 1.9990159273147583, + "learning_rate": 1e-06, + "loss": 1.0781, + "mean_token_accuracy": 0.6768321990966797, + "num_tokens": 128612332.0, + "step": 4967 + }, + { + "epoch": 0.5455743465846694, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.335655450820923, + "learning_rate": 1e-06, + "loss": 1.0256, + "mean_token_accuracy": 0.6894076466560364, + "num_tokens": 128636853.0, + "step": 4968 + }, + { + "epoch": 0.5456841642872831, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.0607857704162598, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7240082621574402, + "num_tokens": 128664436.0, + "step": 4969 + }, + { + "epoch": 0.5457939819898968, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.0718765258789062, + "learning_rate": 1e-06, + "loss": 1.0618, + "mean_token_accuracy": 0.6859296560287476, + "num_tokens": 128694240.0, + "step": 4970 + }, + { + "epoch": 0.5459037996925105, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.2890281677246094, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6961228251457214, + "num_tokens": 128719182.0, + "step": 4971 + }, + { + "epoch": 0.5460136173951241, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.080166816711426, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7009691596031189, + "num_tokens": 128746729.0, + "step": 4972 + }, + { + "epoch": 0.5461234350977378, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.275346517562866, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6939456462860107, + "num_tokens": 128772278.0, + "step": 4973 + }, + { + "epoch": 0.5462332528003514, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.20991587638855, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6910156011581421, + "num_tokens": 128799048.0, + "step": 4974 + }, + { + "epoch": 0.546343070502965, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.1203694343566895, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7026910185813904, + "num_tokens": 128825845.0, + "step": 4975 + }, + { + "epoch": 0.5464528882055787, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.196336507797241, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7068532705307007, + "num_tokens": 128854929.0, + "step": 4976 + }, + { + "epoch": 0.5465627059081924, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.3537909984588623, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7059116363525391, + "num_tokens": 128878255.0, + "step": 4977 + }, + { + "epoch": 0.5466725236108061, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 1.9784443378448486, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6910404562950134, + "num_tokens": 128910671.0, + "step": 4978 + }, + { + "epoch": 0.5467823413134197, + "ewc_loss": 4.57763671875e-05, + "grad_norm": 2.076648235321045, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6929792165756226, + "num_tokens": 128941001.0, + "step": 4979 + }, + { + "epoch": 0.5468921590160334, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.2235829830169678, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7195816040039062, + "num_tokens": 128965677.0, + "step": 4980 + }, + { + "epoch": 0.547001976718647, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.310800552368164, + "learning_rate": 1e-06, + "loss": 1.0726, + "mean_token_accuracy": 0.679271936416626, + "num_tokens": 128992259.0, + "step": 4981 + }, + { + "epoch": 0.5471117944212607, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 1.8660897016525269, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7142989039421082, + "num_tokens": 129026500.0, + "step": 4982 + }, + { + "epoch": 0.5472216121238743, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.139281988143921, + "learning_rate": 1e-06, + "loss": 1.0833, + "mean_token_accuracy": 0.6763970851898193, + "num_tokens": 129055142.0, + "step": 4983 + }, + { + "epoch": 0.5473314298264881, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 1.966435194015503, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7064255475997925, + "num_tokens": 129087650.0, + "step": 4984 + }, + { + "epoch": 0.5474412475291017, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 1.9383940696716309, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7008187770843506, + "num_tokens": 129117894.0, + "step": 4985 + }, + { + "epoch": 0.5475510652317154, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.1492409706115723, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.697957456111908, + "num_tokens": 129144524.0, + "step": 4986 + }, + { + "epoch": 0.547660882934329, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.4229423999786377, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6914054155349731, + "num_tokens": 129167161.0, + "step": 4987 + }, + { + "epoch": 0.5477707006369427, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.287416696548462, + "learning_rate": 1e-06, + "loss": 1.0921, + "mean_token_accuracy": 0.6775975823402405, + "num_tokens": 129192187.0, + "step": 4988 + }, + { + "epoch": 0.5478805183395563, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.080821990966797, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.7016531229019165, + "num_tokens": 129220666.0, + "step": 4989 + }, + { + "epoch": 0.54799033604217, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.345736026763916, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7092280387878418, + "num_tokens": 129242823.0, + "step": 4990 + }, + { + "epoch": 0.5481001537447837, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.307997465133667, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.68873530626297, + "num_tokens": 129267928.0, + "step": 4991 + }, + { + "epoch": 0.5482099714473974, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.3539986610412598, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6893075704574585, + "num_tokens": 129292838.0, + "step": 4992 + }, + { + "epoch": 0.548319789150011, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.3783321380615234, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7093901634216309, + "num_tokens": 129314953.0, + "step": 4993 + }, + { + "epoch": 0.5484296068526247, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.2150559425354004, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.719584584236145, + "num_tokens": 129339660.0, + "step": 4994 + }, + { + "epoch": 0.5485394245552383, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.3791866302490234, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7141515612602234, + "num_tokens": 129363111.0, + "step": 4995 + }, + { + "epoch": 0.5486492422578519, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.2319905757904053, + "learning_rate": 1e-06, + "loss": 1.0394, + "mean_token_accuracy": 0.6879896521568298, + "num_tokens": 129390318.0, + "step": 4996 + }, + { + "epoch": 0.5487590599604656, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.2396059036254883, + "learning_rate": 1e-06, + "loss": 1.041, + "mean_token_accuracy": 0.6939119696617126, + "num_tokens": 129416078.0, + "step": 4997 + }, + { + "epoch": 0.5488688776630792, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.626713991165161, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6893044710159302, + "num_tokens": 129437106.0, + "step": 4998 + }, + { + "epoch": 0.548978695365693, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.0067076683044434, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7044927477836609, + "num_tokens": 129469652.0, + "step": 4999 + }, + { + "epoch": 0.5490885130683066, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.0907516479492188, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.709870457649231, + "num_tokens": 129496588.0, + "step": 5000 + }, + { + "epoch": 0.5491983307709203, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.2427289485931396, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.70685213804245, + "num_tokens": 129521502.0, + "step": 5001 + }, + { + "epoch": 0.5493081484735339, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.3123390674591064, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7215061783790588, + "num_tokens": 129544748.0, + "step": 5002 + }, + { + "epoch": 0.5494179661761476, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.0364651679992676, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7017664909362793, + "num_tokens": 129574117.0, + "step": 5003 + }, + { + "epoch": 0.5495277838787612, + "ewc_loss": 4.601478576660156e-05, + "grad_norm": 2.7738373279571533, + "learning_rate": 1e-06, + "loss": 1.1078, + "mean_token_accuracy": 0.6862044930458069, + "num_tokens": 129595211.0, + "step": 5004 + }, + { + "epoch": 0.5496376015813749, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.3281304836273193, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6996197700500488, + "num_tokens": 129619967.0, + "step": 5005 + }, + { + "epoch": 0.5497474192839886, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.1748969554901123, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7124237418174744, + "num_tokens": 129645294.0, + "step": 5006 + }, + { + "epoch": 0.5498572369866023, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.5297093391418457, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7039002180099487, + "num_tokens": 129664580.0, + "step": 5007 + }, + { + "epoch": 0.5499670546892159, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.126323699951172, + "learning_rate": 1e-06, + "loss": 1.0083, + "mean_token_accuracy": 0.6941173672676086, + "num_tokens": 129691874.0, + "step": 5008 + }, + { + "epoch": 0.5500768723918296, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.1903083324432373, + "learning_rate": 1e-06, + "loss": 1.0608, + "mean_token_accuracy": 0.6849595308303833, + "num_tokens": 129719829.0, + "step": 5009 + }, + { + "epoch": 0.5501866900944432, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.2088100910186768, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7285019159317017, + "num_tokens": 129743967.0, + "step": 5010 + }, + { + "epoch": 0.5502965077970569, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.3373568058013916, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7168493866920471, + "num_tokens": 129767756.0, + "step": 5011 + }, + { + "epoch": 0.5504063254996705, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.313289165496826, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7080357074737549, + "num_tokens": 129790423.0, + "step": 5012 + }, + { + "epoch": 0.5505161432022843, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.285587787628174, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6931288242340088, + "num_tokens": 129815734.0, + "step": 5013 + }, + { + "epoch": 0.5506259609048979, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.0724215507507324, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7129672765731812, + "num_tokens": 129842341.0, + "step": 5014 + }, + { + "epoch": 0.5507357786075116, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 1.9021549224853516, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.701841413974762, + "num_tokens": 129873370.0, + "step": 5015 + }, + { + "epoch": 0.5508455963101252, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.319413661956787, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7023835182189941, + "num_tokens": 129897220.0, + "step": 5016 + }, + { + "epoch": 0.5509554140127388, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.189429759979248, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6998797059059143, + "num_tokens": 129922265.0, + "step": 5017 + }, + { + "epoch": 0.5510652317153525, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.3437039852142334, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7036128044128418, + "num_tokens": 129946103.0, + "step": 5018 + }, + { + "epoch": 0.5511750494179661, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.168292760848999, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6999648213386536, + "num_tokens": 129973655.0, + "step": 5019 + }, + { + "epoch": 0.5512848671205799, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.1979851722717285, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7267589569091797, + "num_tokens": 129998339.0, + "step": 5020 + }, + { + "epoch": 0.5513946848231935, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.01761531829834, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6893044710159302, + "num_tokens": 130031600.0, + "step": 5021 + }, + { + "epoch": 0.5515045025258072, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.2909631729125977, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6962093114852905, + "num_tokens": 130055600.0, + "step": 5022 + }, + { + "epoch": 0.5516143202284208, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.3287370204925537, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7190834283828735, + "num_tokens": 130077860.0, + "step": 5023 + }, + { + "epoch": 0.5517241379310345, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.283998489379883, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.7133194208145142, + "num_tokens": 130102686.0, + "step": 5024 + }, + { + "epoch": 0.5518339556336481, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 2.40138578414917, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6921321153640747, + "num_tokens": 130125007.0, + "step": 5025 + }, + { + "epoch": 0.5519437733362618, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 1.960577368736267, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.7022767066955566, + "num_tokens": 130154385.0, + "step": 5026 + }, + { + "epoch": 0.5520535910388754, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.241244077682495, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7178946733474731, + "num_tokens": 130179220.0, + "step": 5027 + }, + { + "epoch": 0.5521634087414892, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.2902629375457764, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7023926973342896, + "num_tokens": 130204245.0, + "step": 5028 + }, + { + "epoch": 0.5522732264441028, + "ewc_loss": 4.6253204345703125e-05, + "grad_norm": 1.9077767133712769, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7210406064987183, + "num_tokens": 130237468.0, + "step": 5029 + }, + { + "epoch": 0.5523830441467165, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.271578550338745, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6960195899009705, + "num_tokens": 130261337.0, + "step": 5030 + }, + { + "epoch": 0.5524928618493301, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.001293182373047, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6901613473892212, + "num_tokens": 130292001.0, + "step": 5031 + }, + { + "epoch": 0.5526026795519438, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.4131059646606445, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.6993131637573242, + "num_tokens": 130315471.0, + "step": 5032 + }, + { + "epoch": 0.5527124972545574, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.1532235145568848, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.697153627872467, + "num_tokens": 130343018.0, + "step": 5033 + }, + { + "epoch": 0.552822314957171, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.2351672649383545, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6924073696136475, + "num_tokens": 130368875.0, + "step": 5034 + }, + { + "epoch": 0.5529321326597848, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.048644781112671, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6984437704086304, + "num_tokens": 130396092.0, + "step": 5035 + }, + { + "epoch": 0.5530419503623984, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.3043689727783203, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6930155754089355, + "num_tokens": 130420468.0, + "step": 5036 + }, + { + "epoch": 0.5531517680650121, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.5277929306030273, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7059375047683716, + "num_tokens": 130440923.0, + "step": 5037 + }, + { + "epoch": 0.5532615857676257, + "ewc_loss": 4.649162292480469e-05, + "grad_norm": 2.147916793823242, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7301337122917175, + "num_tokens": 130465065.0, + "step": 5038 + }, + { + "epoch": 0.5533714034702394, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.4924237728118896, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6873812675476074, + "num_tokens": 130487004.0, + "step": 5039 + }, + { + "epoch": 0.553481221172853, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.302145004272461, + "learning_rate": 1e-06, + "loss": 1.0559, + "mean_token_accuracy": 0.6775067448616028, + "num_tokens": 130514627.0, + "step": 5040 + }, + { + "epoch": 0.5535910388754667, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.443206548690796, + "learning_rate": 1e-06, + "loss": 1.0274, + "mean_token_accuracy": 0.6888797879219055, + "num_tokens": 130536856.0, + "step": 5041 + }, + { + "epoch": 0.5537008565780804, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.13454008102417, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7069698572158813, + "num_tokens": 130562912.0, + "step": 5042 + }, + { + "epoch": 0.5538106742806941, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1179516315460205, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7059299349784851, + "num_tokens": 130589178.0, + "step": 5043 + }, + { + "epoch": 0.5539204919833077, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1239101886749268, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.702755331993103, + "num_tokens": 130617557.0, + "step": 5044 + }, + { + "epoch": 0.5540303096859214, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.2375874519348145, + "learning_rate": 1e-06, + "loss": 1.053, + "mean_token_accuracy": 0.682947039604187, + "num_tokens": 130645531.0, + "step": 5045 + }, + { + "epoch": 0.554140127388535, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.591632127761841, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7049307823181152, + "num_tokens": 130666322.0, + "step": 5046 + }, + { + "epoch": 0.5542499450911487, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.3541250228881836, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7155972123146057, + "num_tokens": 130689313.0, + "step": 5047 + }, + { + "epoch": 0.5543597627937623, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.0722222328186035, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7064963579177856, + "num_tokens": 130717803.0, + "step": 5048 + }, + { + "epoch": 0.5544695804963761, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 1.976961612701416, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6932076215744019, + "num_tokens": 130747161.0, + "step": 5049 + }, + { + "epoch": 0.5545793981989897, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.4173057079315186, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.6997419595718384, + "num_tokens": 130768925.0, + "step": 5050 + }, + { + "epoch": 0.5546892159016034, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.3488965034484863, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7051256895065308, + "num_tokens": 130792964.0, + "step": 5051 + }, + { + "epoch": 0.554799033604217, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.531545400619507, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7182326912879944, + "num_tokens": 130813028.0, + "step": 5052 + }, + { + "epoch": 0.5549088513068307, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.175428867340088, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7018091082572937, + "num_tokens": 130838883.0, + "step": 5053 + }, + { + "epoch": 0.5550186690094443, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1314520835876465, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6955809593200684, + "num_tokens": 130866937.0, + "step": 5054 + }, + { + "epoch": 0.555128486712058, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.2395949363708496, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6903165578842163, + "num_tokens": 130891268.0, + "step": 5055 + }, + { + "epoch": 0.5552383044146716, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1569743156433105, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6985721588134766, + "num_tokens": 130920607.0, + "step": 5056 + }, + { + "epoch": 0.5553481221172853, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1847951412200928, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.7040905356407166, + "num_tokens": 130947053.0, + "step": 5057 + }, + { + "epoch": 0.555457939819899, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.573164224624634, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.693890392780304, + "num_tokens": 130968392.0, + "step": 5058 + }, + { + "epoch": 0.5555677575225126, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.370809316635132, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7290706038475037, + "num_tokens": 130990351.0, + "step": 5059 + }, + { + "epoch": 0.5556775752251263, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.164435386657715, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7094786763191223, + "num_tokens": 131016161.0, + "step": 5060 + }, + { + "epoch": 0.5557873929277399, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.439321279525757, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7105101346969604, + "num_tokens": 131037117.0, + "step": 5061 + }, + { + "epoch": 0.5558972106303536, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.3440778255462646, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6977801322937012, + "num_tokens": 131060013.0, + "step": 5062 + }, + { + "epoch": 0.5560070283329672, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.3687829971313477, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.6951367855072021, + "num_tokens": 131084337.0, + "step": 5063 + }, + { + "epoch": 0.556116846035581, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.0462539196014404, + "learning_rate": 1e-06, + "loss": 1.1081, + "mean_token_accuracy": 0.6706446409225464, + "num_tokens": 131116657.0, + "step": 5064 + }, + { + "epoch": 0.5562266637381946, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.310840129852295, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7064145803451538, + "num_tokens": 131140951.0, + "step": 5065 + }, + { + "epoch": 0.5563364814408083, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.1782121658325195, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7015626430511475, + "num_tokens": 131166105.0, + "step": 5066 + }, + { + "epoch": 0.5564462991434219, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.0665009021759033, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.692913830280304, + "num_tokens": 131197081.0, + "step": 5067 + }, + { + "epoch": 0.5565561168460356, + "ewc_loss": 4.673004150390625e-05, + "grad_norm": 2.4492557048797607, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7271988391876221, + "num_tokens": 131218222.0, + "step": 5068 + }, + { + "epoch": 0.5566659345486492, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.2339982986450195, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.695227861404419, + "num_tokens": 131242354.0, + "step": 5069 + }, + { + "epoch": 0.5567757522512629, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.183807373046875, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7076190114021301, + "num_tokens": 131267390.0, + "step": 5070 + }, + { + "epoch": 0.5568855699538766, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.0771820545196533, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7042286396026611, + "num_tokens": 131295486.0, + "step": 5071 + }, + { + "epoch": 0.5569953876564903, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.369649887084961, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7060407996177673, + "num_tokens": 131317985.0, + "step": 5072 + }, + { + "epoch": 0.5571052053591039, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.6691412925720215, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7195157408714294, + "num_tokens": 131336640.0, + "step": 5073 + }, + { + "epoch": 0.5572150230617176, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.329684019088745, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6896829605102539, + "num_tokens": 131361817.0, + "step": 5074 + }, + { + "epoch": 0.5573248407643312, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.082010269165039, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7304806709289551, + "num_tokens": 131388140.0, + "step": 5075 + }, + { + "epoch": 0.5574346584669448, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1696248054504395, + "learning_rate": 1e-06, + "loss": 1.0683, + "mean_token_accuracy": 0.6841257214546204, + "num_tokens": 131414579.0, + "step": 5076 + }, + { + "epoch": 0.5575444761695585, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 1.9530020952224731, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7182782292366028, + "num_tokens": 131445760.0, + "step": 5077 + }, + { + "epoch": 0.5576542938721722, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1119894981384277, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6994941830635071, + "num_tokens": 131472746.0, + "step": 5078 + }, + { + "epoch": 0.5577641115747859, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.296945571899414, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7143025994300842, + "num_tokens": 131494842.0, + "step": 5079 + }, + { + "epoch": 0.5578739292773995, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.4258387088775635, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7013039588928223, + "num_tokens": 131517296.0, + "step": 5080 + }, + { + "epoch": 0.5579837469800132, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.046513080596924, + "learning_rate": 1e-06, + "loss": 1.0562, + "mean_token_accuracy": 0.680712103843689, + "num_tokens": 131546387.0, + "step": 5081 + }, + { + "epoch": 0.5580935646826268, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.6660940647125244, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7129721641540527, + "num_tokens": 131564509.0, + "step": 5082 + }, + { + "epoch": 0.5582033823852405, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.361215591430664, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6993919014930725, + "num_tokens": 131588213.0, + "step": 5083 + }, + { + "epoch": 0.5583132000878541, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1535425186157227, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6930060386657715, + "num_tokens": 131616361.0, + "step": 5084 + }, + { + "epoch": 0.5584230177904678, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.0434458255767822, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6900177597999573, + "num_tokens": 131646562.0, + "step": 5085 + }, + { + "epoch": 0.5585328354930815, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.223647356033325, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7104179263114929, + "num_tokens": 131670099.0, + "step": 5086 + }, + { + "epoch": 0.5586426531956952, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1006221771240234, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6919231414794922, + "num_tokens": 131697822.0, + "step": 5087 + }, + { + "epoch": 0.5587524708983088, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.261122465133667, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.712009608745575, + "num_tokens": 131721692.0, + "step": 5088 + }, + { + "epoch": 0.5588622886009225, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.484447956085205, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7044183015823364, + "num_tokens": 131743548.0, + "step": 5089 + }, + { + "epoch": 0.5589721063035361, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.245744466781616, + "learning_rate": 1e-06, + "loss": 1.1377, + "mean_token_accuracy": 0.6673961877822876, + "num_tokens": 131769790.0, + "step": 5090 + }, + { + "epoch": 0.5590819240061498, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.096963405609131, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6955296993255615, + "num_tokens": 131799241.0, + "step": 5091 + }, + { + "epoch": 0.5591917417087634, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 1.750690221786499, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6876307725906372, + "num_tokens": 131839024.0, + "step": 5092 + }, + { + "epoch": 0.5593015594113772, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1085727214813232, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6834912300109863, + "num_tokens": 131866780.0, + "step": 5093 + }, + { + "epoch": 0.5594113771139908, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.1405506134033203, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6970990896224976, + "num_tokens": 131894181.0, + "step": 5094 + }, + { + "epoch": 0.5595211948166045, + "ewc_loss": 4.7206878662109375e-05, + "grad_norm": 2.6979594230651855, + "learning_rate": 1e-06, + "loss": 1.009, + "mean_token_accuracy": 0.6944491863250732, + "num_tokens": 131914061.0, + "step": 5095 + }, + { + "epoch": 0.5596310125192181, + "ewc_loss": 4.696846008300781e-05, + "grad_norm": 2.0711886882781982, + "learning_rate": 1e-06, + "loss": 1.0949, + "mean_token_accuracy": 0.6757335662841797, + "num_tokens": 131944439.0, + "step": 5096 + }, + { + "epoch": 0.5597408302218317, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.173330545425415, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7234692573547363, + "num_tokens": 131969212.0, + "step": 5097 + }, + { + "epoch": 0.5598506479244454, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.196650505065918, + "learning_rate": 1e-06, + "loss": 1.1445, + "mean_token_accuracy": 0.6592423319816589, + "num_tokens": 131998108.0, + "step": 5098 + }, + { + "epoch": 0.559960465627059, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 1.919474720954895, + "learning_rate": 1e-06, + "loss": 1.0447, + "mean_token_accuracy": 0.6776663064956665, + "num_tokens": 132031655.0, + "step": 5099 + }, + { + "epoch": 0.5600702833296728, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.4543867111206055, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7153445482254028, + "num_tokens": 132051282.0, + "step": 5100 + }, + { + "epoch": 0.5601801010322864, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.140289545059204, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.6970322132110596, + "num_tokens": 132075903.0, + "step": 5101 + }, + { + "epoch": 0.5602899187349001, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1805074214935303, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6978164315223694, + "num_tokens": 132101575.0, + "step": 5102 + }, + { + "epoch": 0.5603997364375137, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2036893367767334, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.700695276260376, + "num_tokens": 132127711.0, + "step": 5103 + }, + { + "epoch": 0.5605095541401274, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.3009591102600098, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.71654212474823, + "num_tokens": 132150607.0, + "step": 5104 + }, + { + "epoch": 0.560619371842741, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2319910526275635, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.7006881833076477, + "num_tokens": 132174942.0, + "step": 5105 + }, + { + "epoch": 0.5607291895453547, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2813541889190674, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7194836139678955, + "num_tokens": 132197226.0, + "step": 5106 + }, + { + "epoch": 0.5608390072479684, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1323776245117188, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7173964977264404, + "num_tokens": 132224700.0, + "step": 5107 + }, + { + "epoch": 0.5609488249505821, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.267479658126831, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6901206970214844, + "num_tokens": 132249511.0, + "step": 5108 + }, + { + "epoch": 0.5610586426531957, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2219245433807373, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.686108410358429, + "num_tokens": 132276127.0, + "step": 5109 + }, + { + "epoch": 0.5611684603558094, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.011046886444092, + "learning_rate": 1e-06, + "loss": 1.0716, + "mean_token_accuracy": 0.6771970987319946, + "num_tokens": 132306672.0, + "step": 5110 + }, + { + "epoch": 0.561278278058423, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1517438888549805, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.7041347026824951, + "num_tokens": 132333910.0, + "step": 5111 + }, + { + "epoch": 0.5613880957610367, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.138427734375, + "learning_rate": 1e-06, + "loss": 1.071, + "mean_token_accuracy": 0.6857141852378845, + "num_tokens": 132363479.0, + "step": 5112 + }, + { + "epoch": 0.5614979134636503, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.131068468093872, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7051354646682739, + "num_tokens": 132390880.0, + "step": 5113 + }, + { + "epoch": 0.5616077311662641, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.223595380783081, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7041676044464111, + "num_tokens": 132416496.0, + "step": 5114 + }, + { + "epoch": 0.5617175488688777, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.231037139892578, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7101624011993408, + "num_tokens": 132441545.0, + "step": 5115 + }, + { + "epoch": 0.5618273665714913, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1259989738464355, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7199884653091431, + "num_tokens": 132466047.0, + "step": 5116 + }, + { + "epoch": 0.561937184274105, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.273263692855835, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7223415374755859, + "num_tokens": 132487881.0, + "step": 5117 + }, + { + "epoch": 0.5620470019767186, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1874654293060303, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7188565135002136, + "num_tokens": 132515071.0, + "step": 5118 + }, + { + "epoch": 0.5621568196793323, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.326748847961426, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6821324825286865, + "num_tokens": 132539234.0, + "step": 5119 + }, + { + "epoch": 0.5622666373819459, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.127441883087158, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6891570091247559, + "num_tokens": 132569368.0, + "step": 5120 + }, + { + "epoch": 0.5623764550845596, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.094879627227783, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7039176225662231, + "num_tokens": 132595696.0, + "step": 5121 + }, + { + "epoch": 0.5624862727871733, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 1.9750783443450928, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6920828819274902, + "num_tokens": 132627634.0, + "step": 5122 + }, + { + "epoch": 0.562596090489787, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.227060556411743, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6989203691482544, + "num_tokens": 132652802.0, + "step": 5123 + }, + { + "epoch": 0.5627059081924006, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2824771404266357, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7271190881729126, + "num_tokens": 132677330.0, + "step": 5124 + }, + { + "epoch": 0.5628157258950143, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.112023115158081, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7043726444244385, + "num_tokens": 132703704.0, + "step": 5125 + }, + { + "epoch": 0.5629255435976279, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.422790050506592, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7130605578422546, + "num_tokens": 132731686.0, + "step": 5126 + }, + { + "epoch": 0.5630353613002416, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.0109994411468506, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7124490737915039, + "num_tokens": 132761771.0, + "step": 5127 + }, + { + "epoch": 0.5631451790028552, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.2311577796936035, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6930130124092102, + "num_tokens": 132786439.0, + "step": 5128 + }, + { + "epoch": 0.563254996705469, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.163247585296631, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.726015567779541, + "num_tokens": 132811005.0, + "step": 5129 + }, + { + "epoch": 0.5633648144080826, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.1725151538848877, + "learning_rate": 1e-06, + "loss": 1.1107, + "mean_token_accuracy": 0.671497106552124, + "num_tokens": 132839399.0, + "step": 5130 + }, + { + "epoch": 0.5634746321106963, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.0291919708251953, + "learning_rate": 1e-06, + "loss": 1.0359, + "mean_token_accuracy": 0.6896036267280579, + "num_tokens": 132867618.0, + "step": 5131 + }, + { + "epoch": 0.5635844498133099, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 1.9189025163650513, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6883595585823059, + "num_tokens": 132901567.0, + "step": 5132 + }, + { + "epoch": 0.5636942675159236, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.303814172744751, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7161145210266113, + "num_tokens": 132925044.0, + "step": 5133 + }, + { + "epoch": 0.5638040852185372, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.5343165397644043, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7207315564155579, + "num_tokens": 132944209.0, + "step": 5134 + }, + { + "epoch": 0.5639139029211508, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.0305397510528564, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.7019768357276917, + "num_tokens": 132973354.0, + "step": 5135 + }, + { + "epoch": 0.5640237206237646, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.494539976119995, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7329148054122925, + "num_tokens": 132992898.0, + "step": 5136 + }, + { + "epoch": 0.5641335383263782, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.541881561279297, + "learning_rate": 1e-06, + "loss": 1.0121, + "mean_token_accuracy": 0.6997470259666443, + "num_tokens": 133013935.0, + "step": 5137 + }, + { + "epoch": 0.5642433560289919, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.7427189350128174, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7259020209312439, + "num_tokens": 133033719.0, + "step": 5138 + }, + { + "epoch": 0.5643531737316055, + "ewc_loss": 4.744529724121094e-05, + "grad_norm": 2.0956156253814697, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6975412964820862, + "num_tokens": 133062200.0, + "step": 5139 + }, + { + "epoch": 0.5644629914342192, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.094799518585205, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.6977138519287109, + "num_tokens": 133090360.0, + "step": 5140 + }, + { + "epoch": 0.5645728091368328, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.192803144454956, + "learning_rate": 1e-06, + "loss": 1.0463, + "mean_token_accuracy": 0.6930897235870361, + "num_tokens": 133117850.0, + "step": 5141 + }, + { + "epoch": 0.5646826268394465, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.2540996074676514, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7152428030967712, + "num_tokens": 133142324.0, + "step": 5142 + }, + { + "epoch": 0.5647924445420602, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.4320356845855713, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7426260709762573, + "num_tokens": 133163820.0, + "step": 5143 + }, + { + "epoch": 0.5649022622446739, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.1400043964385986, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7130106687545776, + "num_tokens": 133190171.0, + "step": 5144 + }, + { + "epoch": 0.5650120799472875, + "ewc_loss": 4.76837158203125e-05, + "grad_norm": 2.40718936920166, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7152606248855591, + "num_tokens": 133211488.0, + "step": 5145 + }, + { + "epoch": 0.5651218976499012, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.475890636444092, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7057311534881592, + "num_tokens": 133232643.0, + "step": 5146 + }, + { + "epoch": 0.5652317153525148, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.67910099029541, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7115800380706787, + "num_tokens": 133250041.0, + "step": 5147 + }, + { + "epoch": 0.5653415330551285, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.292391300201416, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.692388117313385, + "num_tokens": 133274638.0, + "step": 5148 + }, + { + "epoch": 0.5654513507577421, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.16123104095459, + "learning_rate": 1e-06, + "loss": 1.0545, + "mean_token_accuracy": 0.691293478012085, + "num_tokens": 133307670.0, + "step": 5149 + }, + { + "epoch": 0.5655611684603558, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2143263816833496, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7043971419334412, + "num_tokens": 133332989.0, + "step": 5150 + }, + { + "epoch": 0.5656709861629695, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.077040433883667, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7080602645874023, + "num_tokens": 133360939.0, + "step": 5151 + }, + { + "epoch": 0.5657808038655832, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1018662452697754, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7120633125305176, + "num_tokens": 133388730.0, + "step": 5152 + }, + { + "epoch": 0.5658906215681968, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.0790653228759766, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.700877845287323, + "num_tokens": 133416971.0, + "step": 5153 + }, + { + "epoch": 0.5660004392708105, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.141065835952759, + "learning_rate": 1e-06, + "loss": 1.0437, + "mean_token_accuracy": 0.6883112192153931, + "num_tokens": 133444888.0, + "step": 5154 + }, + { + "epoch": 0.5661102569734241, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.177771806716919, + "learning_rate": 1e-06, + "loss": 1.0752, + "mean_token_accuracy": 0.6835225224494934, + "num_tokens": 133471556.0, + "step": 5155 + }, + { + "epoch": 0.5662200746760377, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9909071922302246, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6948909759521484, + "num_tokens": 133504244.0, + "step": 5156 + }, + { + "epoch": 0.5663298923786514, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9982868432998657, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6855334043502808, + "num_tokens": 133533015.0, + "step": 5157 + }, + { + "epoch": 0.5664397100812651, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.5363826751708984, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6940849423408508, + "num_tokens": 133553724.0, + "step": 5158 + }, + { + "epoch": 0.5665495277838788, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9178669452667236, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6866286993026733, + "num_tokens": 133586088.0, + "step": 5159 + }, + { + "epoch": 0.5666593454864924, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2628870010375977, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.6983727216720581, + "num_tokens": 133610842.0, + "step": 5160 + }, + { + "epoch": 0.5667691631891061, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3900256156921387, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6999978423118591, + "num_tokens": 133632718.0, + "step": 5161 + }, + { + "epoch": 0.5668789808917197, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1516849994659424, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6946347951889038, + "num_tokens": 133657697.0, + "step": 5162 + }, + { + "epoch": 0.5669887985943334, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2296438217163086, + "learning_rate": 1e-06, + "loss": 0.9908, + "mean_token_accuracy": 0.7025574445724487, + "num_tokens": 133680641.0, + "step": 5163 + }, + { + "epoch": 0.567098616296947, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.042482376098633, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7229815721511841, + "num_tokens": 133707898.0, + "step": 5164 + }, + { + "epoch": 0.5672084339995608, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.4115614891052246, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.7045373916625977, + "num_tokens": 133729956.0, + "step": 5165 + }, + { + "epoch": 0.5673182517021744, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.0339391231536865, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7324050664901733, + "num_tokens": 133756836.0, + "step": 5166 + }, + { + "epoch": 0.5674280694047881, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3599727153778076, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7183672189712524, + "num_tokens": 133778886.0, + "step": 5167 + }, + { + "epoch": 0.5675378871074017, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.131100654602051, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.6837121248245239, + "num_tokens": 133809059.0, + "step": 5168 + }, + { + "epoch": 0.5676477048100154, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.92362642288208, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7050027847290039, + "num_tokens": 133839123.0, + "step": 5169 + }, + { + "epoch": 0.567757522512629, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9882689714431763, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7349931597709656, + "num_tokens": 133866533.0, + "step": 5170 + }, + { + "epoch": 0.5678673402152427, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.0786924362182617, + "learning_rate": 1e-06, + "loss": 1.0406, + "mean_token_accuracy": 0.6889039278030396, + "num_tokens": 133897919.0, + "step": 5171 + }, + { + "epoch": 0.5679771579178564, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9788085222244263, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6966785192489624, + "num_tokens": 133928244.0, + "step": 5172 + }, + { + "epoch": 0.5680869756204701, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.120715379714966, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6921103000640869, + "num_tokens": 133956029.0, + "step": 5173 + }, + { + "epoch": 0.5681967933230837, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.074862241744995, + "learning_rate": 1e-06, + "loss": 1.0454, + "mean_token_accuracy": 0.6831769943237305, + "num_tokens": 133986082.0, + "step": 5174 + }, + { + "epoch": 0.5683066110256974, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2128379344940186, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6977142095565796, + "num_tokens": 134011899.0, + "step": 5175 + }, + { + "epoch": 0.568416428728311, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.4344167709350586, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7066258192062378, + "num_tokens": 134032637.0, + "step": 5176 + }, + { + "epoch": 0.5685262464309246, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3024744987487793, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.7051832675933838, + "num_tokens": 134056987.0, + "step": 5177 + }, + { + "epoch": 0.5686360641335383, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.4307568073272705, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6894846558570862, + "num_tokens": 134081097.0, + "step": 5178 + }, + { + "epoch": 0.5687458818361519, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3628361225128174, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7089968919754028, + "num_tokens": 134104706.0, + "step": 5179 + }, + { + "epoch": 0.5688556995387657, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1740283966064453, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6857221126556396, + "num_tokens": 134131050.0, + "step": 5180 + }, + { + "epoch": 0.5689655172413793, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2508273124694824, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6932462453842163, + "num_tokens": 134157262.0, + "step": 5181 + }, + { + "epoch": 0.569075334943993, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.0363667011260986, + "learning_rate": 1e-06, + "loss": 1.1354, + "mean_token_accuracy": 0.6681901812553406, + "num_tokens": 134188743.0, + "step": 5182 + }, + { + "epoch": 0.5691851526466066, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1360676288604736, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7038466334342957, + "num_tokens": 134215418.0, + "step": 5183 + }, + { + "epoch": 0.5692949703492203, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.827677845954895, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7082276344299316, + "num_tokens": 134248000.0, + "step": 5184 + }, + { + "epoch": 0.5694047880518339, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3824803829193115, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7113446593284607, + "num_tokens": 134271086.0, + "step": 5185 + }, + { + "epoch": 0.5695146057544476, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.402580499649048, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7159532308578491, + "num_tokens": 134293514.0, + "step": 5186 + }, + { + "epoch": 0.5696244234570613, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.097752094268799, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6853072643280029, + "num_tokens": 134324497.0, + "step": 5187 + }, + { + "epoch": 0.569734241159675, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2753102779388428, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7082582712173462, + "num_tokens": 134348726.0, + "step": 5188 + }, + { + "epoch": 0.5698440588622886, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.13124680519104, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.715985894203186, + "num_tokens": 134375115.0, + "step": 5189 + }, + { + "epoch": 0.5699538765649023, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.4028515815734863, + "learning_rate": 1e-06, + "loss": 1.0713, + "mean_token_accuracy": 0.6835784912109375, + "num_tokens": 134397922.0, + "step": 5190 + }, + { + "epoch": 0.5700636942675159, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1462724208831787, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7213519811630249, + "num_tokens": 134425011.0, + "step": 5191 + }, + { + "epoch": 0.5701735119701296, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.302776336669922, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6944706439971924, + "num_tokens": 134452007.0, + "step": 5192 + }, + { + "epoch": 0.5702833296727432, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.087571144104004, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.7013728618621826, + "num_tokens": 134480460.0, + "step": 5193 + }, + { + "epoch": 0.570393147375357, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.230717897415161, + "learning_rate": 1e-06, + "loss": 1.0903, + "mean_token_accuracy": 0.6765998601913452, + "num_tokens": 134507478.0, + "step": 5194 + }, + { + "epoch": 0.5705029650779706, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.311418294906616, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6886111497879028, + "num_tokens": 134531741.0, + "step": 5195 + }, + { + "epoch": 0.5706127827805842, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1079390048980713, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7094672918319702, + "num_tokens": 134557636.0, + "step": 5196 + }, + { + "epoch": 0.5707226004831979, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3236896991729736, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7030643224716187, + "num_tokens": 134579402.0, + "step": 5197 + }, + { + "epoch": 0.5708324181858115, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1283657550811768, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7029099464416504, + "num_tokens": 134606547.0, + "step": 5198 + }, + { + "epoch": 0.5709422358884252, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3251187801361084, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7199108600616455, + "num_tokens": 134630304.0, + "step": 5199 + }, + { + "epoch": 0.5710520535910388, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3095316886901855, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7015745639801025, + "num_tokens": 134654483.0, + "step": 5200 + }, + { + "epoch": 0.5711618712936526, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9902430772781372, + "learning_rate": 1e-06, + "loss": 1.0682, + "mean_token_accuracy": 0.6804168224334717, + "num_tokens": 134687064.0, + "step": 5201 + }, + { + "epoch": 0.5712716889962662, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.225548267364502, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7074905633926392, + "num_tokens": 134713343.0, + "step": 5202 + }, + { + "epoch": 0.5713815066988799, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.413907289505005, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6888061761856079, + "num_tokens": 134736151.0, + "step": 5203 + }, + { + "epoch": 0.5714913244014935, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.304056406021118, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7187316417694092, + "num_tokens": 134760650.0, + "step": 5204 + }, + { + "epoch": 0.5716011421041072, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.5426411628723145, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7102874517440796, + "num_tokens": 134780341.0, + "step": 5205 + }, + { + "epoch": 0.5717109598067208, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1972341537475586, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7002639770507812, + "num_tokens": 134804164.0, + "step": 5206 + }, + { + "epoch": 0.5718207775093345, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.154815673828125, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7015738487243652, + "num_tokens": 134830883.0, + "step": 5207 + }, + { + "epoch": 0.5719305952119481, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.2236547470092773, + "learning_rate": 1e-06, + "loss": 1.059, + "mean_token_accuracy": 0.6842418909072876, + "num_tokens": 134857646.0, + "step": 5208 + }, + { + "epoch": 0.5720404129145619, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.0640156269073486, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7070615887641907, + "num_tokens": 134886570.0, + "step": 5209 + }, + { + "epoch": 0.5721502306171755, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.3307130336761475, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7189586758613586, + "num_tokens": 134909165.0, + "step": 5210 + }, + { + "epoch": 0.5722600483197892, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1685631275177, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.7033206224441528, + "num_tokens": 134935723.0, + "step": 5211 + }, + { + "epoch": 0.5723698660224028, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.657749891281128, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7117548584938049, + "num_tokens": 134952760.0, + "step": 5212 + }, + { + "epoch": 0.5724796837250165, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.153892993927002, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7038607597351074, + "num_tokens": 134978712.0, + "step": 5213 + }, + { + "epoch": 0.5725895014276301, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.119431734085083, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6838171482086182, + "num_tokens": 135007948.0, + "step": 5214 + }, + { + "epoch": 0.5726993191302437, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.945985198020935, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6881365776062012, + "num_tokens": 135039118.0, + "step": 5215 + }, + { + "epoch": 0.5728091368328575, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.866076946258545, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7043893933296204, + "num_tokens": 135076922.0, + "step": 5216 + }, + { + "epoch": 0.5729189545354711, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.034456729888916, + "learning_rate": 1e-06, + "loss": 1.0649, + "mean_token_accuracy": 0.6856540441513062, + "num_tokens": 135106676.0, + "step": 5217 + }, + { + "epoch": 0.5730287722380848, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.345043659210205, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7161391973495483, + "num_tokens": 135127502.0, + "step": 5218 + }, + { + "epoch": 0.5731385899406984, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.201909065246582, + "learning_rate": 1e-06, + "loss": 1.0623, + "mean_token_accuracy": 0.682465672492981, + "num_tokens": 135152895.0, + "step": 5219 + }, + { + "epoch": 0.5732484076433121, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9252400398254395, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7075172662734985, + "num_tokens": 135185791.0, + "step": 5220 + }, + { + "epoch": 0.5733582253459257, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 1.9598910808563232, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6898202896118164, + "num_tokens": 135216237.0, + "step": 5221 + }, + { + "epoch": 0.5734680430485394, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1689438819885254, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.6811989545822144, + "num_tokens": 135241407.0, + "step": 5222 + }, + { + "epoch": 0.5735778607511531, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.460916757583618, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7210162281990051, + "num_tokens": 135262741.0, + "step": 5223 + }, + { + "epoch": 0.5736876784537668, + "ewc_loss": 4.792213439941406e-05, + "grad_norm": 2.1614580154418945, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7315480709075928, + "num_tokens": 135285964.0, + "step": 5224 + }, + { + "epoch": 0.5737974961563804, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0943551063537598, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6916268467903137, + "num_tokens": 135314741.0, + "step": 5225 + }, + { + "epoch": 0.5739073138589941, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.3884646892547607, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.7013651132583618, + "num_tokens": 135336909.0, + "step": 5226 + }, + { + "epoch": 0.5740171315616077, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0012359619140625, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7258831858634949, + "num_tokens": 135364698.0, + "step": 5227 + }, + { + "epoch": 0.5741269492642214, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2687599658966064, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7027856111526489, + "num_tokens": 135389786.0, + "step": 5228 + }, + { + "epoch": 0.574236766966835, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.275089979171753, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7176298499107361, + "num_tokens": 135413067.0, + "step": 5229 + }, + { + "epoch": 0.5743465846694488, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.444899559020996, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7053478956222534, + "num_tokens": 135436186.0, + "step": 5230 + }, + { + "epoch": 0.5744564023720624, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.360801935195923, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7311288714408875, + "num_tokens": 135457201.0, + "step": 5231 + }, + { + "epoch": 0.5745662200746761, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.271571397781372, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7110335826873779, + "num_tokens": 135481982.0, + "step": 5232 + }, + { + "epoch": 0.5746760377772897, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.3900327682495117, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.693263053894043, + "num_tokens": 135505097.0, + "step": 5233 + }, + { + "epoch": 0.5747858554799034, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0967857837677, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.6965236663818359, + "num_tokens": 135533256.0, + "step": 5234 + }, + { + "epoch": 0.574895673182517, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.053973913192749, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7209076881408691, + "num_tokens": 135559313.0, + "step": 5235 + }, + { + "epoch": 0.5750054908851306, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.241734027862549, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7099200487136841, + "num_tokens": 135583174.0, + "step": 5236 + }, + { + "epoch": 0.5751153085877443, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0867936611175537, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7035015821456909, + "num_tokens": 135610717.0, + "step": 5237 + }, + { + "epoch": 0.575225126290358, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2715561389923096, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6974868774414062, + "num_tokens": 135635496.0, + "step": 5238 + }, + { + "epoch": 0.5753349439929717, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0295908451080322, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7205855846405029, + "num_tokens": 135662797.0, + "step": 5239 + }, + { + "epoch": 0.5754447616955853, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.6473052501678467, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7393285036087036, + "num_tokens": 135681463.0, + "step": 5240 + }, + { + "epoch": 0.575554579398199, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.002351760864258, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6958203315734863, + "num_tokens": 135712559.0, + "step": 5241 + }, + { + "epoch": 0.5756643971008126, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.300401210784912, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7132014036178589, + "num_tokens": 135735649.0, + "step": 5242 + }, + { + "epoch": 0.5757742148034263, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2383034229278564, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6967160701751709, + "num_tokens": 135763058.0, + "step": 5243 + }, + { + "epoch": 0.5758840325060399, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.3207006454467773, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7158807516098022, + "num_tokens": 135785873.0, + "step": 5244 + }, + { + "epoch": 0.5759938502086537, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2928707599639893, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7192379832267761, + "num_tokens": 135809402.0, + "step": 5245 + }, + { + "epoch": 0.5761036679112673, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.4442577362060547, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6884632706642151, + "num_tokens": 135832842.0, + "step": 5246 + }, + { + "epoch": 0.576213485613881, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2365033626556396, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6967830061912537, + "num_tokens": 135859345.0, + "step": 5247 + }, + { + "epoch": 0.5763233033164946, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.2052536010742188, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.705449104309082, + "num_tokens": 135885536.0, + "step": 5248 + }, + { + "epoch": 0.5764331210191083, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.215466022491455, + "learning_rate": 1e-06, + "loss": 1.0517, + "mean_token_accuracy": 0.6814209818840027, + "num_tokens": 135912140.0, + "step": 5249 + }, + { + "epoch": 0.5765429387217219, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.1006243228912354, + "learning_rate": 1e-06, + "loss": 1.1023, + "mean_token_accuracy": 0.6790887117385864, + "num_tokens": 135942276.0, + "step": 5250 + }, + { + "epoch": 0.5766527564243356, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.057316541671753, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7067378759384155, + "num_tokens": 135969380.0, + "step": 5251 + }, + { + "epoch": 0.5767625741269493, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 1.9769136905670166, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7253157496452332, + "num_tokens": 135997935.0, + "step": 5252 + }, + { + "epoch": 0.576872391829563, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.192638397216797, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6955305337905884, + "num_tokens": 136022158.0, + "step": 5253 + }, + { + "epoch": 0.5769822095321766, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.1110916137695312, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7086113691329956, + "num_tokens": 136048515.0, + "step": 5254 + }, + { + "epoch": 0.5770920272347903, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.3545947074890137, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7086266279220581, + "num_tokens": 136072347.0, + "step": 5255 + }, + { + "epoch": 0.5772018449374039, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.3114523887634277, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6820881962776184, + "num_tokens": 136098672.0, + "step": 5256 + }, + { + "epoch": 0.5773116626400175, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.241605043411255, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6984877586364746, + "num_tokens": 136123095.0, + "step": 5257 + }, + { + "epoch": 0.5774214803426312, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.1752076148986816, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.6827887296676636, + "num_tokens": 136150673.0, + "step": 5258 + }, + { + "epoch": 0.577531298045245, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 1.9011902809143066, + "learning_rate": 1e-06, + "loss": 1.0559, + "mean_token_accuracy": 0.6880571842193604, + "num_tokens": 136185183.0, + "step": 5259 + }, + { + "epoch": 0.5776411157478586, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 1.8676656484603882, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7099477648735046, + "num_tokens": 136217660.0, + "step": 5260 + }, + { + "epoch": 0.5777509334504722, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.33332896232605, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7129453420639038, + "num_tokens": 136240480.0, + "step": 5261 + }, + { + "epoch": 0.5778607511530859, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.3747189044952393, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6881338953971863, + "num_tokens": 136265258.0, + "step": 5262 + }, + { + "epoch": 0.5779705688556995, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.195735454559326, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7044186592102051, + "num_tokens": 136291840.0, + "step": 5263 + }, + { + "epoch": 0.5780803865583132, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.100529909133911, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.71241694688797, + "num_tokens": 136320057.0, + "step": 5264 + }, + { + "epoch": 0.5781902042609268, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.0405330657958984, + "learning_rate": 1e-06, + "loss": 1.0405, + "mean_token_accuracy": 0.6825841069221497, + "num_tokens": 136350046.0, + "step": 5265 + }, + { + "epoch": 0.5783000219635406, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.135695457458496, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.7117376923561096, + "num_tokens": 136378683.0, + "step": 5266 + }, + { + "epoch": 0.5784098396661542, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.0586819648742676, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.681780219078064, + "num_tokens": 136408297.0, + "step": 5267 + }, + { + "epoch": 0.5785196573687679, + "ewc_loss": 4.8160552978515625e-05, + "grad_norm": 2.0308940410614014, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6793519258499146, + "num_tokens": 136438358.0, + "step": 5268 + }, + { + "epoch": 0.5786294750713815, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.140355348587036, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.7043857574462891, + "num_tokens": 136464614.0, + "step": 5269 + }, + { + "epoch": 0.5787392927739952, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.2097702026367188, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7032384872436523, + "num_tokens": 136487921.0, + "step": 5270 + }, + { + "epoch": 0.5788491104766088, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.2691755294799805, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7005807161331177, + "num_tokens": 136512898.0, + "step": 5271 + }, + { + "epoch": 0.5789589281792225, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.1686882972717285, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.726481556892395, + "num_tokens": 136539048.0, + "step": 5272 + }, + { + "epoch": 0.5790687458818361, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.145437002182007, + "learning_rate": 1e-06, + "loss": 1.0702, + "mean_token_accuracy": 0.6760463714599609, + "num_tokens": 136567431.0, + "step": 5273 + }, + { + "epoch": 0.5791785635844499, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.174406051635742, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7138481140136719, + "num_tokens": 136593366.0, + "step": 5274 + }, + { + "epoch": 0.5792883812870635, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.492441177368164, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6916657090187073, + "num_tokens": 136614886.0, + "step": 5275 + }, + { + "epoch": 0.5793981989896771, + "ewc_loss": 4.839897155761719e-05, + "grad_norm": 2.094041585922241, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.6867113709449768, + "num_tokens": 136644948.0, + "step": 5276 + }, + { + "epoch": 0.5795080166922908, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.462005615234375, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.7045392394065857, + "num_tokens": 136667190.0, + "step": 5277 + }, + { + "epoch": 0.5796178343949044, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.186481237411499, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7194595336914062, + "num_tokens": 136691470.0, + "step": 5278 + }, + { + "epoch": 0.5797276520975181, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.0272297859191895, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.700748860836029, + "num_tokens": 136720589.0, + "step": 5279 + }, + { + "epoch": 0.5798374698001317, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 1.9426628351211548, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6866821050643921, + "num_tokens": 136751274.0, + "step": 5280 + }, + { + "epoch": 0.5799472875027455, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.6339426040649414, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.6955968141555786, + "num_tokens": 136772880.0, + "step": 5281 + }, + { + "epoch": 0.5800571052053591, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.118591070175171, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6987779140472412, + "num_tokens": 136800349.0, + "step": 5282 + }, + { + "epoch": 0.5801669229079728, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.5888521671295166, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7229123115539551, + "num_tokens": 136819943.0, + "step": 5283 + }, + { + "epoch": 0.5802767406105864, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.02620005607605, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.7018282413482666, + "num_tokens": 136849882.0, + "step": 5284 + }, + { + "epoch": 0.5803865583132001, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.0662026405334473, + "learning_rate": 1e-06, + "loss": 1.0685, + "mean_token_accuracy": 0.6748858094215393, + "num_tokens": 136878293.0, + "step": 5285 + }, + { + "epoch": 0.5804963760158137, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1508848667144775, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6946674585342407, + "num_tokens": 136906377.0, + "step": 5286 + }, + { + "epoch": 0.5806061937184274, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.063185214996338, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.6925761699676514, + "num_tokens": 136937276.0, + "step": 5287 + }, + { + "epoch": 0.5807160114210411, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.178304433822632, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.702502429485321, + "num_tokens": 136962371.0, + "step": 5288 + }, + { + "epoch": 0.5808258291236548, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.0279431343078613, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7170517444610596, + "num_tokens": 136989852.0, + "step": 5289 + }, + { + "epoch": 0.5809356468262684, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.562749147415161, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.71034836769104, + "num_tokens": 137008127.0, + "step": 5290 + }, + { + "epoch": 0.5810454645288821, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1684722900390625, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7436285614967346, + "num_tokens": 137031663.0, + "step": 5291 + }, + { + "epoch": 0.5811552822314957, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1599724292755127, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6955764293670654, + "num_tokens": 137059073.0, + "step": 5292 + }, + { + "epoch": 0.5812650999341094, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.39143967628479, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7295323014259338, + "num_tokens": 137079263.0, + "step": 5293 + }, + { + "epoch": 0.581374917636723, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.011470079421997, + "learning_rate": 1e-06, + "loss": 1.0628, + "mean_token_accuracy": 0.6827753782272339, + "num_tokens": 137111579.0, + "step": 5294 + }, + { + "epoch": 0.5814847353393368, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.110560894012451, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7207561731338501, + "num_tokens": 137136444.0, + "step": 5295 + }, + { + "epoch": 0.5815945530419504, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.148019790649414, + "learning_rate": 1e-06, + "loss": 1.1313, + "mean_token_accuracy": 0.6600973606109619, + "num_tokens": 137165499.0, + "step": 5296 + }, + { + "epoch": 0.581704370744564, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.017707347869873, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6955879926681519, + "num_tokens": 137194506.0, + "step": 5297 + }, + { + "epoch": 0.5818141884471777, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1352407932281494, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7019190192222595, + "num_tokens": 137222222.0, + "step": 5298 + }, + { + "epoch": 0.5819240061497913, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.223320245742798, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6893661022186279, + "num_tokens": 137248256.0, + "step": 5299 + }, + { + "epoch": 0.582033823852405, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.145688772201538, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7356923222541809, + "num_tokens": 137272022.0, + "step": 5300 + }, + { + "epoch": 0.5821436415550186, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.336475372314453, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.702607274055481, + "num_tokens": 137294545.0, + "step": 5301 + }, + { + "epoch": 0.5822534592576323, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.487205982208252, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7106276750564575, + "num_tokens": 137314838.0, + "step": 5302 + }, + { + "epoch": 0.582363276960246, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.118328332901001, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6987422704696655, + "num_tokens": 137340879.0, + "step": 5303 + }, + { + "epoch": 0.5824730946628597, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1759307384490967, + "learning_rate": 1e-06, + "loss": 1.077, + "mean_token_accuracy": 0.6930184364318848, + "num_tokens": 137368376.0, + "step": 5304 + }, + { + "epoch": 0.5825829123654733, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1066055297851562, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6972324848175049, + "num_tokens": 137396813.0, + "step": 5305 + }, + { + "epoch": 0.582692730068087, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.226699113845825, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7039795517921448, + "num_tokens": 137421814.0, + "step": 5306 + }, + { + "epoch": 0.5828025477707006, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1007003784179688, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7053961157798767, + "num_tokens": 137448987.0, + "step": 5307 + }, + { + "epoch": 0.5829123654733143, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 1.927878499031067, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7030797600746155, + "num_tokens": 137478806.0, + "step": 5308 + }, + { + "epoch": 0.5830221831759279, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.211512804031372, + "learning_rate": 1e-06, + "loss": 1.0293, + "mean_token_accuracy": 0.6843690872192383, + "num_tokens": 137504325.0, + "step": 5309 + }, + { + "epoch": 0.5831320008785417, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.4943618774414062, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7043315172195435, + "num_tokens": 137525204.0, + "step": 5310 + }, + { + "epoch": 0.5832418185811553, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.188471555709839, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7087147831916809, + "num_tokens": 137551618.0, + "step": 5311 + }, + { + "epoch": 0.583351636283769, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.205693483352661, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7062469124794006, + "num_tokens": 137577095.0, + "step": 5312 + }, + { + "epoch": 0.5834614539863826, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 1.9992190599441528, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7163500189781189, + "num_tokens": 137606454.0, + "step": 5313 + }, + { + "epoch": 0.5835712716889963, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.0796444416046143, + "learning_rate": 1e-06, + "loss": 1.0798, + "mean_token_accuracy": 0.6719244718551636, + "num_tokens": 137634449.0, + "step": 5314 + }, + { + "epoch": 0.5836810893916099, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.1847305297851562, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.7029779553413391, + "num_tokens": 137658097.0, + "step": 5315 + }, + { + "epoch": 0.5837909070942235, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.320673704147339, + "learning_rate": 1e-06, + "loss": 1.0315, + "mean_token_accuracy": 0.6873934268951416, + "num_tokens": 137683043.0, + "step": 5316 + }, + { + "epoch": 0.5839007247968373, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.4396450519561768, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.719561755657196, + "num_tokens": 137703335.0, + "step": 5317 + }, + { + "epoch": 0.584010542499451, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.5221478939056396, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7084130048751831, + "num_tokens": 137722898.0, + "step": 5318 + }, + { + "epoch": 0.5841203602020646, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 1.8747825622558594, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7075712084770203, + "num_tokens": 137755357.0, + "step": 5319 + }, + { + "epoch": 0.5842301779046782, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.0312042236328125, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7022406458854675, + "num_tokens": 137785996.0, + "step": 5320 + }, + { + "epoch": 0.5843399956072919, + "ewc_loss": 4.863739013671875e-05, + "grad_norm": 2.2081339359283447, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.7035884857177734, + "num_tokens": 137812226.0, + "step": 5321 + }, + { + "epoch": 0.5844498133099055, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 1.930915355682373, + "learning_rate": 1e-06, + "loss": 1.1331, + "mean_token_accuracy": 0.6641705632209778, + "num_tokens": 137845770.0, + "step": 5322 + }, + { + "epoch": 0.5845596310125192, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.101191520690918, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6969964504241943, + "num_tokens": 137874955.0, + "step": 5323 + }, + { + "epoch": 0.5846694487151329, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 1.8995245695114136, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7099513411521912, + "num_tokens": 137908104.0, + "step": 5324 + }, + { + "epoch": 0.5847792664177466, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.26698637008667, + "learning_rate": 1e-06, + "loss": 1.0664, + "mean_token_accuracy": 0.6829545497894287, + "num_tokens": 137933143.0, + "step": 5325 + }, + { + "epoch": 0.5848890841203602, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.0055134296417236, + "learning_rate": 1e-06, + "loss": 1.0565, + "mean_token_accuracy": 0.6816959381103516, + "num_tokens": 137963416.0, + "step": 5326 + }, + { + "epoch": 0.5849989018229739, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.189330816268921, + "learning_rate": 1e-06, + "loss": 1.0869, + "mean_token_accuracy": 0.6703282594680786, + "num_tokens": 137990308.0, + "step": 5327 + }, + { + "epoch": 0.5851087195255875, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.2756872177124023, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7249271273612976, + "num_tokens": 138012509.0, + "step": 5328 + }, + { + "epoch": 0.5852185372282012, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.5223052501678467, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7200674414634705, + "num_tokens": 138032504.0, + "step": 5329 + }, + { + "epoch": 0.5853283549308148, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.095672845840454, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7270107865333557, + "num_tokens": 138057651.0, + "step": 5330 + }, + { + "epoch": 0.5854381726334285, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.0944175720214844, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7197110652923584, + "num_tokens": 138086199.0, + "step": 5331 + }, + { + "epoch": 0.5855479903360422, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.1154632568359375, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7159194350242615, + "num_tokens": 138113881.0, + "step": 5332 + }, + { + "epoch": 0.5856578080386559, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.4351189136505127, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7203257083892822, + "num_tokens": 138135806.0, + "step": 5333 + }, + { + "epoch": 0.5857676257412695, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.316004514694214, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.711211085319519, + "num_tokens": 138159020.0, + "step": 5334 + }, + { + "epoch": 0.5858774434438832, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.1233274936676025, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7018526792526245, + "num_tokens": 138184601.0, + "step": 5335 + }, + { + "epoch": 0.5859872611464968, + "ewc_loss": 4.887580871582031e-05, + "grad_norm": 2.150392770767212, + "learning_rate": 1e-06, + "loss": 1.0311, + "mean_token_accuracy": 0.6856993436813354, + "num_tokens": 138212102.0, + "step": 5336 + }, + { + "epoch": 0.5860970788491104, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.279181957244873, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.684502124786377, + "num_tokens": 138238813.0, + "step": 5337 + }, + { + "epoch": 0.5862068965517241, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.448225975036621, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7000305652618408, + "num_tokens": 138260141.0, + "step": 5338 + }, + { + "epoch": 0.5863167142543378, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.1857919692993164, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7263782024383545, + "num_tokens": 138283651.0, + "step": 5339 + }, + { + "epoch": 0.5864265319569515, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 1.886191725730896, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.697726845741272, + "num_tokens": 138315435.0, + "step": 5340 + }, + { + "epoch": 0.5865363496595651, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.068023204803467, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7088659405708313, + "num_tokens": 138342803.0, + "step": 5341 + }, + { + "epoch": 0.5866461673621788, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.387526273727417, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6966531276702881, + "num_tokens": 138365491.0, + "step": 5342 + }, + { + "epoch": 0.5867559850647924, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 1.9578543901443481, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7200864553451538, + "num_tokens": 138395065.0, + "step": 5343 + }, + { + "epoch": 0.5868658027674061, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.327176809310913, + "learning_rate": 1e-06, + "loss": 1.0555, + "mean_token_accuracy": 0.6829604506492615, + "num_tokens": 138420604.0, + "step": 5344 + }, + { + "epoch": 0.5869756204700197, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.1509909629821777, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.7033690214157104, + "num_tokens": 138448634.0, + "step": 5345 + }, + { + "epoch": 0.5870854381726335, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.067426919937134, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.697285532951355, + "num_tokens": 138477350.0, + "step": 5346 + }, + { + "epoch": 0.5871952558752471, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.514585018157959, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7081102132797241, + "num_tokens": 138499062.0, + "step": 5347 + }, + { + "epoch": 0.5873050735778608, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.397036552429199, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7029397487640381, + "num_tokens": 138520145.0, + "step": 5348 + }, + { + "epoch": 0.5874148912804744, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.4598281383514404, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7022032737731934, + "num_tokens": 138541908.0, + "step": 5349 + }, + { + "epoch": 0.5875247089830881, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.07114839553833, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7385565042495728, + "num_tokens": 138566648.0, + "step": 5350 + }, + { + "epoch": 0.5876345266857017, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.06551456451416, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7164123058319092, + "num_tokens": 138593706.0, + "step": 5351 + }, + { + "epoch": 0.5877443443883154, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.100048065185547, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7055656313896179, + "num_tokens": 138622965.0, + "step": 5352 + }, + { + "epoch": 0.5878541620909291, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.2255373001098633, + "learning_rate": 1e-06, + "loss": 1.0252, + "mean_token_accuracy": 0.6902586221694946, + "num_tokens": 138649693.0, + "step": 5353 + }, + { + "epoch": 0.5879639797935428, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 1.8353487253189087, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6958736181259155, + "num_tokens": 138681641.0, + "step": 5354 + }, + { + "epoch": 0.5880737974961564, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.2599048614501953, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6970859169960022, + "num_tokens": 138705096.0, + "step": 5355 + }, + { + "epoch": 0.58818361519877, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.4795315265655518, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7118791341781616, + "num_tokens": 138725318.0, + "step": 5356 + }, + { + "epoch": 0.5882934329013837, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.1956677436828613, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7102882862091064, + "num_tokens": 138752583.0, + "step": 5357 + }, + { + "epoch": 0.5884032506039973, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.3006980419158936, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.711700975894928, + "num_tokens": 138775165.0, + "step": 5358 + }, + { + "epoch": 0.588513068306611, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.0552639961242676, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7110456824302673, + "num_tokens": 138803464.0, + "step": 5359 + }, + { + "epoch": 0.5886228860092246, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.210540294647217, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6943613886833191, + "num_tokens": 138830517.0, + "step": 5360 + }, + { + "epoch": 0.5887327037118384, + "ewc_loss": 4.9114227294921875e-05, + "grad_norm": 2.4161899089813232, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.7028375267982483, + "num_tokens": 138852743.0, + "step": 5361 + }, + { + "epoch": 0.588842521414452, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.2333807945251465, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7068217396736145, + "num_tokens": 138877770.0, + "step": 5362 + }, + { + "epoch": 0.5889523391170657, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.015540599822998, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6968748569488525, + "num_tokens": 138907114.0, + "step": 5363 + }, + { + "epoch": 0.5890621568196793, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.100508689880371, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7134106159210205, + "num_tokens": 138934234.0, + "step": 5364 + }, + { + "epoch": 0.589171974522293, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.5653247833251953, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7091363072395325, + "num_tokens": 138953750.0, + "step": 5365 + }, + { + "epoch": 0.5892817922249066, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.9489893913269043, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7036874890327454, + "num_tokens": 138985148.0, + "step": 5366 + }, + { + "epoch": 0.5893916099275203, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.184248924255371, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7198284864425659, + "num_tokens": 139009650.0, + "step": 5367 + }, + { + "epoch": 0.589501427630134, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.322437286376953, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7267221212387085, + "num_tokens": 139030423.0, + "step": 5368 + }, + { + "epoch": 0.5896112453327477, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.236337661743164, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7026827335357666, + "num_tokens": 139056646.0, + "step": 5369 + }, + { + "epoch": 0.5897210630353613, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.4813294410705566, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7181863784790039, + "num_tokens": 139076987.0, + "step": 5370 + }, + { + "epoch": 0.589830880737975, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.104776382446289, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7251759767532349, + "num_tokens": 139101805.0, + "step": 5371 + }, + { + "epoch": 0.5899406984405886, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.9781572818756104, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.699070394039154, + "num_tokens": 139129225.0, + "step": 5372 + }, + { + "epoch": 0.5900505161432023, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.1028249263763428, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7122934460639954, + "num_tokens": 139155600.0, + "step": 5373 + }, + { + "epoch": 0.5901603338458159, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.2007253170013428, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6889708042144775, + "num_tokens": 139183045.0, + "step": 5374 + }, + { + "epoch": 0.5902701515484297, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.8691543340682983, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7022407054901123, + "num_tokens": 139216042.0, + "step": 5375 + }, + { + "epoch": 0.5903799692510433, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.2698755264282227, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7065781950950623, + "num_tokens": 139240011.0, + "step": 5376 + }, + { + "epoch": 0.590489786953657, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.0968523025512695, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7131234407424927, + "num_tokens": 139268069.0, + "step": 5377 + }, + { + "epoch": 0.5905996046562706, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.4142963886260986, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7231649160385132, + "num_tokens": 139289186.0, + "step": 5378 + }, + { + "epoch": 0.5907094223588842, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.1369552612304688, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7073227167129517, + "num_tokens": 139318862.0, + "step": 5379 + }, + { + "epoch": 0.5908192400614979, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.376936197280884, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7035312652587891, + "num_tokens": 139342120.0, + "step": 5380 + }, + { + "epoch": 0.5909290577641115, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.028900623321533, + "learning_rate": 1e-06, + "loss": 1.119, + "mean_token_accuracy": 0.6660592555999756, + "num_tokens": 139374511.0, + "step": 5381 + }, + { + "epoch": 0.5910388754667253, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.4606940746307373, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7258966565132141, + "num_tokens": 139395596.0, + "step": 5382 + }, + { + "epoch": 0.5911486931693389, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.2104108333587646, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6947760581970215, + "num_tokens": 139420950.0, + "step": 5383 + }, + { + "epoch": 0.5912585108719526, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.1024668216705322, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7135452032089233, + "num_tokens": 139449011.0, + "step": 5384 + }, + { + "epoch": 0.5913683285745662, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.8988213539123535, + "learning_rate": 1e-06, + "loss": 1.0557, + "mean_token_accuracy": 0.6915345191955566, + "num_tokens": 139480828.0, + "step": 5385 + }, + { + "epoch": 0.5914781462771799, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.18597674369812, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7023041248321533, + "num_tokens": 139507020.0, + "step": 5386 + }, + { + "epoch": 0.5915879639797935, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.0859458446502686, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7012791633605957, + "num_tokens": 139534335.0, + "step": 5387 + }, + { + "epoch": 0.5916977816824072, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.073268175125122, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7091892957687378, + "num_tokens": 139561396.0, + "step": 5388 + }, + { + "epoch": 0.5918075993850208, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.1107568740844727, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6890813112258911, + "num_tokens": 139588667.0, + "step": 5389 + }, + { + "epoch": 0.5919174170876346, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.9893150329589844, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6873911619186401, + "num_tokens": 139621085.0, + "step": 5390 + }, + { + "epoch": 0.5920272347902482, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.1999335289001465, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7249457240104675, + "num_tokens": 139644535.0, + "step": 5391 + }, + { + "epoch": 0.5921370524928619, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.924646258354187, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.7047696113586426, + "num_tokens": 139675145.0, + "step": 5392 + }, + { + "epoch": 0.5922468701954755, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.2908291816711426, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7165418863296509, + "num_tokens": 139698271.0, + "step": 5393 + }, + { + "epoch": 0.5923566878980892, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.034681797027588, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7076703310012817, + "num_tokens": 139728259.0, + "step": 5394 + }, + { + "epoch": 0.5924665056007028, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 2.134598970413208, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7168624997138977, + "num_tokens": 139751081.0, + "step": 5395 + }, + { + "epoch": 0.5925763233033164, + "ewc_loss": 4.935264587402344e-05, + "grad_norm": 1.9401097297668457, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.697251558303833, + "num_tokens": 139783392.0, + "step": 5396 + }, + { + "epoch": 0.5926861410059302, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.1553802490234375, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7077010869979858, + "num_tokens": 139809224.0, + "step": 5397 + }, + { + "epoch": 0.5927959587085438, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.160965919494629, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7313287854194641, + "num_tokens": 139835349.0, + "step": 5398 + }, + { + "epoch": 0.5929057764111575, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.2558252811431885, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7237367630004883, + "num_tokens": 139858710.0, + "step": 5399 + }, + { + "epoch": 0.5930155941137711, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.1825473308563232, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6931929588317871, + "num_tokens": 139883838.0, + "step": 5400 + }, + { + "epoch": 0.5931254118163848, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.3433334827423096, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6977130174636841, + "num_tokens": 139908596.0, + "step": 5401 + }, + { + "epoch": 0.5932352295189984, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.117987871170044, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.7053282856941223, + "num_tokens": 139937133.0, + "step": 5402 + }, + { + "epoch": 0.5933450472216121, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.217777729034424, + "learning_rate": 1e-06, + "loss": 1.0307, + "mean_token_accuracy": 0.6896116733551025, + "num_tokens": 139962948.0, + "step": 5403 + }, + { + "epoch": 0.5934548649242258, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.0452234745025635, + "learning_rate": 1e-06, + "loss": 1.1131, + "mean_token_accuracy": 0.6638280153274536, + "num_tokens": 139993617.0, + "step": 5404 + }, + { + "epoch": 0.5935646826268395, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.0637147426605225, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6965396404266357, + "num_tokens": 140021652.0, + "step": 5405 + }, + { + "epoch": 0.5936745003294531, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.172633171081543, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6950977444648743, + "num_tokens": 140049633.0, + "step": 5406 + }, + { + "epoch": 0.5937843180320668, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.463874101638794, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7374244928359985, + "num_tokens": 140070854.0, + "step": 5407 + }, + { + "epoch": 0.5938941357346804, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.2682247161865234, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7094451189041138, + "num_tokens": 140094943.0, + "step": 5408 + }, + { + "epoch": 0.5940039534372941, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.5065438747406006, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7040911316871643, + "num_tokens": 140114169.0, + "step": 5409 + }, + { + "epoch": 0.5941137711399077, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.2054567337036133, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.7009567022323608, + "num_tokens": 140139128.0, + "step": 5410 + }, + { + "epoch": 0.5942235888425215, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.161449909210205, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7186996340751648, + "num_tokens": 140165992.0, + "step": 5411 + }, + { + "epoch": 0.5943334065451351, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.339832067489624, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6910239458084106, + "num_tokens": 140190420.0, + "step": 5412 + }, + { + "epoch": 0.5944432242477488, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.050798177719116, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7144982814788818, + "num_tokens": 140218466.0, + "step": 5413 + }, + { + "epoch": 0.5945530419503624, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.2558422088623047, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.720422625541687, + "num_tokens": 140241243.0, + "step": 5414 + }, + { + "epoch": 0.594662859652976, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.119225025177002, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7096243500709534, + "num_tokens": 140269198.0, + "step": 5415 + }, + { + "epoch": 0.5947726773555897, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.270927667617798, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7269502282142639, + "num_tokens": 140291919.0, + "step": 5416 + }, + { + "epoch": 0.5948824950582033, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.3517134189605713, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7182317972183228, + "num_tokens": 140314477.0, + "step": 5417 + }, + { + "epoch": 0.5949923127608171, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.309957981109619, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7063327431678772, + "num_tokens": 140337521.0, + "step": 5418 + }, + { + "epoch": 0.5951021304634307, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.280229091644287, + "learning_rate": 1e-06, + "loss": 1.0244, + "mean_token_accuracy": 0.701447606086731, + "num_tokens": 140360715.0, + "step": 5419 + }, + { + "epoch": 0.5952119481660444, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.3046984672546387, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7045378088951111, + "num_tokens": 140384016.0, + "step": 5420 + }, + { + "epoch": 0.595321765868658, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 2.3545422554016113, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7269020080566406, + "num_tokens": 140408011.0, + "step": 5421 + }, + { + "epoch": 0.5954315835712717, + "ewc_loss": 4.9591064453125e-05, + "grad_norm": 1.8889720439910889, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.7038241624832153, + "num_tokens": 140439623.0, + "step": 5422 + }, + { + "epoch": 0.5955414012738853, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2407443523406982, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6882890462875366, + "num_tokens": 140463762.0, + "step": 5423 + }, + { + "epoch": 0.595651218976499, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1744935512542725, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6870728731155396, + "num_tokens": 140489649.0, + "step": 5424 + }, + { + "epoch": 0.5957610366791126, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4523212909698486, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7067667245864868, + "num_tokens": 140509941.0, + "step": 5425 + }, + { + "epoch": 0.5958708543817264, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2011327743530273, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6976048946380615, + "num_tokens": 140535871.0, + "step": 5426 + }, + { + "epoch": 0.59598067208434, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1583216190338135, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7095478773117065, + "num_tokens": 140562930.0, + "step": 5427 + }, + { + "epoch": 0.5960904897869537, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.030937910079956, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7141404747962952, + "num_tokens": 140590643.0, + "step": 5428 + }, + { + "epoch": 0.5962003074895673, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.299041509628296, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6944639682769775, + "num_tokens": 140615200.0, + "step": 5429 + }, + { + "epoch": 0.596310125192181, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.947994351387024, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7176535129547119, + "num_tokens": 140645253.0, + "step": 5430 + }, + { + "epoch": 0.5964199428947946, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2279393672943115, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7226740121841431, + "num_tokens": 140667899.0, + "step": 5431 + }, + { + "epoch": 0.5965297605974083, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.3313100337982178, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6912375092506409, + "num_tokens": 140691109.0, + "step": 5432 + }, + { + "epoch": 0.596639578300022, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.149298667907715, + "learning_rate": 1e-06, + "loss": 1.0263, + "mean_token_accuracy": 0.6954379081726074, + "num_tokens": 140717798.0, + "step": 5433 + }, + { + "epoch": 0.5967493960026357, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.9357835054397583, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.6917010545730591, + "num_tokens": 140749119.0, + "step": 5434 + }, + { + "epoch": 0.5968592137052493, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.166626453399658, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7143100500106812, + "num_tokens": 140774407.0, + "step": 5435 + }, + { + "epoch": 0.596969031407863, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1999425888061523, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.696756899356842, + "num_tokens": 140800384.0, + "step": 5436 + }, + { + "epoch": 0.5970788491104766, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2723288536071777, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.718449592590332, + "num_tokens": 140823824.0, + "step": 5437 + }, + { + "epoch": 0.5971886668130902, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.3280138969421387, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7045611143112183, + "num_tokens": 140845377.0, + "step": 5438 + }, + { + "epoch": 0.5972984845157039, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.118462324142456, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7178939580917358, + "num_tokens": 140869853.0, + "step": 5439 + }, + { + "epoch": 0.5974083022183176, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2576143741607666, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7216216921806335, + "num_tokens": 140893376.0, + "step": 5440 + }, + { + "epoch": 0.5975181199209313, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.9159352779388428, + "learning_rate": 1e-06, + "loss": 1.0736, + "mean_token_accuracy": 0.6725603342056274, + "num_tokens": 140930049.0, + "step": 5441 + }, + { + "epoch": 0.5976279376235449, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.0001626014709473, + "learning_rate": 1e-06, + "loss": 1.0748, + "mean_token_accuracy": 0.6795419454574585, + "num_tokens": 140960774.0, + "step": 5442 + }, + { + "epoch": 0.5977377553261586, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.8729711771011353, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.699087381362915, + "num_tokens": 140991688.0, + "step": 5443 + }, + { + "epoch": 0.5978475730287722, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1354401111602783, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7119555473327637, + "num_tokens": 141018534.0, + "step": 5444 + }, + { + "epoch": 0.5979573907313859, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4358389377593994, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.6944815516471863, + "num_tokens": 141040847.0, + "step": 5445 + }, + { + "epoch": 0.5980672084339995, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.144482374191284, + "learning_rate": 1e-06, + "loss": 1.0286, + "mean_token_accuracy": 0.6931429505348206, + "num_tokens": 141068941.0, + "step": 5446 + }, + { + "epoch": 0.5981770261366133, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.5253093242645264, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7292104959487915, + "num_tokens": 141088818.0, + "step": 5447 + }, + { + "epoch": 0.5982868438392269, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.101392984390259, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7412407398223877, + "num_tokens": 141111050.0, + "step": 5448 + }, + { + "epoch": 0.5983966615418406, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.188363552093506, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7243276834487915, + "num_tokens": 141134434.0, + "step": 5449 + }, + { + "epoch": 0.5985064792444542, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.422029733657837, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7176249027252197, + "num_tokens": 141156074.0, + "step": 5450 + }, + { + "epoch": 0.5986162969470679, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1076102256774902, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6989456415176392, + "num_tokens": 141184073.0, + "step": 5451 + }, + { + "epoch": 0.5987261146496815, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.243251085281372, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7110613584518433, + "num_tokens": 141207261.0, + "step": 5452 + }, + { + "epoch": 0.5988359323522952, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4815433025360107, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7178730964660645, + "num_tokens": 141227943.0, + "step": 5453 + }, + { + "epoch": 0.5989457500549088, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.9327253103256226, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.6982672214508057, + "num_tokens": 141259938.0, + "step": 5454 + }, + { + "epoch": 0.5990555677575226, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4074547290802, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7039926052093506, + "num_tokens": 141281372.0, + "step": 5455 + }, + { + "epoch": 0.5991653854601362, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4376258850097656, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7204533815383911, + "num_tokens": 141301792.0, + "step": 5456 + }, + { + "epoch": 0.5992752031627498, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1555535793304443, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7028858661651611, + "num_tokens": 141330241.0, + "step": 5457 + }, + { + "epoch": 0.5993850208653635, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.0949692726135254, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.6935725212097168, + "num_tokens": 141360496.0, + "step": 5458 + }, + { + "epoch": 0.5994948385679771, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.175875186920166, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6907973289489746, + "num_tokens": 141386431.0, + "step": 5459 + }, + { + "epoch": 0.5996046562705908, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.0906710624694824, + "learning_rate": 1e-06, + "loss": 1.1037, + "mean_token_accuracy": 0.6680858135223389, + "num_tokens": 141415944.0, + "step": 5460 + }, + { + "epoch": 0.5997144739732044, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.077821969985962, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6997438073158264, + "num_tokens": 141444005.0, + "step": 5461 + }, + { + "epoch": 0.5998242916758182, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1645500659942627, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6899508237838745, + "num_tokens": 141470558.0, + "step": 5462 + }, + { + "epoch": 0.5999341093784318, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.0623908042907715, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6887359023094177, + "num_tokens": 141499619.0, + "step": 5463 + }, + { + "epoch": 0.6000439270810455, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.4347286224365234, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7215118408203125, + "num_tokens": 141519971.0, + "step": 5464 + }, + { + "epoch": 0.6001537447836591, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.081090211868286, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7239841222763062, + "num_tokens": 141547413.0, + "step": 5465 + }, + { + "epoch": 0.6002635624862728, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 1.9026058912277222, + "learning_rate": 1e-06, + "loss": 1.0701, + "mean_token_accuracy": 0.679171621799469, + "num_tokens": 141582883.0, + "step": 5466 + }, + { + "epoch": 0.6003733801888864, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.2117130756378174, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7046909332275391, + "num_tokens": 141609111.0, + "step": 5467 + }, + { + "epoch": 0.6004831978915001, + "ewc_loss": 4.982948303222656e-05, + "grad_norm": 2.1360602378845215, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7231947183609009, + "num_tokens": 141632586.0, + "step": 5468 + }, + { + "epoch": 0.6005930155941138, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.4690699577331543, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7146041393280029, + "num_tokens": 141653030.0, + "step": 5469 + }, + { + "epoch": 0.6007028332967275, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.068772554397583, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.692496657371521, + "num_tokens": 141680762.0, + "step": 5470 + }, + { + "epoch": 0.6008126509993411, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.332911729812622, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6908146739006042, + "num_tokens": 141705139.0, + "step": 5471 + }, + { + "epoch": 0.6009224687019548, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.188136339187622, + "learning_rate": 1e-06, + "loss": 1.0657, + "mean_token_accuracy": 0.68021160364151, + "num_tokens": 141732248.0, + "step": 5472 + }, + { + "epoch": 0.6010322864045684, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.3840954303741455, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6882895231246948, + "num_tokens": 141756347.0, + "step": 5473 + }, + { + "epoch": 0.601142104107182, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 1.8523811101913452, + "learning_rate": 1e-06, + "loss": 1.0425, + "mean_token_accuracy": 0.6884924173355103, + "num_tokens": 141790860.0, + "step": 5474 + }, + { + "epoch": 0.6012519218097957, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.1080498695373535, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7033215761184692, + "num_tokens": 141818768.0, + "step": 5475 + }, + { + "epoch": 0.6013617395124095, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 1.977253794670105, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7175378799438477, + "num_tokens": 141846167.0, + "step": 5476 + }, + { + "epoch": 0.6014715572150231, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.448756694793701, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7072733044624329, + "num_tokens": 141867109.0, + "step": 5477 + }, + { + "epoch": 0.6015813749176367, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.1214239597320557, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.688948392868042, + "num_tokens": 141894244.0, + "step": 5478 + }, + { + "epoch": 0.6016911926202504, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.3592185974121094, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.713119387626648, + "num_tokens": 141918529.0, + "step": 5479 + }, + { + "epoch": 0.601801010322864, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.244615077972412, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7110798358917236, + "num_tokens": 141941287.0, + "step": 5480 + }, + { + "epoch": 0.6019108280254777, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.334496259689331, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6836040019989014, + "num_tokens": 141965353.0, + "step": 5481 + }, + { + "epoch": 0.6020206457280913, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.223247528076172, + "learning_rate": 1e-06, + "loss": 1.1035, + "mean_token_accuracy": 0.6755358576774597, + "num_tokens": 141992824.0, + "step": 5482 + }, + { + "epoch": 0.602130463430705, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.361355781555176, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6906013488769531, + "num_tokens": 142016379.0, + "step": 5483 + }, + { + "epoch": 0.6022402811333187, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 1.905390977859497, + "learning_rate": 1e-06, + "loss": 1.0475, + "mean_token_accuracy": 0.6830679178237915, + "num_tokens": 142051236.0, + "step": 5484 + }, + { + "epoch": 0.6023500988359324, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.285452127456665, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6934252381324768, + "num_tokens": 142076418.0, + "step": 5485 + }, + { + "epoch": 0.602459916538546, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.565667152404785, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7209630608558655, + "num_tokens": 142096091.0, + "step": 5486 + }, + { + "epoch": 0.6025697342411597, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.098536968231201, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7063657641410828, + "num_tokens": 142124345.0, + "step": 5487 + }, + { + "epoch": 0.6026795519437733, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.1969335079193115, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6881976127624512, + "num_tokens": 142154203.0, + "step": 5488 + }, + { + "epoch": 0.602789369646387, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.2310054302215576, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.6998592615127563, + "num_tokens": 142178440.0, + "step": 5489 + }, + { + "epoch": 0.6028991873490006, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.422074317932129, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7081326246261597, + "num_tokens": 142203260.0, + "step": 5490 + }, + { + "epoch": 0.6030090050516144, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.5904769897460938, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7135876417160034, + "num_tokens": 142222642.0, + "step": 5491 + }, + { + "epoch": 0.603118822754228, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.143073797225952, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7128515839576721, + "num_tokens": 142248504.0, + "step": 5492 + }, + { + "epoch": 0.6032286404568417, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.2200233936309814, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7069306373596191, + "num_tokens": 142275888.0, + "step": 5493 + }, + { + "epoch": 0.6033384581594553, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.2027621269226074, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6970342993736267, + "num_tokens": 142302175.0, + "step": 5494 + }, + { + "epoch": 0.603448275862069, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.3678529262542725, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7094902992248535, + "num_tokens": 142326863.0, + "step": 5495 + }, + { + "epoch": 0.6035580935646826, + "ewc_loss": 5.0067901611328125e-05, + "grad_norm": 2.0415127277374268, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7048376798629761, + "num_tokens": 142355356.0, + "step": 5496 + }, + { + "epoch": 0.6036679112672962, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.236365556716919, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7065790891647339, + "num_tokens": 142379442.0, + "step": 5497 + }, + { + "epoch": 0.60377772896991, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.3344035148620605, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.703534722328186, + "num_tokens": 142402486.0, + "step": 5498 + }, + { + "epoch": 0.6038875466725236, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.263828754425049, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6941541433334351, + "num_tokens": 142427756.0, + "step": 5499 + }, + { + "epoch": 0.6039973643751373, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.1035122871398926, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7036323547363281, + "num_tokens": 142456504.0, + "step": 5500 + }, + { + "epoch": 0.6041071820777509, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.089853286743164, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7192245721817017, + "num_tokens": 142481999.0, + "step": 5501 + }, + { + "epoch": 0.6042169997803646, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.1357674598693848, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7143626809120178, + "num_tokens": 142508085.0, + "step": 5502 + }, + { + "epoch": 0.6043268174829782, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.3538599014282227, + "learning_rate": 1e-06, + "loss": 1.0542, + "mean_token_accuracy": 0.6830271482467651, + "num_tokens": 142533502.0, + "step": 5503 + }, + { + "epoch": 0.6044366351855919, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.126215934753418, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.707141637802124, + "num_tokens": 142562232.0, + "step": 5504 + }, + { + "epoch": 0.6045464528882056, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.290613889694214, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.719007134437561, + "num_tokens": 142584535.0, + "step": 5505 + }, + { + "epoch": 0.6046562705908193, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.0668423175811768, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.69820237159729, + "num_tokens": 142613548.0, + "step": 5506 + }, + { + "epoch": 0.6047660882934329, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.4071738719940186, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7043637037277222, + "num_tokens": 142634501.0, + "step": 5507 + }, + { + "epoch": 0.6048759059960466, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.1521964073181152, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7109842300415039, + "num_tokens": 142660053.0, + "step": 5508 + }, + { + "epoch": 0.6049857236986602, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.071002960205078, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.7019457817077637, + "num_tokens": 142686587.0, + "step": 5509 + }, + { + "epoch": 0.6050955414012739, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.240814208984375, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.6998583674430847, + "num_tokens": 142710842.0, + "step": 5510 + }, + { + "epoch": 0.6052053591038875, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 1.903430461883545, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.6988663673400879, + "num_tokens": 142744140.0, + "step": 5511 + }, + { + "epoch": 0.6053151768065012, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.18749737739563, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7067173719406128, + "num_tokens": 142769131.0, + "step": 5512 + }, + { + "epoch": 0.6054249945091149, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.1418371200561523, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.7094932198524475, + "num_tokens": 142798422.0, + "step": 5513 + }, + { + "epoch": 0.6055348122117286, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.3524200916290283, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7126266956329346, + "num_tokens": 142820626.0, + "step": 5514 + }, + { + "epoch": 0.6056446299143422, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.21341872215271, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.714138388633728, + "num_tokens": 142842981.0, + "step": 5515 + }, + { + "epoch": 0.6057544476169558, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.491655111312866, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7085992097854614, + "num_tokens": 142862928.0, + "step": 5516 + }, + { + "epoch": 0.6058642653195695, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.2123281955718994, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.6942272782325745, + "num_tokens": 142891872.0, + "step": 5517 + }, + { + "epoch": 0.6059740830221831, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.0600216388702393, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.70097815990448, + "num_tokens": 142919028.0, + "step": 5518 + }, + { + "epoch": 0.6060839007247968, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 1.973547339439392, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7176454067230225, + "num_tokens": 142947374.0, + "step": 5519 + }, + { + "epoch": 0.6061937184274105, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.2054686546325684, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6874179840087891, + "num_tokens": 142974095.0, + "step": 5520 + }, + { + "epoch": 0.6063035361300242, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.4828226566314697, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6889868378639221, + "num_tokens": 142995967.0, + "step": 5521 + }, + { + "epoch": 0.6064133538326378, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.0721704959869385, + "learning_rate": 1e-06, + "loss": 1.0608, + "mean_token_accuracy": 0.6793885231018066, + "num_tokens": 143025221.0, + "step": 5522 + }, + { + "epoch": 0.6065231715352515, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.0977623462677, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7040936350822449, + "num_tokens": 143051527.0, + "step": 5523 + }, + { + "epoch": 0.6066329892378651, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.3745548725128174, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6966140866279602, + "num_tokens": 143075544.0, + "step": 5524 + }, + { + "epoch": 0.6067428069404788, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.1479263305664062, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7209033966064453, + "num_tokens": 143099855.0, + "step": 5525 + }, + { + "epoch": 0.6068526246430924, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 1.9837324619293213, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7144585251808167, + "num_tokens": 143129817.0, + "step": 5526 + }, + { + "epoch": 0.6069624423457062, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.47361421585083, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6976654529571533, + "num_tokens": 143150698.0, + "step": 5527 + }, + { + "epoch": 0.6070722600483198, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.4689865112304688, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7056772708892822, + "num_tokens": 143171043.0, + "step": 5528 + }, + { + "epoch": 0.6071820777509335, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.3128623962402344, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7146488428115845, + "num_tokens": 143193800.0, + "step": 5529 + }, + { + "epoch": 0.6072918954535471, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.0860886573791504, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6966682076454163, + "num_tokens": 143222251.0, + "step": 5530 + }, + { + "epoch": 0.6074017131561608, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.423032522201538, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7093954086303711, + "num_tokens": 143245623.0, + "step": 5531 + }, + { + "epoch": 0.6075115308587744, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.189100742340088, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7200597524642944, + "num_tokens": 143271328.0, + "step": 5532 + }, + { + "epoch": 0.607621348561388, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.5468287467956543, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7180346250534058, + "num_tokens": 143290653.0, + "step": 5533 + }, + { + "epoch": 0.6077311662640018, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.136164903640747, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7237575650215149, + "num_tokens": 143316964.0, + "step": 5534 + }, + { + "epoch": 0.6078409839666155, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.4015040397644043, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7228532433509827, + "num_tokens": 143339087.0, + "step": 5535 + }, + { + "epoch": 0.6079508016692291, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 1.8929022550582886, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.7004939317703247, + "num_tokens": 143372999.0, + "step": 5536 + }, + { + "epoch": 0.6080606193718427, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.2522010803222656, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.704017698764801, + "num_tokens": 143398034.0, + "step": 5537 + }, + { + "epoch": 0.6081704370744564, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.363555431365967, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.719869077205658, + "num_tokens": 143419511.0, + "step": 5538 + }, + { + "epoch": 0.60828025477707, + "ewc_loss": 5.030632019042969e-05, + "grad_norm": 2.053114891052246, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6902707815170288, + "num_tokens": 143447620.0, + "step": 5539 + }, + { + "epoch": 0.6083900724796837, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.3797073364257812, + "learning_rate": 1e-06, + "loss": 1.045, + "mean_token_accuracy": 0.6944726705551147, + "num_tokens": 143470927.0, + "step": 5540 + }, + { + "epoch": 0.6084998901822973, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.2148654460906982, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7042796611785889, + "num_tokens": 143495362.0, + "step": 5541 + }, + { + "epoch": 0.6086097078849111, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.1037044525146484, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.714430570602417, + "num_tokens": 143522298.0, + "step": 5542 + }, + { + "epoch": 0.6087195255875247, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.5787129402160645, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7190392017364502, + "num_tokens": 143540762.0, + "step": 5543 + }, + { + "epoch": 0.6088293432901384, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.406367063522339, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7300960421562195, + "num_tokens": 143562935.0, + "step": 5544 + }, + { + "epoch": 0.608939160992752, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.3768486976623535, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6967364549636841, + "num_tokens": 143586717.0, + "step": 5545 + }, + { + "epoch": 0.6090489786953657, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.122779607772827, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6906872391700745, + "num_tokens": 143613936.0, + "step": 5546 + }, + { + "epoch": 0.6091587963979793, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.117583990097046, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7087500691413879, + "num_tokens": 143641181.0, + "step": 5547 + }, + { + "epoch": 0.609268614100593, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.160404682159424, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7043615579605103, + "num_tokens": 143666141.0, + "step": 5548 + }, + { + "epoch": 0.6093784318032067, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.105262041091919, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7001304030418396, + "num_tokens": 143694096.0, + "step": 5549 + }, + { + "epoch": 0.6094882495058204, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.02139949798584, + "learning_rate": 1e-06, + "loss": 1.0351, + "mean_token_accuracy": 0.6970069408416748, + "num_tokens": 143724483.0, + "step": 5550 + }, + { + "epoch": 0.609598067208434, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.1788196563720703, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7102154493331909, + "num_tokens": 143748850.0, + "step": 5551 + }, + { + "epoch": 0.6097078849110477, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.124436140060425, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6975712776184082, + "num_tokens": 143776175.0, + "step": 5552 + }, + { + "epoch": 0.6098177026136613, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.191725015640259, + "learning_rate": 1e-06, + "loss": 1.067, + "mean_token_accuracy": 0.689401388168335, + "num_tokens": 143804042.0, + "step": 5553 + }, + { + "epoch": 0.609927520316275, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.167505979537964, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.707496166229248, + "num_tokens": 143831258.0, + "step": 5554 + }, + { + "epoch": 0.6100373380188886, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 1.991486668586731, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6930471658706665, + "num_tokens": 143861686.0, + "step": 5555 + }, + { + "epoch": 0.6101471557215024, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.406489610671997, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7012515664100647, + "num_tokens": 143883454.0, + "step": 5556 + }, + { + "epoch": 0.610256973424116, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.2241718769073486, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7071963548660278, + "num_tokens": 143907384.0, + "step": 5557 + }, + { + "epoch": 0.6103667911267296, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.193868637084961, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.693172037601471, + "num_tokens": 143934663.0, + "step": 5558 + }, + { + "epoch": 0.6104766088293433, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.0571510791778564, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7109242081642151, + "num_tokens": 143962719.0, + "step": 5559 + }, + { + "epoch": 0.6105864265319569, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.5222980976104736, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7116566896438599, + "num_tokens": 143983202.0, + "step": 5560 + }, + { + "epoch": 0.6106962442345706, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.4417598247528076, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7148270606994629, + "num_tokens": 144004714.0, + "step": 5561 + }, + { + "epoch": 0.6108060619371842, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.5395021438598633, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6985982060432434, + "num_tokens": 144025861.0, + "step": 5562 + }, + { + "epoch": 0.610915879639798, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.34818959236145, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.714577317237854, + "num_tokens": 144048018.0, + "step": 5563 + }, + { + "epoch": 0.6110256973424116, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.099729537963867, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7264083623886108, + "num_tokens": 144073098.0, + "step": 5564 + }, + { + "epoch": 0.6111355150450253, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 1.9941387176513672, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.6918507814407349, + "num_tokens": 144103692.0, + "step": 5565 + }, + { + "epoch": 0.6112453327476389, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.1515204906463623, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.6998656988143921, + "num_tokens": 144130071.0, + "step": 5566 + }, + { + "epoch": 0.6113551504502526, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 1.8772454261779785, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7126039862632751, + "num_tokens": 144160736.0, + "step": 5567 + }, + { + "epoch": 0.6114649681528662, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.155233383178711, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7161167860031128, + "num_tokens": 144187160.0, + "step": 5568 + }, + { + "epoch": 0.6115747858554799, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.2962305545806885, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6923259496688843, + "num_tokens": 144215373.0, + "step": 5569 + }, + { + "epoch": 0.6116846035580935, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 1.9324891567230225, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7229700088500977, + "num_tokens": 144246010.0, + "step": 5570 + }, + { + "epoch": 0.6117944212607073, + "ewc_loss": 5.054473876953125e-05, + "grad_norm": 2.271416425704956, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6892908811569214, + "num_tokens": 144271534.0, + "step": 5571 + }, + { + "epoch": 0.6119042389633209, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.0751888751983643, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6982244253158569, + "num_tokens": 144296962.0, + "step": 5572 + }, + { + "epoch": 0.6120140566659346, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.3933041095733643, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.71735018491745, + "num_tokens": 144319660.0, + "step": 5573 + }, + { + "epoch": 0.6121238743685482, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.0659821033477783, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7053929567337036, + "num_tokens": 144345996.0, + "step": 5574 + }, + { + "epoch": 0.6122336920711619, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.2100934982299805, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7186269760131836, + "num_tokens": 144370199.0, + "step": 5575 + }, + { + "epoch": 0.6123435097737755, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1573150157928467, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6840878129005432, + "num_tokens": 144398358.0, + "step": 5576 + }, + { + "epoch": 0.6124533274763891, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.2990713119506836, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7111048698425293, + "num_tokens": 144422666.0, + "step": 5577 + }, + { + "epoch": 0.6125631451790029, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 1.9970026016235352, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6908957958221436, + "num_tokens": 144453770.0, + "step": 5578 + }, + { + "epoch": 0.6126729628816165, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 1.9283310174942017, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6982212662696838, + "num_tokens": 144487826.0, + "step": 5579 + }, + { + "epoch": 0.6127827805842302, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.265244245529175, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7213664054870605, + "num_tokens": 144510583.0, + "step": 5580 + }, + { + "epoch": 0.6128925982868438, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1541879177093506, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.705815315246582, + "num_tokens": 144537090.0, + "step": 5581 + }, + { + "epoch": 0.6130024159894575, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 1.7873002290725708, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7071904540061951, + "num_tokens": 144573721.0, + "step": 5582 + }, + { + "epoch": 0.6131122336920711, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1164345741271973, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.7004858255386353, + "num_tokens": 144600228.0, + "step": 5583 + }, + { + "epoch": 0.6132220513946848, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.190847396850586, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6980513334274292, + "num_tokens": 144624825.0, + "step": 5584 + }, + { + "epoch": 0.6133318690972985, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.024160146713257, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7074676156044006, + "num_tokens": 144653870.0, + "step": 5585 + }, + { + "epoch": 0.6134416867999122, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.026082754135132, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7006697654724121, + "num_tokens": 144682126.0, + "step": 5586 + }, + { + "epoch": 0.6135515045025258, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.087254762649536, + "learning_rate": 1e-06, + "loss": 1.0649, + "mean_token_accuracy": 0.6790720224380493, + "num_tokens": 144711108.0, + "step": 5587 + }, + { + "epoch": 0.6136613222051395, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.3223938941955566, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7377254962921143, + "num_tokens": 144732059.0, + "step": 5588 + }, + { + "epoch": 0.6137711399077531, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.2554452419281006, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7188476324081421, + "num_tokens": 144756402.0, + "step": 5589 + }, + { + "epoch": 0.6138809576103668, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.477874517440796, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.6997695565223694, + "num_tokens": 144778478.0, + "step": 5590 + }, + { + "epoch": 0.6139907753129804, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1760404109954834, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6977558135986328, + "num_tokens": 144803171.0, + "step": 5591 + }, + { + "epoch": 0.6141005930155942, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.2060320377349854, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6961426138877869, + "num_tokens": 144829122.0, + "step": 5592 + }, + { + "epoch": 0.6142104107182078, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.3093554973602295, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7036976218223572, + "num_tokens": 144853840.0, + "step": 5593 + }, + { + "epoch": 0.6143202284208215, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.158459186553955, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7284178733825684, + "num_tokens": 144879799.0, + "step": 5594 + }, + { + "epoch": 0.6144300461234351, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.200188636779785, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7298554182052612, + "num_tokens": 144903876.0, + "step": 5595 + }, + { + "epoch": 0.6145398638260487, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1232287883758545, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6890860795974731, + "num_tokens": 144932095.0, + "step": 5596 + }, + { + "epoch": 0.6146496815286624, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.2982258796691895, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7159457802772522, + "num_tokens": 144955621.0, + "step": 5597 + }, + { + "epoch": 0.614759499231276, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.1108903884887695, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.703066349029541, + "num_tokens": 144985577.0, + "step": 5598 + }, + { + "epoch": 0.6148693169338898, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.451876640319824, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7296180725097656, + "num_tokens": 145005958.0, + "step": 5599 + }, + { + "epoch": 0.6149791346365034, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.4587719440460205, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7178716063499451, + "num_tokens": 145025793.0, + "step": 5600 + }, + { + "epoch": 0.6150889523391171, + "ewc_loss": 5.078315734863281e-05, + "grad_norm": 2.023871421813965, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.6998544335365295, + "num_tokens": 145056561.0, + "step": 5601 + }, + { + "epoch": 0.6151987700417307, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.0335190296173096, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7210404872894287, + "num_tokens": 145084877.0, + "step": 5602 + }, + { + "epoch": 0.6153085877443444, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.3123648166656494, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7276418209075928, + "num_tokens": 145108660.0, + "step": 5603 + }, + { + "epoch": 0.615418405446958, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2717490196228027, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7076081037521362, + "num_tokens": 145132828.0, + "step": 5604 + }, + { + "epoch": 0.6155282231495717, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2880961894989014, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7015729546546936, + "num_tokens": 145154962.0, + "step": 5605 + }, + { + "epoch": 0.6156380408521853, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.408238649368286, + "learning_rate": 1e-06, + "loss": 1.0285, + "mean_token_accuracy": 0.6840066909790039, + "num_tokens": 145182406.0, + "step": 5606 + }, + { + "epoch": 0.6157478585547991, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.166551351547241, + "learning_rate": 1e-06, + "loss": 1.0213, + "mean_token_accuracy": 0.6866700649261475, + "num_tokens": 145208430.0, + "step": 5607 + }, + { + "epoch": 0.6158576762574127, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2344272136688232, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7110263109207153, + "num_tokens": 145236623.0, + "step": 5608 + }, + { + "epoch": 0.6159674939600264, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.191856861114502, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7078409194946289, + "num_tokens": 145261618.0, + "step": 5609 + }, + { + "epoch": 0.61607731166264, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.126375913619995, + "learning_rate": 1e-06, + "loss": 1.0494, + "mean_token_accuracy": 0.6963167786598206, + "num_tokens": 145291071.0, + "step": 5610 + }, + { + "epoch": 0.6161871293652537, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.8884042501449585, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7351141571998596, + "num_tokens": 145323857.0, + "step": 5611 + }, + { + "epoch": 0.6162969470678673, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.1382319927215576, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6909396052360535, + "num_tokens": 145353905.0, + "step": 5612 + }, + { + "epoch": 0.616406764770481, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.9486134052276611, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7216508984565735, + "num_tokens": 145384056.0, + "step": 5613 + }, + { + "epoch": 0.6165165824730947, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.151963710784912, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7270817160606384, + "num_tokens": 145410449.0, + "step": 5614 + }, + { + "epoch": 0.6166264001757084, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2702932357788086, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7270199060440063, + "num_tokens": 145433886.0, + "step": 5615 + }, + { + "epoch": 0.616736217878322, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.460008382797241, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6875678300857544, + "num_tokens": 145455812.0, + "step": 5616 + }, + { + "epoch": 0.6168460355809356, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.1177327632904053, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.68783038854599, + "num_tokens": 145483198.0, + "step": 5617 + }, + { + "epoch": 0.6169558532835493, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.239323854446411, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7067861557006836, + "num_tokens": 145508062.0, + "step": 5618 + }, + { + "epoch": 0.6170656709861629, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2018191814422607, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6959974765777588, + "num_tokens": 145533803.0, + "step": 5619 + }, + { + "epoch": 0.6171754886887766, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.101529121398926, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.6902147531509399, + "num_tokens": 145561163.0, + "step": 5620 + }, + { + "epoch": 0.6172853063913903, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.0730767250061035, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7055551409721375, + "num_tokens": 145587924.0, + "step": 5621 + }, + { + "epoch": 0.617395124094004, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.0779972076416016, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7127640247344971, + "num_tokens": 145614186.0, + "step": 5622 + }, + { + "epoch": 0.6175049417966176, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2292697429656982, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7153463363647461, + "num_tokens": 145636444.0, + "step": 5623 + }, + { + "epoch": 0.6176147594992313, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.040806531906128, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7031429409980774, + "num_tokens": 145665371.0, + "step": 5624 + }, + { + "epoch": 0.6177245772018449, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.1827621459960938, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7033588290214539, + "num_tokens": 145689331.0, + "step": 5625 + }, + { + "epoch": 0.6178343949044586, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.478480577468872, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7024749517440796, + "num_tokens": 145709752.0, + "step": 5626 + }, + { + "epoch": 0.6179442126070722, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.3269128799438477, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7152040600776672, + "num_tokens": 145732554.0, + "step": 5627 + }, + { + "epoch": 0.618054030309686, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.8480387926101685, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.6835995316505432, + "num_tokens": 145768213.0, + "step": 5628 + }, + { + "epoch": 0.6181638480122996, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.451195478439331, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7156780958175659, + "num_tokens": 145788714.0, + "step": 5629 + }, + { + "epoch": 0.6182736657149133, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.398108720779419, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7188469171524048, + "num_tokens": 145811217.0, + "step": 5630 + }, + { + "epoch": 0.6183834834175269, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2887120246887207, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6978345513343811, + "num_tokens": 145834633.0, + "step": 5631 + }, + { + "epoch": 0.6184933011201406, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.44767427444458, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7063688635826111, + "num_tokens": 145855250.0, + "step": 5632 + }, + { + "epoch": 0.6186031188227542, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.937376856803894, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6951632499694824, + "num_tokens": 145888420.0, + "step": 5633 + }, + { + "epoch": 0.6187129365253679, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.271700859069824, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7098926901817322, + "num_tokens": 145911814.0, + "step": 5634 + }, + { + "epoch": 0.6188227542279815, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.175736427307129, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7440876960754395, + "num_tokens": 145937059.0, + "step": 5635 + }, + { + "epoch": 0.6189325719305953, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2028720378875732, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7031406164169312, + "num_tokens": 145964732.0, + "step": 5636 + }, + { + "epoch": 0.6190423896332089, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.183450698852539, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.7009153366088867, + "num_tokens": 145990592.0, + "step": 5637 + }, + { + "epoch": 0.6191522073358225, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.9938652515411377, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6929137706756592, + "num_tokens": 146020581.0, + "step": 5638 + }, + { + "epoch": 0.6192620250384362, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.253007411956787, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.6929523348808289, + "num_tokens": 146045852.0, + "step": 5639 + }, + { + "epoch": 0.6193718427410498, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2156646251678467, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7426677942276001, + "num_tokens": 146069988.0, + "step": 5640 + }, + { + "epoch": 0.6194816604436635, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.4819765090942383, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7068623900413513, + "num_tokens": 146090334.0, + "step": 5641 + }, + { + "epoch": 0.6195914781462771, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2087550163269043, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7127435207366943, + "num_tokens": 146113724.0, + "step": 5642 + }, + { + "epoch": 0.6197012958488909, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.223146677017212, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.712393581867218, + "num_tokens": 146138973.0, + "step": 5643 + }, + { + "epoch": 0.6198111135515045, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 1.9172800779342651, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7315171957015991, + "num_tokens": 146168630.0, + "step": 5644 + }, + { + "epoch": 0.6199209312541182, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.005059242248535, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6907075643539429, + "num_tokens": 146198815.0, + "step": 5645 + }, + { + "epoch": 0.6200307489567318, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.48498272895813, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7257041335105896, + "num_tokens": 146219566.0, + "step": 5646 + }, + { + "epoch": 0.6201405666593455, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.273355484008789, + "learning_rate": 1e-06, + "loss": 1.0635, + "mean_token_accuracy": 0.6835379600524902, + "num_tokens": 146243718.0, + "step": 5647 + }, + { + "epoch": 0.6202503843619591, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.224311590194702, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6968396902084351, + "num_tokens": 146268636.0, + "step": 5648 + }, + { + "epoch": 0.6203602020645728, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.3087079524993896, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7154629230499268, + "num_tokens": 146291335.0, + "step": 5649 + }, + { + "epoch": 0.6204700197671865, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.429208993911743, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7160961627960205, + "num_tokens": 146311255.0, + "step": 5650 + }, + { + "epoch": 0.6205798374698002, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.284193754196167, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.6987994909286499, + "num_tokens": 146335164.0, + "step": 5651 + }, + { + "epoch": 0.6206896551724138, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.5774776935577393, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7268839478492737, + "num_tokens": 146354287.0, + "step": 5652 + }, + { + "epoch": 0.6207994728750275, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.170950412750244, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7203408479690552, + "num_tokens": 146380335.0, + "step": 5653 + }, + { + "epoch": 0.6209092905776411, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.3654706478118896, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6999796628952026, + "num_tokens": 146401453.0, + "step": 5654 + }, + { + "epoch": 0.6210191082802548, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.0008652210235596, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6996787190437317, + "num_tokens": 146431847.0, + "step": 5655 + }, + { + "epoch": 0.6211289259828684, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.21769118309021, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.7040908932685852, + "num_tokens": 146456537.0, + "step": 5656 + }, + { + "epoch": 0.6212387436854822, + "ewc_loss": 5.1021575927734375e-05, + "grad_norm": 2.2495899200439453, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.7074412107467651, + "num_tokens": 146479809.0, + "step": 5657 + }, + { + "epoch": 0.6213485613880958, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.1119260787963867, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6959154605865479, + "num_tokens": 146505965.0, + "step": 5658 + }, + { + "epoch": 0.6214583790907094, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.236328125, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7104701995849609, + "num_tokens": 146531551.0, + "step": 5659 + }, + { + "epoch": 0.6215681967933231, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 1.931504249572754, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6805707216262817, + "num_tokens": 146563551.0, + "step": 5660 + }, + { + "epoch": 0.6216780144959367, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.137770652770996, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6917600035667419, + "num_tokens": 146591120.0, + "step": 5661 + }, + { + "epoch": 0.6217878321985504, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.4280502796173096, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.719866156578064, + "num_tokens": 146611965.0, + "step": 5662 + }, + { + "epoch": 0.621897649901164, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 1.8019030094146729, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7045543193817139, + "num_tokens": 146648461.0, + "step": 5663 + }, + { + "epoch": 0.6220074676037777, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.1306474208831787, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.6999502182006836, + "num_tokens": 146676427.0, + "step": 5664 + }, + { + "epoch": 0.6221172853063914, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.2296054363250732, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7191542387008667, + "num_tokens": 146700165.0, + "step": 5665 + }, + { + "epoch": 0.6222271030090051, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.2067365646362305, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.700660228729248, + "num_tokens": 146724996.0, + "step": 5666 + }, + { + "epoch": 0.6223369207116187, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.2638258934020996, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7116521596908569, + "num_tokens": 146750905.0, + "step": 5667 + }, + { + "epoch": 0.6224467384142324, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.1458990573883057, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7055915594100952, + "num_tokens": 146776383.0, + "step": 5668 + }, + { + "epoch": 0.622556556116846, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.167003631591797, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7319971919059753, + "num_tokens": 146800241.0, + "step": 5669 + }, + { + "epoch": 0.6226663738194597, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 1.9689230918884277, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7089070677757263, + "num_tokens": 146831318.0, + "step": 5670 + }, + { + "epoch": 0.6227761915220733, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.391343116760254, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.700324535369873, + "num_tokens": 146853834.0, + "step": 5671 + }, + { + "epoch": 0.6228860092246871, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.5400519371032715, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7344661951065063, + "num_tokens": 146872499.0, + "step": 5672 + }, + { + "epoch": 0.6229958269273007, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.239018440246582, + "learning_rate": 1e-06, + "loss": 1.068, + "mean_token_accuracy": 0.6826282739639282, + "num_tokens": 146897374.0, + "step": 5673 + }, + { + "epoch": 0.6231056446299144, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 1.9623597860336304, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6938285827636719, + "num_tokens": 146927243.0, + "step": 5674 + }, + { + "epoch": 0.623215462332528, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.2798192501068115, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7155757546424866, + "num_tokens": 146949952.0, + "step": 5675 + }, + { + "epoch": 0.6233252800351416, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.138728380203247, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6891736388206482, + "num_tokens": 146979182.0, + "step": 5676 + }, + { + "epoch": 0.6234350977377553, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.024904489517212, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6977636814117432, + "num_tokens": 147009131.0, + "step": 5677 + }, + { + "epoch": 0.6235449154403689, + "ewc_loss": 5.14984130859375e-05, + "grad_norm": 2.106684923171997, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6980140209197998, + "num_tokens": 147038561.0, + "step": 5678 + }, + { + "epoch": 0.6236547331429827, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.693185329437256, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7309970259666443, + "num_tokens": 147055735.0, + "step": 5679 + }, + { + "epoch": 0.6237645508455963, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.353961706161499, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7167279720306396, + "num_tokens": 147076421.0, + "step": 5680 + }, + { + "epoch": 0.62387436854821, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.264204740524292, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7091119885444641, + "num_tokens": 147099457.0, + "step": 5681 + }, + { + "epoch": 0.6239841862508236, + "ewc_loss": 5.173683166503906e-05, + "grad_norm": 2.2401320934295654, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6950381398200989, + "num_tokens": 147125823.0, + "step": 5682 + }, + { + "epoch": 0.6240940039534373, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.0458526611328125, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7068793773651123, + "num_tokens": 147154551.0, + "step": 5683 + }, + { + "epoch": 0.6242038216560509, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.3674368858337402, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7035805583000183, + "num_tokens": 147177596.0, + "step": 5684 + }, + { + "epoch": 0.6243136393586646, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.3758163452148438, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6968204975128174, + "num_tokens": 147199678.0, + "step": 5685 + }, + { + "epoch": 0.6244234570612783, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.174791097640991, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7294354438781738, + "num_tokens": 147223851.0, + "step": 5686 + }, + { + "epoch": 0.624533274763892, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.0901992321014404, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7182793021202087, + "num_tokens": 147251227.0, + "step": 5687 + }, + { + "epoch": 0.6246430924665056, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.2780098915100098, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.687980055809021, + "num_tokens": 147277361.0, + "step": 5688 + }, + { + "epoch": 0.6247529101691193, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.3506546020507812, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7045825719833374, + "num_tokens": 147297440.0, + "step": 5689 + }, + { + "epoch": 0.6248627278717329, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.133253812789917, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7089875936508179, + "num_tokens": 147324500.0, + "step": 5690 + }, + { + "epoch": 0.6249725455743466, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.1235077381134033, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6947246193885803, + "num_tokens": 147352353.0, + "step": 5691 + }, + { + "epoch": 0.6250823632769602, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.540987253189087, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7135127782821655, + "num_tokens": 147371328.0, + "step": 5692 + }, + { + "epoch": 0.6251921809795739, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.4131972789764404, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7193116545677185, + "num_tokens": 147390884.0, + "step": 5693 + }, + { + "epoch": 0.6253019986821876, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.5693557262420654, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7209411859512329, + "num_tokens": 147410305.0, + "step": 5694 + }, + { + "epoch": 0.6254118163848013, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.2305238246917725, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7055704593658447, + "num_tokens": 147434813.0, + "step": 5695 + }, + { + "epoch": 0.6255216340874149, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 1.9666049480438232, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.7004741430282593, + "num_tokens": 147466456.0, + "step": 5696 + }, + { + "epoch": 0.6256314517900285, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.6408638954162598, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7140908241271973, + "num_tokens": 147485784.0, + "step": 5697 + }, + { + "epoch": 0.6257412694926422, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.171172857284546, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7257628440856934, + "num_tokens": 147509484.0, + "step": 5698 + }, + { + "epoch": 0.6258510871952558, + "ewc_loss": 5.1975250244140625e-05, + "grad_norm": 2.4179744720458984, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.707223653793335, + "num_tokens": 147530235.0, + "step": 5699 + }, + { + "epoch": 0.6259609048978695, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2348265647888184, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6864417791366577, + "num_tokens": 147556317.0, + "step": 5700 + }, + { + "epoch": 0.6260707226004832, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.218763589859009, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.7088754773139954, + "num_tokens": 147580805.0, + "step": 5701 + }, + { + "epoch": 0.6261805403030969, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.262909173965454, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7149701118469238, + "num_tokens": 147605713.0, + "step": 5702 + }, + { + "epoch": 0.6262903580057105, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.206735134124756, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6945884823799133, + "num_tokens": 147632335.0, + "step": 5703 + }, + { + "epoch": 0.6264001757083242, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.0575990676879883, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.6996724605560303, + "num_tokens": 147658970.0, + "step": 5704 + }, + { + "epoch": 0.6265099934109378, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.147806406021118, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6999287605285645, + "num_tokens": 147686015.0, + "step": 5705 + }, + { + "epoch": 0.6266198111135515, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.4335036277770996, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7070999145507812, + "num_tokens": 147708507.0, + "step": 5706 + }, + { + "epoch": 0.6267296288161651, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.160006284713745, + "learning_rate": 1e-06, + "loss": 1.0402, + "mean_token_accuracy": 0.6825231313705444, + "num_tokens": 147736217.0, + "step": 5707 + }, + { + "epoch": 0.6268394465187789, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2828242778778076, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7034127712249756, + "num_tokens": 147760731.0, + "step": 5708 + }, + { + "epoch": 0.6269492642213925, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.1684625148773193, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.688879132270813, + "num_tokens": 147786980.0, + "step": 5709 + }, + { + "epoch": 0.6270590819240062, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2930750846862793, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7213371992111206, + "num_tokens": 147807755.0, + "step": 5710 + }, + { + "epoch": 0.6271688996266198, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.129589080810547, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6974177360534668, + "num_tokens": 147834274.0, + "step": 5711 + }, + { + "epoch": 0.6272787173292335, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.098477840423584, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7187511920928955, + "num_tokens": 147860854.0, + "step": 5712 + }, + { + "epoch": 0.6273885350318471, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2227370738983154, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6830880641937256, + "num_tokens": 147887369.0, + "step": 5713 + }, + { + "epoch": 0.6274983527344608, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.046584129333496, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7239886522293091, + "num_tokens": 147915582.0, + "step": 5714 + }, + { + "epoch": 0.6276081704370745, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.0757944583892822, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6929159164428711, + "num_tokens": 147943778.0, + "step": 5715 + }, + { + "epoch": 0.6277179881396882, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2394351959228516, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6938259601593018, + "num_tokens": 147967687.0, + "step": 5716 + }, + { + "epoch": 0.6278278058423018, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.097125291824341, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6978933215141296, + "num_tokens": 147996893.0, + "step": 5717 + }, + { + "epoch": 0.6279376235449154, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 1.933929681777954, + "learning_rate": 1e-06, + "loss": 1.0634, + "mean_token_accuracy": 0.6812057495117188, + "num_tokens": 148030636.0, + "step": 5718 + }, + { + "epoch": 0.6280474412475291, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 2.2365550994873047, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7122164368629456, + "num_tokens": 148054012.0, + "step": 5719 + }, + { + "epoch": 0.6281572589501427, + "ewc_loss": 5.221366882324219e-05, + "grad_norm": 1.980597734451294, + "learning_rate": 1e-06, + "loss": 1.0755, + "mean_token_accuracy": 0.6792458295822144, + "num_tokens": 148085309.0, + "step": 5720 + }, + { + "epoch": 0.6282670766527564, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.296604633331299, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.6953748464584351, + "num_tokens": 148109119.0, + "step": 5721 + }, + { + "epoch": 0.62837689435537, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.3914711475372314, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6976442337036133, + "num_tokens": 148132385.0, + "step": 5722 + }, + { + "epoch": 0.6284867120579838, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 1.8657443523406982, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7026553153991699, + "num_tokens": 148165336.0, + "step": 5723 + }, + { + "epoch": 0.6285965297605974, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.0035157203674316, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6993659734725952, + "num_tokens": 148195345.0, + "step": 5724 + }, + { + "epoch": 0.6287063474632111, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.293944835662842, + "learning_rate": 1e-06, + "loss": 1.0657, + "mean_token_accuracy": 0.6865960359573364, + "num_tokens": 148218406.0, + "step": 5725 + }, + { + "epoch": 0.6288161651658247, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 1.977932095527649, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7059370279312134, + "num_tokens": 148249170.0, + "step": 5726 + }, + { + "epoch": 0.6289259828684384, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.44459867477417, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7215392589569092, + "num_tokens": 148269680.0, + "step": 5727 + }, + { + "epoch": 0.629035800571052, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.37213397026062, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7252612709999084, + "num_tokens": 148291251.0, + "step": 5728 + }, + { + "epoch": 0.6291456182736657, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.1912295818328857, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7142011523246765, + "num_tokens": 148318381.0, + "step": 5729 + }, + { + "epoch": 0.6292554359762794, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.0711886882781982, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7128742933273315, + "num_tokens": 148345808.0, + "step": 5730 + }, + { + "epoch": 0.6293652536788931, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.392153263092041, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7242335677146912, + "num_tokens": 148367241.0, + "step": 5731 + }, + { + "epoch": 0.6294750713815067, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.2246592044830322, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6854246854782104, + "num_tokens": 148394134.0, + "step": 5732 + }, + { + "epoch": 0.6295848890841204, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.2168262004852295, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7137061357498169, + "num_tokens": 148418898.0, + "step": 5733 + }, + { + "epoch": 0.629694706786734, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.1914143562316895, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.7024936676025391, + "num_tokens": 148445447.0, + "step": 5734 + }, + { + "epoch": 0.6298045244893477, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.294642210006714, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6993308067321777, + "num_tokens": 148468281.0, + "step": 5735 + }, + { + "epoch": 0.6299143421919613, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.1740360260009766, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7081867456436157, + "num_tokens": 148494389.0, + "step": 5736 + }, + { + "epoch": 0.630024159894575, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.2229084968566895, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7055369019508362, + "num_tokens": 148521662.0, + "step": 5737 + }, + { + "epoch": 0.6301339775971887, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.549826145172119, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7169139385223389, + "num_tokens": 148541457.0, + "step": 5738 + }, + { + "epoch": 0.6302437952998023, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.134129047393799, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6880679726600647, + "num_tokens": 148569634.0, + "step": 5739 + }, + { + "epoch": 0.630353613002416, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 1.9741562604904175, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6992725729942322, + "num_tokens": 148600795.0, + "step": 5740 + }, + { + "epoch": 0.6304634307050296, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.023987054824829, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7135852575302124, + "num_tokens": 148629434.0, + "step": 5741 + }, + { + "epoch": 0.6305732484076433, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.052008628845215, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7053489089012146, + "num_tokens": 148658548.0, + "step": 5742 + }, + { + "epoch": 0.6306830661102569, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.148186445236206, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7135194540023804, + "num_tokens": 148685449.0, + "step": 5743 + }, + { + "epoch": 0.6307928838128707, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.2170956134796143, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.698119044303894, + "num_tokens": 148709970.0, + "step": 5744 + }, + { + "epoch": 0.6309027015154843, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.152649402618408, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7044646739959717, + "num_tokens": 148737062.0, + "step": 5745 + }, + { + "epoch": 0.631012519218098, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 1.974570393562317, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.705176055431366, + "num_tokens": 148767671.0, + "step": 5746 + }, + { + "epoch": 0.6311223369207116, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.0501160621643066, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.6764997243881226, + "num_tokens": 148798121.0, + "step": 5747 + }, + { + "epoch": 0.6312321546233253, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.18467378616333, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.744011402130127, + "num_tokens": 148822303.0, + "step": 5748 + }, + { + "epoch": 0.6313419723259389, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.250221014022827, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7081146240234375, + "num_tokens": 148846091.0, + "step": 5749 + }, + { + "epoch": 0.6314517900285526, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.031332015991211, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.7056936621665955, + "num_tokens": 148875284.0, + "step": 5750 + }, + { + "epoch": 0.6315616077311663, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.257612943649292, + "learning_rate": 1e-06, + "loss": 1.0722, + "mean_token_accuracy": 0.6927140355110168, + "num_tokens": 148903453.0, + "step": 5751 + }, + { + "epoch": 0.63167142543378, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.224435567855835, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7025312185287476, + "num_tokens": 148930804.0, + "step": 5752 + }, + { + "epoch": 0.6317812431363936, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.115222454071045, + "learning_rate": 1e-06, + "loss": 1.0632, + "mean_token_accuracy": 0.6818941831588745, + "num_tokens": 148961324.0, + "step": 5753 + }, + { + "epoch": 0.6318910608390073, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.079946994781494, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7077812552452087, + "num_tokens": 148989833.0, + "step": 5754 + }, + { + "epoch": 0.6320008785416209, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 1.9489299058914185, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.7004164457321167, + "num_tokens": 149021071.0, + "step": 5755 + }, + { + "epoch": 0.6321106962442345, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.1558315753936768, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7165971994400024, + "num_tokens": 149044826.0, + "step": 5756 + }, + { + "epoch": 0.6322205139468482, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.0389351844787598, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6988572478294373, + "num_tokens": 149071338.0, + "step": 5757 + }, + { + "epoch": 0.6323303316494618, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.069770574569702, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.708534836769104, + "num_tokens": 149100176.0, + "step": 5758 + }, + { + "epoch": 0.6324401493520756, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.230137586593628, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6916258931159973, + "num_tokens": 149126871.0, + "step": 5759 + }, + { + "epoch": 0.6325499670546892, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.452547550201416, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7034332156181335, + "num_tokens": 149149003.0, + "step": 5760 + }, + { + "epoch": 0.6326597847573029, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.493619203567505, + "learning_rate": 1e-06, + "loss": 1.046, + "mean_token_accuracy": 0.6808760166168213, + "num_tokens": 149172255.0, + "step": 5761 + }, + { + "epoch": 0.6327696024599165, + "ewc_loss": 5.269050598144531e-05, + "grad_norm": 2.338820695877075, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.694343626499176, + "num_tokens": 149194487.0, + "step": 5762 + }, + { + "epoch": 0.6328794201625302, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.176185369491577, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7031266689300537, + "num_tokens": 149219638.0, + "step": 5763 + }, + { + "epoch": 0.6329892378651438, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.706397771835327, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.6972064971923828, + "num_tokens": 149238091.0, + "step": 5764 + }, + { + "epoch": 0.6330990555677575, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.256488561630249, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7058933973312378, + "num_tokens": 149261916.0, + "step": 5765 + }, + { + "epoch": 0.6332088732703712, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.330538034439087, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7253235578536987, + "num_tokens": 149283449.0, + "step": 5766 + }, + { + "epoch": 0.6333186909729849, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.129068613052368, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.709360659122467, + "num_tokens": 149309221.0, + "step": 5767 + }, + { + "epoch": 0.6334285086755985, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.004479169845581, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6919918060302734, + "num_tokens": 149338800.0, + "step": 5768 + }, + { + "epoch": 0.6335383263782122, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.119418144226074, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7046294212341309, + "num_tokens": 149367427.0, + "step": 5769 + }, + { + "epoch": 0.6336481440808258, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.4710755348205566, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7071269750595093, + "num_tokens": 149391851.0, + "step": 5770 + }, + { + "epoch": 0.6337579617834395, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.3379576206207275, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7131344079971313, + "num_tokens": 149415251.0, + "step": 5771 + }, + { + "epoch": 0.6338677794860531, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.0564229488372803, + "learning_rate": 1e-06, + "loss": 1.0461, + "mean_token_accuracy": 0.6847431659698486, + "num_tokens": 149444650.0, + "step": 5772 + }, + { + "epoch": 0.6339775971886669, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 1.9433492422103882, + "learning_rate": 1e-06, + "loss": 1.0798, + "mean_token_accuracy": 0.6740125417709351, + "num_tokens": 149475849.0, + "step": 5773 + }, + { + "epoch": 0.6340874148912805, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.0761122703552246, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6932984590530396, + "num_tokens": 149503813.0, + "step": 5774 + }, + { + "epoch": 0.6341972325938942, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.2834153175354004, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6971768140792847, + "num_tokens": 149527726.0, + "step": 5775 + }, + { + "epoch": 0.6343070502965078, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 1.9570409059524536, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6937182545661926, + "num_tokens": 149560705.0, + "step": 5776 + }, + { + "epoch": 0.6344168679991214, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.02840256690979, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7052644491195679, + "num_tokens": 149587557.0, + "step": 5777 + }, + { + "epoch": 0.6345266857017351, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.1393184661865234, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6972044706344604, + "num_tokens": 149613308.0, + "step": 5778 + }, + { + "epoch": 0.6346365034043487, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.2606122493743896, + "learning_rate": 1e-06, + "loss": 1.0637, + "mean_token_accuracy": 0.6749433279037476, + "num_tokens": 149639459.0, + "step": 5779 + }, + { + "epoch": 0.6347463211069625, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.0424387454986572, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7016414403915405, + "num_tokens": 149669146.0, + "step": 5780 + }, + { + "epoch": 0.6348561388095761, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.1162354946136475, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7093281745910645, + "num_tokens": 149695518.0, + "step": 5781 + }, + { + "epoch": 0.6349659565121898, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.421005964279175, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7161080241203308, + "num_tokens": 149718383.0, + "step": 5782 + }, + { + "epoch": 0.6350757742148034, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.354386568069458, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7055661082267761, + "num_tokens": 149741179.0, + "step": 5783 + }, + { + "epoch": 0.6351855919174171, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.386122703552246, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6972888708114624, + "num_tokens": 149763981.0, + "step": 5784 + }, + { + "epoch": 0.6352954096200307, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 1.9890109300613403, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.6976497769355774, + "num_tokens": 149794146.0, + "step": 5785 + }, + { + "epoch": 0.6354052273226444, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.2365691661834717, + "learning_rate": 1e-06, + "loss": 1.0562, + "mean_token_accuracy": 0.6919013261795044, + "num_tokens": 149819811.0, + "step": 5786 + }, + { + "epoch": 0.635515045025258, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.0600154399871826, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.6950613260269165, + "num_tokens": 149848238.0, + "step": 5787 + }, + { + "epoch": 0.6356248627278718, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.2150955200195312, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7064225673675537, + "num_tokens": 149874801.0, + "step": 5788 + }, + { + "epoch": 0.6357346804304854, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.4179317951202393, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7111541628837585, + "num_tokens": 149897048.0, + "step": 5789 + }, + { + "epoch": 0.6358444981330991, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.3901381492614746, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6904465556144714, + "num_tokens": 149919727.0, + "step": 5790 + }, + { + "epoch": 0.6359543158357127, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.136669874191284, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.704526424407959, + "num_tokens": 149945061.0, + "step": 5791 + }, + { + "epoch": 0.6360641335383264, + "ewc_loss": 5.316734313964844e-05, + "grad_norm": 2.4704933166503906, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7187325358390808, + "num_tokens": 149966169.0, + "step": 5792 + }, + { + "epoch": 0.63617395124094, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.105196475982666, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7229292392730713, + "num_tokens": 149992607.0, + "step": 5793 + }, + { + "epoch": 0.6362837689435537, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.206644296646118, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.6852338314056396, + "num_tokens": 150019636.0, + "step": 5794 + }, + { + "epoch": 0.6363935866461674, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.1075518131256104, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7001364827156067, + "num_tokens": 150046268.0, + "step": 5795 + }, + { + "epoch": 0.636503404348781, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.1569576263427734, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6903256773948669, + "num_tokens": 150072283.0, + "step": 5796 + }, + { + "epoch": 0.6366132220513947, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.1177725791931152, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.7028976678848267, + "num_tokens": 150099010.0, + "step": 5797 + }, + { + "epoch": 0.6367230397540083, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.117274761199951, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7098515033721924, + "num_tokens": 150124807.0, + "step": 5798 + }, + { + "epoch": 0.636832857456622, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.104846954345703, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7096966505050659, + "num_tokens": 150153282.0, + "step": 5799 + }, + { + "epoch": 0.6369426751592356, + "ewc_loss": 5.340576171875e-05, + "grad_norm": 2.3940560817718506, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6942258477210999, + "num_tokens": 150175704.0, + "step": 5800 + }, + { + "epoch": 0.6370524928618493, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.1700339317321777, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7148720026016235, + "num_tokens": 150201266.0, + "step": 5801 + }, + { + "epoch": 0.637162310564463, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.243993043899536, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7141621112823486, + "num_tokens": 150223486.0, + "step": 5802 + }, + { + "epoch": 0.6372721282670767, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.1964168548583984, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7049649953842163, + "num_tokens": 150249127.0, + "step": 5803 + }, + { + "epoch": 0.6373819459696903, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.274491548538208, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.721807599067688, + "num_tokens": 150272652.0, + "step": 5804 + }, + { + "epoch": 0.637491763672304, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.3224029541015625, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7169942855834961, + "num_tokens": 150296163.0, + "step": 5805 + }, + { + "epoch": 0.6376015813749176, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.0740861892700195, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.692439615726471, + "num_tokens": 150326248.0, + "step": 5806 + }, + { + "epoch": 0.6377113990775313, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.1629281044006348, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7151163220405579, + "num_tokens": 150350743.0, + "step": 5807 + }, + { + "epoch": 0.6378212167801449, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.271389961242676, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7249008417129517, + "num_tokens": 150374351.0, + "step": 5808 + }, + { + "epoch": 0.6379310344827587, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.004757881164551, + "learning_rate": 1e-06, + "loss": 1.0455, + "mean_token_accuracy": 0.6835986375808716, + "num_tokens": 150403427.0, + "step": 5809 + }, + { + "epoch": 0.6380408521853723, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.3107433319091797, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7187318205833435, + "num_tokens": 150425915.0, + "step": 5810 + }, + { + "epoch": 0.638150669887986, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.320335865020752, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7131119966506958, + "num_tokens": 150448372.0, + "step": 5811 + }, + { + "epoch": 0.6382604875905996, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.2975642681121826, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7212974429130554, + "num_tokens": 150470554.0, + "step": 5812 + }, + { + "epoch": 0.6383703052932133, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.3610565662384033, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6850875616073608, + "num_tokens": 150493751.0, + "step": 5813 + }, + { + "epoch": 0.6384801229958269, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.264570474624634, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6917885541915894, + "num_tokens": 150518023.0, + "step": 5814 + }, + { + "epoch": 0.6385899406984406, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 1.90707528591156, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7221362590789795, + "num_tokens": 150549241.0, + "step": 5815 + }, + { + "epoch": 0.6386997584010542, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.026146173477173, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.7052287459373474, + "num_tokens": 150580556.0, + "step": 5816 + }, + { + "epoch": 0.638809576103668, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.0447514057159424, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6916280388832092, + "num_tokens": 150611486.0, + "step": 5817 + }, + { + "epoch": 0.6389193938062816, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.2504727840423584, + "learning_rate": 1e-06, + "loss": 1.0628, + "mean_token_accuracy": 0.6766292452812195, + "num_tokens": 150636022.0, + "step": 5818 + }, + { + "epoch": 0.6390292115088952, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.2969181537628174, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.69803386926651, + "num_tokens": 150659070.0, + "step": 5819 + }, + { + "epoch": 0.6391390292115089, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.143064260482788, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7296615839004517, + "num_tokens": 150685946.0, + "step": 5820 + }, + { + "epoch": 0.6392488469141225, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.2508060932159424, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7139261364936829, + "num_tokens": 150710552.0, + "step": 5821 + }, + { + "epoch": 0.6393586646167362, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.182544231414795, + "learning_rate": 1e-06, + "loss": 1.0761, + "mean_token_accuracy": 0.6865947842597961, + "num_tokens": 150739809.0, + "step": 5822 + }, + { + "epoch": 0.6394684823193498, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.4624314308166504, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7084642648696899, + "num_tokens": 150761127.0, + "step": 5823 + }, + { + "epoch": 0.6395783000219636, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.099923849105835, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7599598169326782, + "num_tokens": 150786097.0, + "step": 5824 + }, + { + "epoch": 0.6396881177245772, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.4144012928009033, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6938173770904541, + "num_tokens": 150807894.0, + "step": 5825 + }, + { + "epoch": 0.6397979354271909, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.283975124359131, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7072099447250366, + "num_tokens": 150831323.0, + "step": 5826 + }, + { + "epoch": 0.6399077531298045, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.121286630630493, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6979882717132568, + "num_tokens": 150858194.0, + "step": 5827 + }, + { + "epoch": 0.6400175708324182, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.15374755859375, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7087085843086243, + "num_tokens": 150882564.0, + "step": 5828 + }, + { + "epoch": 0.6401273885350318, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.200814723968506, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.725068211555481, + "num_tokens": 150907187.0, + "step": 5829 + }, + { + "epoch": 0.6402372062376455, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.1568214893341064, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6874552965164185, + "num_tokens": 150934895.0, + "step": 5830 + }, + { + "epoch": 0.6403470239402592, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.288738250732422, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6964584589004517, + "num_tokens": 150958210.0, + "step": 5831 + }, + { + "epoch": 0.6404568416428729, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2883613109588623, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7348300218582153, + "num_tokens": 150980567.0, + "step": 5832 + }, + { + "epoch": 0.6405666593454865, + "ewc_loss": 5.364418029785156e-05, + "grad_norm": 2.048004150390625, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6935856938362122, + "num_tokens": 151009372.0, + "step": 5833 + }, + { + "epoch": 0.6406764770481002, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.216683864593506, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6964302062988281, + "num_tokens": 151035712.0, + "step": 5834 + }, + { + "epoch": 0.6407862947507138, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1170828342437744, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7237790822982788, + "num_tokens": 151061318.0, + "step": 5835 + }, + { + "epoch": 0.6408961124533274, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1365809440612793, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7038203477859497, + "num_tokens": 151087793.0, + "step": 5836 + }, + { + "epoch": 0.6410059301559411, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.3041179180145264, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6941413879394531, + "num_tokens": 151111689.0, + "step": 5837 + }, + { + "epoch": 0.6411157478585549, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.9612925052642822, + "learning_rate": 1e-06, + "loss": 1.0546, + "mean_token_accuracy": 0.6888910531997681, + "num_tokens": 151144752.0, + "step": 5838 + }, + { + "epoch": 0.6412255655611685, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.002556324005127, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7102281451225281, + "num_tokens": 151175374.0, + "step": 5839 + }, + { + "epoch": 0.6413353832637821, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.9053032398223877, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6977142691612244, + "num_tokens": 151207581.0, + "step": 5840 + }, + { + "epoch": 0.6414452009663958, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1588573455810547, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7230280637741089, + "num_tokens": 151231272.0, + "step": 5841 + }, + { + "epoch": 0.6415550186690094, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.030479669570923, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6968784332275391, + "num_tokens": 151259848.0, + "step": 5842 + }, + { + "epoch": 0.6416648363716231, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2933905124664307, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7165434956550598, + "num_tokens": 151283535.0, + "step": 5843 + }, + { + "epoch": 0.6417746540742367, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.325465202331543, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7174220085144043, + "num_tokens": 151307785.0, + "step": 5844 + }, + { + "epoch": 0.6418844717768504, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.223083972930908, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6899026036262512, + "num_tokens": 151334314.0, + "step": 5845 + }, + { + "epoch": 0.6419942894794641, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.474684953689575, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7182841897010803, + "num_tokens": 151355616.0, + "step": 5846 + }, + { + "epoch": 0.6421041071820778, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2717154026031494, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7285805940628052, + "num_tokens": 151377054.0, + "step": 5847 + }, + { + "epoch": 0.6422139248846914, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1799635887145996, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7110546827316284, + "num_tokens": 151402658.0, + "step": 5848 + }, + { + "epoch": 0.6423237425873051, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1817333698272705, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7048720121383667, + "num_tokens": 151429171.0, + "step": 5849 + }, + { + "epoch": 0.6424335602899187, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1129236221313477, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7039690613746643, + "num_tokens": 151458553.0, + "step": 5850 + }, + { + "epoch": 0.6425433779925324, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2796812057495117, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7181774973869324, + "num_tokens": 151481152.0, + "step": 5851 + }, + { + "epoch": 0.642653195695146, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.944836974143982, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7072749137878418, + "num_tokens": 151511318.0, + "step": 5852 + }, + { + "epoch": 0.6427630133977598, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2639689445495605, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7128445506095886, + "num_tokens": 151534627.0, + "step": 5853 + }, + { + "epoch": 0.6428728311003734, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.21728515625, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6984845399856567, + "num_tokens": 151560458.0, + "step": 5854 + }, + { + "epoch": 0.6429826488029871, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.354032278060913, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.702735424041748, + "num_tokens": 151582610.0, + "step": 5855 + }, + { + "epoch": 0.6430924665056007, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2329578399658203, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.719295859336853, + "num_tokens": 151608983.0, + "step": 5856 + }, + { + "epoch": 0.6432022842082143, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.5662901401519775, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7236505746841431, + "num_tokens": 151627493.0, + "step": 5857 + }, + { + "epoch": 0.643312101910828, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.7843762636184692, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7415066957473755, + "num_tokens": 151661357.0, + "step": 5858 + }, + { + "epoch": 0.6434219196134416, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.892016053199768, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6975265741348267, + "num_tokens": 151696267.0, + "step": 5859 + }, + { + "epoch": 0.6435317373160554, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0372767448425293, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7030613422393799, + "num_tokens": 151724986.0, + "step": 5860 + }, + { + "epoch": 0.643641555018669, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0477380752563477, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7209022045135498, + "num_tokens": 151752918.0, + "step": 5861 + }, + { + "epoch": 0.6437513727212827, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1609201431274414, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7236660718917847, + "num_tokens": 151777243.0, + "step": 5862 + }, + { + "epoch": 0.6438611904238963, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1735031604766846, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6995912790298462, + "num_tokens": 151805063.0, + "step": 5863 + }, + { + "epoch": 0.64397100812651, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.225541591644287, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7108522653579712, + "num_tokens": 151831224.0, + "step": 5864 + }, + { + "epoch": 0.6440808258291236, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.016022205352783, + "learning_rate": 1e-06, + "loss": 1.0451, + "mean_token_accuracy": 0.6807587146759033, + "num_tokens": 151862421.0, + "step": 5865 + }, + { + "epoch": 0.6441906435317373, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.287093162536621, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6980817914009094, + "num_tokens": 151888055.0, + "step": 5866 + }, + { + "epoch": 0.644300461234351, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.4594013690948486, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7290032505989075, + "num_tokens": 151907627.0, + "step": 5867 + }, + { + "epoch": 0.6444102789369647, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.249690294265747, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7056257724761963, + "num_tokens": 151931472.0, + "step": 5868 + }, + { + "epoch": 0.6445200966395783, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.19280743598938, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7154879570007324, + "num_tokens": 151958960.0, + "step": 5869 + }, + { + "epoch": 0.644629914342192, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0349676609039307, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7102237939834595, + "num_tokens": 151986428.0, + "step": 5870 + }, + { + "epoch": 0.6447397320448056, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.3314707279205322, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.707834780216217, + "num_tokens": 152009759.0, + "step": 5871 + }, + { + "epoch": 0.6448495497474193, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0407304763793945, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.6868426203727722, + "num_tokens": 152039419.0, + "step": 5872 + }, + { + "epoch": 0.6449593674500329, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.3000197410583496, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7106781005859375, + "num_tokens": 152061413.0, + "step": 5873 + }, + { + "epoch": 0.6450691851526466, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.23683500289917, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6872217655181885, + "num_tokens": 152086909.0, + "step": 5874 + }, + { + "epoch": 0.6451790028552603, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.31438946723938, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7120916247367859, + "num_tokens": 152110852.0, + "step": 5875 + }, + { + "epoch": 0.645288820557874, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1573104858398438, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.690413236618042, + "num_tokens": 152139170.0, + "step": 5876 + }, + { + "epoch": 0.6453986382604876, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.377898693084717, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6960039138793945, + "num_tokens": 152162709.0, + "step": 5877 + }, + { + "epoch": 0.6455084559631012, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.196190118789673, + "learning_rate": 1e-06, + "loss": 1.0273, + "mean_token_accuracy": 0.695699155330658, + "num_tokens": 152190478.0, + "step": 5878 + }, + { + "epoch": 0.6456182736657149, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1266419887542725, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7016522288322449, + "num_tokens": 152218556.0, + "step": 5879 + }, + { + "epoch": 0.6457280913683285, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.269364833831787, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7104003429412842, + "num_tokens": 152242928.0, + "step": 5880 + }, + { + "epoch": 0.6458379090709422, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.3092923164367676, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7106884717941284, + "num_tokens": 152266912.0, + "step": 5881 + }, + { + "epoch": 0.6459477267735559, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.9761284589767456, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.6980417966842651, + "num_tokens": 152297859.0, + "step": 5882 + }, + { + "epoch": 0.6460575444761696, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.3872711658477783, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7086939215660095, + "num_tokens": 152320490.0, + "step": 5883 + }, + { + "epoch": 0.6461673621787832, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1518137454986572, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6898092031478882, + "num_tokens": 152346999.0, + "step": 5884 + }, + { + "epoch": 0.6462771798813969, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.094238519668579, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7058478593826294, + "num_tokens": 152377770.0, + "step": 5885 + }, + { + "epoch": 0.6463869975840105, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0660128593444824, + "learning_rate": 1e-06, + "loss": 1.0481, + "mean_token_accuracy": 0.6848558783531189, + "num_tokens": 152408079.0, + "step": 5886 + }, + { + "epoch": 0.6464968152866242, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.349341869354248, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6981087923049927, + "num_tokens": 152430651.0, + "step": 5887 + }, + { + "epoch": 0.6466066329892378, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0985469818115234, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6939389705657959, + "num_tokens": 152459557.0, + "step": 5888 + }, + { + "epoch": 0.6467164506918516, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.6345458030700684, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7320644855499268, + "num_tokens": 152478085.0, + "step": 5889 + }, + { + "epoch": 0.6468262683944652, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.127091884613037, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7250496745109558, + "num_tokens": 152506522.0, + "step": 5890 + }, + { + "epoch": 0.6469360860970789, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0864624977111816, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6803343296051025, + "num_tokens": 152535998.0, + "step": 5891 + }, + { + "epoch": 0.6470459037996925, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1864173412323, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6921619176864624, + "num_tokens": 152562365.0, + "step": 5892 + }, + { + "epoch": 0.6471557215023062, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0045087337493896, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.686711311340332, + "num_tokens": 152591858.0, + "step": 5893 + }, + { + "epoch": 0.6472655392049198, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.25388503074646, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7293040752410889, + "num_tokens": 152616172.0, + "step": 5894 + }, + { + "epoch": 0.6473753569075335, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.2878615856170654, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6986356973648071, + "num_tokens": 152639601.0, + "step": 5895 + }, + { + "epoch": 0.6474851746101472, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1646668910980225, + "learning_rate": 1e-06, + "loss": 1.0628, + "mean_token_accuracy": 0.6761893630027771, + "num_tokens": 152667239.0, + "step": 5896 + }, + { + "epoch": 0.6475949923127609, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.004091262817383, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6868818402290344, + "num_tokens": 152697231.0, + "step": 5897 + }, + { + "epoch": 0.6477048100153745, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.047717571258545, + "learning_rate": 1e-06, + "loss": 1.0293, + "mean_token_accuracy": 0.6952304840087891, + "num_tokens": 152728080.0, + "step": 5898 + }, + { + "epoch": 0.6478146277179881, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.8413159847259521, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6855560541152954, + "num_tokens": 152762345.0, + "step": 5899 + }, + { + "epoch": 0.6479244454206018, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.0893287658691406, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6959670186042786, + "num_tokens": 152789650.0, + "step": 5900 + }, + { + "epoch": 0.6480342631232154, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 1.991503357887268, + "learning_rate": 1e-06, + "loss": 1.0875, + "mean_token_accuracy": 0.6795666813850403, + "num_tokens": 152821126.0, + "step": 5901 + }, + { + "epoch": 0.6481440808258291, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.277923345565796, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6993958950042725, + "num_tokens": 152846609.0, + "step": 5902 + }, + { + "epoch": 0.6482538985284428, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.108025312423706, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7004388570785522, + "num_tokens": 152872815.0, + "step": 5903 + }, + { + "epoch": 0.6483637162310565, + "ewc_loss": 5.3882598876953125e-05, + "grad_norm": 2.1053221225738525, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6911680698394775, + "num_tokens": 152900870.0, + "step": 5904 + }, + { + "epoch": 0.6484735339336701, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.138583183288574, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7212190628051758, + "num_tokens": 152926475.0, + "step": 5905 + }, + { + "epoch": 0.6485833516362838, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 1.9315454959869385, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6925972700119019, + "num_tokens": 152957867.0, + "step": 5906 + }, + { + "epoch": 0.6486931693388974, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.198625087738037, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7152325510978699, + "num_tokens": 152982578.0, + "step": 5907 + }, + { + "epoch": 0.6488029870415111, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.268944025039673, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7166765332221985, + "num_tokens": 153006120.0, + "step": 5908 + }, + { + "epoch": 0.6489128047441247, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.121666431427002, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7288237810134888, + "num_tokens": 153032777.0, + "step": 5909 + }, + { + "epoch": 0.6490226224467384, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.2029712200164795, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6888420581817627, + "num_tokens": 153061135.0, + "step": 5910 + }, + { + "epoch": 0.6491324401493521, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.0088822841644287, + "learning_rate": 1e-06, + "loss": 1.0998, + "mean_token_accuracy": 0.6620699167251587, + "num_tokens": 153094626.0, + "step": 5911 + }, + { + "epoch": 0.6492422578519658, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.334890842437744, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7023588418960571, + "num_tokens": 153120192.0, + "step": 5912 + }, + { + "epoch": 0.6493520755545794, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.153947114944458, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7077111005783081, + "num_tokens": 153146784.0, + "step": 5913 + }, + { + "epoch": 0.6494618932571931, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.2203569412231445, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.72135329246521, + "num_tokens": 153170056.0, + "step": 5914 + }, + { + "epoch": 0.6495717109598067, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.153554677963257, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7207609415054321, + "num_tokens": 153195030.0, + "step": 5915 + }, + { + "epoch": 0.6496815286624203, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.121760129928589, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7158170938491821, + "num_tokens": 153221898.0, + "step": 5916 + }, + { + "epoch": 0.649791346365034, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 1.9746640920639038, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7010424137115479, + "num_tokens": 153255269.0, + "step": 5917 + }, + { + "epoch": 0.6499011640676478, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.1182727813720703, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7219077944755554, + "num_tokens": 153281596.0, + "step": 5918 + }, + { + "epoch": 0.6500109817702614, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.5879359245300293, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7191311120986938, + "num_tokens": 153299996.0, + "step": 5919 + }, + { + "epoch": 0.650120799472875, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.258380889892578, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7064167857170105, + "num_tokens": 153324129.0, + "step": 5920 + }, + { + "epoch": 0.6502306171754887, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.249662160873413, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6993094682693481, + "num_tokens": 153348145.0, + "step": 5921 + }, + { + "epoch": 0.6503404348781023, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.2704267501831055, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7126368284225464, + "num_tokens": 153371737.0, + "step": 5922 + }, + { + "epoch": 0.650450252580716, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.0188181400299072, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.701886773109436, + "num_tokens": 153402899.0, + "step": 5923 + }, + { + "epoch": 0.6505600702833296, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.0678865909576416, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7309554815292358, + "num_tokens": 153428385.0, + "step": 5924 + }, + { + "epoch": 0.6506698879859434, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 1.8921902179718018, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6849005818367004, + "num_tokens": 153461679.0, + "step": 5925 + }, + { + "epoch": 0.650779705688557, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.124551296234131, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6999759674072266, + "num_tokens": 153491875.0, + "step": 5926 + }, + { + "epoch": 0.6508895233911707, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.4779577255249023, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6990190744400024, + "num_tokens": 153512525.0, + "step": 5927 + }, + { + "epoch": 0.6509993410937843, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.243659257888794, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7058777213096619, + "num_tokens": 153537574.0, + "step": 5928 + }, + { + "epoch": 0.651109158796398, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.285879611968994, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7132106423377991, + "num_tokens": 153561400.0, + "step": 5929 + }, + { + "epoch": 0.6512189764990116, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.2111480236053467, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6928458213806152, + "num_tokens": 153588151.0, + "step": 5930 + }, + { + "epoch": 0.6513287942016253, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.1197733879089355, + "learning_rate": 1e-06, + "loss": 1.1116, + "mean_token_accuracy": 0.669730544090271, + "num_tokens": 153619776.0, + "step": 5931 + }, + { + "epoch": 0.651438611904239, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.5885603427886963, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7055542469024658, + "num_tokens": 153639527.0, + "step": 5932 + }, + { + "epoch": 0.6515484296068527, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 1.9965064525604248, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7252864241600037, + "num_tokens": 153667505.0, + "step": 5933 + }, + { + "epoch": 0.6516582473094663, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.10697865486145, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.6969729661941528, + "num_tokens": 153695390.0, + "step": 5934 + }, + { + "epoch": 0.65176806501208, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.4199235439300537, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7153834104537964, + "num_tokens": 153717638.0, + "step": 5935 + }, + { + "epoch": 0.6518778827146936, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.3624486923217773, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7144283056259155, + "num_tokens": 153738948.0, + "step": 5936 + }, + { + "epoch": 0.6519877004173072, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.6110951900482178, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7303445339202881, + "num_tokens": 153755757.0, + "step": 5937 + }, + { + "epoch": 0.6520975181199209, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.12137508392334, + "learning_rate": 1e-06, + "loss": 1.0481, + "mean_token_accuracy": 0.6843878626823425, + "num_tokens": 153782883.0, + "step": 5938 + }, + { + "epoch": 0.6522073358225345, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.1476333141326904, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7181313633918762, + "num_tokens": 153809530.0, + "step": 5939 + }, + { + "epoch": 0.6523171535251483, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.287687063217163, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7353094816207886, + "num_tokens": 153833445.0, + "step": 5940 + }, + { + "epoch": 0.6524269712277619, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.045384645462036, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6942217350006104, + "num_tokens": 153863494.0, + "step": 5941 + }, + { + "epoch": 0.6525367889303756, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.1753616333007812, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7027865052223206, + "num_tokens": 153889125.0, + "step": 5942 + }, + { + "epoch": 0.6526466066329892, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.384544849395752, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7020437717437744, + "num_tokens": 153911542.0, + "step": 5943 + }, + { + "epoch": 0.6527564243356029, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 1.9664545059204102, + "learning_rate": 1e-06, + "loss": 1.0653, + "mean_token_accuracy": 0.6878362894058228, + "num_tokens": 153942533.0, + "step": 5944 + }, + { + "epoch": 0.6528662420382165, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.231109619140625, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6952261328697205, + "num_tokens": 153967534.0, + "step": 5945 + }, + { + "epoch": 0.6529760597408302, + "ewc_loss": 5.412101745605469e-05, + "grad_norm": 2.0073657035827637, + "learning_rate": 1e-06, + "loss": 1.0861, + "mean_token_accuracy": 0.6764075756072998, + "num_tokens": 153998781.0, + "step": 5946 + }, + { + "epoch": 0.6530858774434439, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.553762674331665, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.6982786059379578, + "num_tokens": 154019262.0, + "step": 5947 + }, + { + "epoch": 0.6531956951460576, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.0985875129699707, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.6997138261795044, + "num_tokens": 154047477.0, + "step": 5948 + }, + { + "epoch": 0.6533055128486712, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.257681131362915, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.721006453037262, + "num_tokens": 154070596.0, + "step": 5949 + }, + { + "epoch": 0.6534153305512849, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.885310649871826, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7155213356018066, + "num_tokens": 154088080.0, + "step": 5950 + }, + { + "epoch": 0.6535251482538985, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.282536268234253, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7191059589385986, + "num_tokens": 154111183.0, + "step": 5951 + }, + { + "epoch": 0.6536349659565122, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.303673028945923, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7055203318595886, + "num_tokens": 154133988.0, + "step": 5952 + }, + { + "epoch": 0.6537447836591258, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.003861665725708, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.706907331943512, + "num_tokens": 154162949.0, + "step": 5953 + }, + { + "epoch": 0.6538546013617396, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.058870553970337, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7036887407302856, + "num_tokens": 154191471.0, + "step": 5954 + }, + { + "epoch": 0.6539644190643532, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.2372305393218994, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7085140943527222, + "num_tokens": 154214934.0, + "step": 5955 + }, + { + "epoch": 0.6540742367669669, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.048032283782959, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7010441422462463, + "num_tokens": 154242775.0, + "step": 5956 + }, + { + "epoch": 0.6541840544695805, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.0936365127563477, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6843831539154053, + "num_tokens": 154271659.0, + "step": 5957 + }, + { + "epoch": 0.6542938721721941, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.4940414428710938, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7357579469680786, + "num_tokens": 154291248.0, + "step": 5958 + }, + { + "epoch": 0.6544036898748078, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.0218284130096436, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7004343271255493, + "num_tokens": 154319496.0, + "step": 5959 + }, + { + "epoch": 0.6545135075774214, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.0065345764160156, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7087026238441467, + "num_tokens": 154348439.0, + "step": 5960 + }, + { + "epoch": 0.6546233252800352, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.2339398860931396, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7289230823516846, + "num_tokens": 154372375.0, + "step": 5961 + }, + { + "epoch": 0.6547331429826488, + "ewc_loss": 5.435943603515625e-05, + "grad_norm": 2.111260175704956, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7174810171127319, + "num_tokens": 154397081.0, + "step": 5962 + }, + { + "epoch": 0.6548429606852625, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.181065797805786, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.6995131969451904, + "num_tokens": 154425629.0, + "step": 5963 + }, + { + "epoch": 0.6549527783878761, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.4429984092712402, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7284402847290039, + "num_tokens": 154447369.0, + "step": 5964 + }, + { + "epoch": 0.6550625960904898, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.4756996631622314, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6998937129974365, + "num_tokens": 154469352.0, + "step": 5965 + }, + { + "epoch": 0.6551724137931034, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.153963327407837, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6916351318359375, + "num_tokens": 154498264.0, + "step": 5966 + }, + { + "epoch": 0.6552822314957171, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.4105188846588135, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.731037437915802, + "num_tokens": 154518816.0, + "step": 5967 + }, + { + "epoch": 0.6553920491983307, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 1.9227718114852905, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.6827428340911865, + "num_tokens": 154553216.0, + "step": 5968 + }, + { + "epoch": 0.6555018669009445, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.367575168609619, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7358380556106567, + "num_tokens": 154573981.0, + "step": 5969 + }, + { + "epoch": 0.6556116846035581, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.076427698135376, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7005940675735474, + "num_tokens": 154602589.0, + "step": 5970 + }, + { + "epoch": 0.6557215023061718, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.6148815155029297, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7062759399414062, + "num_tokens": 154621052.0, + "step": 5971 + }, + { + "epoch": 0.6558313200087854, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.2572498321533203, + "learning_rate": 1e-06, + "loss": 1.0442, + "mean_token_accuracy": 0.6868844032287598, + "num_tokens": 154647699.0, + "step": 5972 + }, + { + "epoch": 0.6559411377113991, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.3119091987609863, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7199167013168335, + "num_tokens": 154669577.0, + "step": 5973 + }, + { + "epoch": 0.6560509554140127, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.2206623554229736, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7136867642402649, + "num_tokens": 154695662.0, + "step": 5974 + }, + { + "epoch": 0.6561607731166264, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.123189687728882, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6931338906288147, + "num_tokens": 154723545.0, + "step": 5975 + }, + { + "epoch": 0.6562705908192401, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0031774044036865, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6879793405532837, + "num_tokens": 154755534.0, + "step": 5976 + }, + { + "epoch": 0.6563804085218538, + "ewc_loss": 5.459785461425781e-05, + "grad_norm": 2.1882166862487793, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6980932354927063, + "num_tokens": 154781348.0, + "step": 5977 + }, + { + "epoch": 0.6564902262244674, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0956788063049316, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6959304809570312, + "num_tokens": 154810297.0, + "step": 5978 + }, + { + "epoch": 0.656600043927081, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.415196418762207, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7121454477310181, + "num_tokens": 154831098.0, + "step": 5979 + }, + { + "epoch": 0.6567098616296947, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.1765642166137695, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6821893453598022, + "num_tokens": 154857877.0, + "step": 5980 + }, + { + "epoch": 0.6568196793323083, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2314536571502686, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7131208777427673, + "num_tokens": 154883350.0, + "step": 5981 + }, + { + "epoch": 0.656929497034922, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.1407554149627686, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6942641735076904, + "num_tokens": 154909297.0, + "step": 5982 + }, + { + "epoch": 0.6570393147375357, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0151422023773193, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7114633321762085, + "num_tokens": 154936292.0, + "step": 5983 + }, + { + "epoch": 0.6571491324401494, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 1.9837470054626465, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.6845953464508057, + "num_tokens": 154970211.0, + "step": 5984 + }, + { + "epoch": 0.657258950142763, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.530879497528076, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7198998928070068, + "num_tokens": 154989482.0, + "step": 5985 + }, + { + "epoch": 0.6573687678453767, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0978879928588867, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7070010304450989, + "num_tokens": 155017921.0, + "step": 5986 + }, + { + "epoch": 0.6574785855479903, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2370493412017822, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6904904842376709, + "num_tokens": 155042883.0, + "step": 5987 + }, + { + "epoch": 0.657588403250604, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0784947872161865, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6933490037918091, + "num_tokens": 155071816.0, + "step": 5988 + }, + { + "epoch": 0.6576982209532176, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2581987380981445, + "learning_rate": 1e-06, + "loss": 1.049, + "mean_token_accuracy": 0.6953182816505432, + "num_tokens": 155098747.0, + "step": 5989 + }, + { + "epoch": 0.6578080386558314, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.4868314266204834, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7198952436447144, + "num_tokens": 155119942.0, + "step": 5990 + }, + { + "epoch": 0.657917856358445, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0977771282196045, + "learning_rate": 1e-06, + "loss": 1.054, + "mean_token_accuracy": 0.6844077110290527, + "num_tokens": 155146995.0, + "step": 5991 + }, + { + "epoch": 0.6580276740610587, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 1.9933191537857056, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6912567615509033, + "num_tokens": 155179232.0, + "step": 5992 + }, + { + "epoch": 0.6581374917636723, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.3293988704681396, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7017672061920166, + "num_tokens": 155202130.0, + "step": 5993 + }, + { + "epoch": 0.658247309466286, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2772557735443115, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.6884349584579468, + "num_tokens": 155227196.0, + "step": 5994 + }, + { + "epoch": 0.6583571271688996, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2447283267974854, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.6886644959449768, + "num_tokens": 155254350.0, + "step": 5995 + }, + { + "epoch": 0.6584669448715132, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.4921538829803467, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7074058651924133, + "num_tokens": 155274055.0, + "step": 5996 + }, + { + "epoch": 0.6585767625741269, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.1288323402404785, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7049992680549622, + "num_tokens": 155299575.0, + "step": 5997 + }, + { + "epoch": 0.6586865802767407, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.506634473800659, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7054403424263, + "num_tokens": 155322020.0, + "step": 5998 + }, + { + "epoch": 0.6587963979793543, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.0289909839630127, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6936511397361755, + "num_tokens": 155351256.0, + "step": 5999 + }, + { + "epoch": 0.6589062156819679, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.113318920135498, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6890687346458435, + "num_tokens": 155380192.0, + "step": 6000 + }, + { + "epoch": 0.6590160333845816, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2101900577545166, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7001549005508423, + "num_tokens": 155404549.0, + "step": 6001 + }, + { + "epoch": 0.6591258510871952, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.085923671722412, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7016024589538574, + "num_tokens": 155432136.0, + "step": 6002 + }, + { + "epoch": 0.6592356687898089, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.045903444290161, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7001634240150452, + "num_tokens": 155462102.0, + "step": 6003 + }, + { + "epoch": 0.6593454864924225, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.207324743270874, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7058888077735901, + "num_tokens": 155486045.0, + "step": 6004 + }, + { + "epoch": 0.6594553041950363, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.2917275428771973, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6983180046081543, + "num_tokens": 155510272.0, + "step": 6005 + }, + { + "epoch": 0.6595651218976499, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 1.9208825826644897, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7236613035202026, + "num_tokens": 155541765.0, + "step": 6006 + }, + { + "epoch": 0.6596749396002636, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 1.924710750579834, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7027801275253296, + "num_tokens": 155574579.0, + "step": 6007 + }, + { + "epoch": 0.6597847573028772, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.014850616455078, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6927255392074585, + "num_tokens": 155603237.0, + "step": 6008 + }, + { + "epoch": 0.6598945750054909, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.5124151706695557, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7080609798431396, + "num_tokens": 155623349.0, + "step": 6009 + }, + { + "epoch": 0.6600043927081045, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.188302516937256, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.685224175453186, + "num_tokens": 155651660.0, + "step": 6010 + }, + { + "epoch": 0.6601142104107182, + "ewc_loss": 5.4836273193359375e-05, + "grad_norm": 2.214543581008911, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7034206390380859, + "num_tokens": 155676112.0, + "step": 6011 + }, + { + "epoch": 0.6602240281133319, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 1.8871874809265137, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7206916809082031, + "num_tokens": 155706650.0, + "step": 6012 + }, + { + "epoch": 0.6603338458159456, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 1.8136630058288574, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7063044309616089, + "num_tokens": 155742884.0, + "step": 6013 + }, + { + "epoch": 0.6604436635185592, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.0500855445861816, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7290169596672058, + "num_tokens": 155769598.0, + "step": 6014 + }, + { + "epoch": 0.6605534812211729, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.2845332622528076, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6971378922462463, + "num_tokens": 155792342.0, + "step": 6015 + }, + { + "epoch": 0.6606632989237865, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.0513415336608887, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7040217518806458, + "num_tokens": 155820917.0, + "step": 6016 + }, + { + "epoch": 0.6607731166264001, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.154477119445801, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7159584760665894, + "num_tokens": 155846192.0, + "step": 6017 + }, + { + "epoch": 0.6608829343290138, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.3059241771698, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6931442022323608, + "num_tokens": 155869072.0, + "step": 6018 + }, + { + "epoch": 0.6609927520316276, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.1661229133605957, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6935011744499207, + "num_tokens": 155898142.0, + "step": 6019 + }, + { + "epoch": 0.6611025697342412, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.0819547176361084, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7201328873634338, + "num_tokens": 155925699.0, + "step": 6020 + }, + { + "epoch": 0.6612123874368548, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.2790825366973877, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.692829966545105, + "num_tokens": 155950313.0, + "step": 6021 + }, + { + "epoch": 0.6613222051394685, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.04475474357605, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7176175117492676, + "num_tokens": 155977689.0, + "step": 6022 + }, + { + "epoch": 0.6614320228420821, + "ewc_loss": 5.507469177246094e-05, + "grad_norm": 2.2039854526519775, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7059879302978516, + "num_tokens": 156003254.0, + "step": 6023 + }, + { + "epoch": 0.6615418405446958, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2154791355133057, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7195477485656738, + "num_tokens": 156027004.0, + "step": 6024 + }, + { + "epoch": 0.6616516582473094, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1059775352478027, + "learning_rate": 1e-06, + "loss": 1.0419, + "mean_token_accuracy": 0.6830988526344299, + "num_tokens": 156057555.0, + "step": 6025 + }, + { + "epoch": 0.6617614759499231, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1832211017608643, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6929478645324707, + "num_tokens": 156083938.0, + "step": 6026 + }, + { + "epoch": 0.6618712936525368, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 1.9673817157745361, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7064642906188965, + "num_tokens": 156115291.0, + "step": 6027 + }, + { + "epoch": 0.6619811113551505, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0343010425567627, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7169362902641296, + "num_tokens": 156141710.0, + "step": 6028 + }, + { + "epoch": 0.6620909290577641, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1332318782806396, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6985326409339905, + "num_tokens": 156167447.0, + "step": 6029 + }, + { + "epoch": 0.6622007467603778, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0216825008392334, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6820673942565918, + "num_tokens": 156197415.0, + "step": 6030 + }, + { + "epoch": 0.6623105644629914, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 1.7321804761886597, + "learning_rate": 1e-06, + "loss": 1.0787, + "mean_token_accuracy": 0.6788212060928345, + "num_tokens": 156235692.0, + "step": 6031 + }, + { + "epoch": 0.6624203821656051, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.058788776397705, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7158765196800232, + "num_tokens": 156263420.0, + "step": 6032 + }, + { + "epoch": 0.6625301998682187, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.3318886756896973, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7224624156951904, + "num_tokens": 156285855.0, + "step": 6033 + }, + { + "epoch": 0.6626400175708325, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0736465454101562, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7023869752883911, + "num_tokens": 156315182.0, + "step": 6034 + }, + { + "epoch": 0.6627498352734461, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1815528869628906, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7154274582862854, + "num_tokens": 156340739.0, + "step": 6035 + }, + { + "epoch": 0.6628596529760598, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.122790813446045, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6974228620529175, + "num_tokens": 156367369.0, + "step": 6036 + }, + { + "epoch": 0.6629694706786734, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2075836658477783, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.693774938583374, + "num_tokens": 156393412.0, + "step": 6037 + }, + { + "epoch": 0.663079288381287, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2302780151367188, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7117112874984741, + "num_tokens": 156419102.0, + "step": 6038 + }, + { + "epoch": 0.6631891060839007, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.112816095352173, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.6933386325836182, + "num_tokens": 156446862.0, + "step": 6039 + }, + { + "epoch": 0.6632989237865143, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2378084659576416, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7066745758056641, + "num_tokens": 156471500.0, + "step": 6040 + }, + { + "epoch": 0.6634087414891281, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1732428073883057, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7095407843589783, + "num_tokens": 156497443.0, + "step": 6041 + }, + { + "epoch": 0.6635185591917417, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0064539909362793, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.6913731098175049, + "num_tokens": 156527999.0, + "step": 6042 + }, + { + "epoch": 0.6636283768943554, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.291703224182129, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6995511054992676, + "num_tokens": 156551836.0, + "step": 6043 + }, + { + "epoch": 0.663738194596969, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1241519451141357, + "learning_rate": 1e-06, + "loss": 1.0199, + "mean_token_accuracy": 0.6866545081138611, + "num_tokens": 156580317.0, + "step": 6044 + }, + { + "epoch": 0.6638480122995827, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.3191373348236084, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7169885039329529, + "num_tokens": 156603323.0, + "step": 6045 + }, + { + "epoch": 0.6639578300021963, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.101945400238037, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7042258977890015, + "num_tokens": 156630293.0, + "step": 6046 + }, + { + "epoch": 0.66406764770481, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0826425552368164, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.6911485195159912, + "num_tokens": 156658155.0, + "step": 6047 + }, + { + "epoch": 0.6641774654074237, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1228764057159424, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6957379579544067, + "num_tokens": 156685519.0, + "step": 6048 + }, + { + "epoch": 0.6642872831100374, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0767648220062256, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7150841951370239, + "num_tokens": 156712257.0, + "step": 6049 + }, + { + "epoch": 0.664397100812651, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.292949914932251, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7030710577964783, + "num_tokens": 156735594.0, + "step": 6050 + }, + { + "epoch": 0.6645069185152647, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.189375162124634, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.6983633041381836, + "num_tokens": 156760714.0, + "step": 6051 + }, + { + "epoch": 0.6646167362178783, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0036020278930664, + "learning_rate": 1e-06, + "loss": 1.098, + "mean_token_accuracy": 0.6803330183029175, + "num_tokens": 156791882.0, + "step": 6052 + }, + { + "epoch": 0.664726553920492, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2829065322875977, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6979318261146545, + "num_tokens": 156816170.0, + "step": 6053 + }, + { + "epoch": 0.6648363716231056, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2175002098083496, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7353061437606812, + "num_tokens": 156838617.0, + "step": 6054 + }, + { + "epoch": 0.6649461893257194, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.0382418632507324, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.687300443649292, + "num_tokens": 156870103.0, + "step": 6055 + }, + { + "epoch": 0.665056007028333, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2626078128814697, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7082343697547913, + "num_tokens": 156895180.0, + "step": 6056 + }, + { + "epoch": 0.6651658247309467, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.2987515926361084, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7072045207023621, + "num_tokens": 156919471.0, + "step": 6057 + }, + { + "epoch": 0.6652756424335603, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.142969846725464, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.7075514197349548, + "num_tokens": 156947861.0, + "step": 6058 + }, + { + "epoch": 0.665385460136174, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.5266661643981934, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7224370837211609, + "num_tokens": 156966379.0, + "step": 6059 + }, + { + "epoch": 0.6654952778387876, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.285243034362793, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7399747967720032, + "num_tokens": 156988824.0, + "step": 6060 + }, + { + "epoch": 0.6656050955414012, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.275334596633911, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6955870389938354, + "num_tokens": 157012835.0, + "step": 6061 + }, + { + "epoch": 0.6657149132440149, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1106252670288086, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7135136723518372, + "num_tokens": 157041090.0, + "step": 6062 + }, + { + "epoch": 0.6658247309466286, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.1658811569213867, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6901421546936035, + "num_tokens": 157069696.0, + "step": 6063 + }, + { + "epoch": 0.6659345486492423, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.490081548690796, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7391089200973511, + "num_tokens": 157088566.0, + "step": 6064 + }, + { + "epoch": 0.6660443663518559, + "ewc_loss": 5.53131103515625e-05, + "grad_norm": 2.117676258087158, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7105842232704163, + "num_tokens": 157114916.0, + "step": 6065 + }, + { + "epoch": 0.6661541840544696, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.095628023147583, + "learning_rate": 1e-06, + "loss": 1.0581, + "mean_token_accuracy": 0.6790770292282104, + "num_tokens": 157146212.0, + "step": 6066 + }, + { + "epoch": 0.6662640017570832, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.430716037750244, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7034768462181091, + "num_tokens": 157168951.0, + "step": 6067 + }, + { + "epoch": 0.6663738194596969, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.5015971660614014, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7118687629699707, + "num_tokens": 157189670.0, + "step": 6068 + }, + { + "epoch": 0.6664836371623105, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.5165305137634277, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7096304893493652, + "num_tokens": 157214174.0, + "step": 6069 + }, + { + "epoch": 0.6665934548649243, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1030194759368896, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6865044832229614, + "num_tokens": 157242797.0, + "step": 6070 + }, + { + "epoch": 0.6667032725675379, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.579608917236328, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7049549221992493, + "num_tokens": 157262190.0, + "step": 6071 + }, + { + "epoch": 0.6668130902701516, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3499279022216797, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7153000235557556, + "num_tokens": 157283822.0, + "step": 6072 + }, + { + "epoch": 0.6669229079727652, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.20731520652771, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6954960823059082, + "num_tokens": 157308337.0, + "step": 6073 + }, + { + "epoch": 0.6670327256753789, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1129493713378906, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7121757864952087, + "num_tokens": 157332784.0, + "step": 6074 + }, + { + "epoch": 0.6671425433779925, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1777219772338867, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.688568115234375, + "num_tokens": 157360868.0, + "step": 6075 + }, + { + "epoch": 0.6672523610806061, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.4357078075408936, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.697671115398407, + "num_tokens": 157382569.0, + "step": 6076 + }, + { + "epoch": 0.6673621787832199, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1153616905212402, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6943159103393555, + "num_tokens": 157409460.0, + "step": 6077 + }, + { + "epoch": 0.6674719964858336, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.019564390182495, + "learning_rate": 1e-06, + "loss": 1.0426, + "mean_token_accuracy": 0.6886745691299438, + "num_tokens": 157440052.0, + "step": 6078 + }, + { + "epoch": 0.6675818141884472, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3105103969573975, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6953778266906738, + "num_tokens": 157464038.0, + "step": 6079 + }, + { + "epoch": 0.6676916318910608, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.0093846321105957, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.6973937749862671, + "num_tokens": 157492927.0, + "step": 6080 + }, + { + "epoch": 0.6678014495936745, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 1.9424948692321777, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7162576913833618, + "num_tokens": 157522518.0, + "step": 6081 + }, + { + "epoch": 0.6679112672962881, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.2302279472351074, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7073952555656433, + "num_tokens": 157547170.0, + "step": 6082 + }, + { + "epoch": 0.6680210849989018, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.145439863204956, + "learning_rate": 1e-06, + "loss": 1.1362, + "mean_token_accuracy": 0.6696594953536987, + "num_tokens": 157575044.0, + "step": 6083 + }, + { + "epoch": 0.6681309027015155, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 1.7574700117111206, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.6869844794273376, + "num_tokens": 157614962.0, + "step": 6084 + }, + { + "epoch": 0.6682407204041292, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.2803754806518555, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7046799659729004, + "num_tokens": 157638797.0, + "step": 6085 + }, + { + "epoch": 0.6683505381067428, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1759774684906006, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7043924927711487, + "num_tokens": 157663629.0, + "step": 6086 + }, + { + "epoch": 0.6684603558093565, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3185622692108154, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.722801923751831, + "num_tokens": 157686851.0, + "step": 6087 + }, + { + "epoch": 0.6685701735119701, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.4667539596557617, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7242093682289124, + "num_tokens": 157706324.0, + "step": 6088 + }, + { + "epoch": 0.6686799912145838, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.376286506652832, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6947934031486511, + "num_tokens": 157730442.0, + "step": 6089 + }, + { + "epoch": 0.6687898089171974, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3080742359161377, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7202740907669067, + "num_tokens": 157753769.0, + "step": 6090 + }, + { + "epoch": 0.6688996266198111, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.0513954162597656, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6968605518341064, + "num_tokens": 157783437.0, + "step": 6091 + }, + { + "epoch": 0.6690094443224248, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.068117141723633, + "learning_rate": 1e-06, + "loss": 1.0742, + "mean_token_accuracy": 0.6736367344856262, + "num_tokens": 157813602.0, + "step": 6092 + }, + { + "epoch": 0.6691192620250385, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.036663293838501, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7120748162269592, + "num_tokens": 157841665.0, + "step": 6093 + }, + { + "epoch": 0.6692290797276521, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3673737049102783, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7056077718734741, + "num_tokens": 157864011.0, + "step": 6094 + }, + { + "epoch": 0.6693388974302658, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.0140364170074463, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7085877656936646, + "num_tokens": 157893302.0, + "step": 6095 + }, + { + "epoch": 0.6694487151328794, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.2523937225341797, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7222306728363037, + "num_tokens": 157917252.0, + "step": 6096 + }, + { + "epoch": 0.669558532835493, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 1.990339994430542, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7065269947052002, + "num_tokens": 157946229.0, + "step": 6097 + }, + { + "epoch": 0.6696683505381067, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.2058892250061035, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6901384592056274, + "num_tokens": 157971749.0, + "step": 6098 + }, + { + "epoch": 0.6697781682407205, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 1.9390641450881958, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.6820933818817139, + "num_tokens": 158004797.0, + "step": 6099 + }, + { + "epoch": 0.6698879859433341, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1355667114257812, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.7024401426315308, + "num_tokens": 158032561.0, + "step": 6100 + }, + { + "epoch": 0.6699978036459477, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.320338726043701, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7274057865142822, + "num_tokens": 158054892.0, + "step": 6101 + }, + { + "epoch": 0.6701076213485614, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.141059398651123, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6892385482788086, + "num_tokens": 158083531.0, + "step": 6102 + }, + { + "epoch": 0.670217439051175, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.4313554763793945, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6942343711853027, + "num_tokens": 158105142.0, + "step": 6103 + }, + { + "epoch": 0.6703272567537887, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.606290102005005, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7068668007850647, + "num_tokens": 158124184.0, + "step": 6104 + }, + { + "epoch": 0.6704370744564023, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.4471611976623535, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7186892032623291, + "num_tokens": 158144656.0, + "step": 6105 + }, + { + "epoch": 0.6705468921590161, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3522472381591797, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7064886689186096, + "num_tokens": 158168309.0, + "step": 6106 + }, + { + "epoch": 0.6706567098616297, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.420985221862793, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7208883762359619, + "num_tokens": 158189567.0, + "step": 6107 + }, + { + "epoch": 0.6707665275642434, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3804590702056885, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7086374163627625, + "num_tokens": 158210476.0, + "step": 6108 + }, + { + "epoch": 0.670876345266857, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.101787805557251, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7201303243637085, + "num_tokens": 158237155.0, + "step": 6109 + }, + { + "epoch": 0.6709861629694707, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.403773307800293, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.6891738176345825, + "num_tokens": 158260792.0, + "step": 6110 + }, + { + "epoch": 0.6710959806720843, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.481999158859253, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7206679582595825, + "num_tokens": 158280874.0, + "step": 6111 + }, + { + "epoch": 0.671205798374698, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.2456820011138916, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7203496694564819, + "num_tokens": 158304659.0, + "step": 6112 + }, + { + "epoch": 0.6713156160773117, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.0728678703308105, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.6862732172012329, + "num_tokens": 158334472.0, + "step": 6113 + }, + { + "epoch": 0.6714254337799254, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 1.9888914823532104, + "learning_rate": 1e-06, + "loss": 1.0411, + "mean_token_accuracy": 0.6874366402626038, + "num_tokens": 158367217.0, + "step": 6114 + }, + { + "epoch": 0.671535251482539, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.000051259994507, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6965228319168091, + "num_tokens": 158396304.0, + "step": 6115 + }, + { + "epoch": 0.6716450691851527, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.1473605632781982, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6940341591835022, + "num_tokens": 158422496.0, + "step": 6116 + }, + { + "epoch": 0.6717548868877663, + "ewc_loss": 5.555152893066406e-05, + "grad_norm": 2.3638012409210205, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7048779129981995, + "num_tokens": 158445506.0, + "step": 6117 + }, + { + "epoch": 0.67186470459038, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.040649175643921, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7146804332733154, + "num_tokens": 158473721.0, + "step": 6118 + }, + { + "epoch": 0.6719745222929936, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.3005270957946777, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7151224613189697, + "num_tokens": 158496941.0, + "step": 6119 + }, + { + "epoch": 0.6720843399956072, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.027930736541748, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7221639156341553, + "num_tokens": 158525505.0, + "step": 6120 + }, + { + "epoch": 0.672194157698221, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.273345947265625, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7055180072784424, + "num_tokens": 158548342.0, + "step": 6121 + }, + { + "epoch": 0.6723039754008346, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0387794971466064, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.693374514579773, + "num_tokens": 158577693.0, + "step": 6122 + }, + { + "epoch": 0.6724137931034483, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.177517890930176, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7171999216079712, + "num_tokens": 158602480.0, + "step": 6123 + }, + { + "epoch": 0.6725236108060619, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0424251556396484, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.705156683921814, + "num_tokens": 158631339.0, + "step": 6124 + }, + { + "epoch": 0.6726334285086756, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.4802989959716797, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.71051025390625, + "num_tokens": 158651126.0, + "step": 6125 + }, + { + "epoch": 0.6727432462112892, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.3330979347229004, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7204259037971497, + "num_tokens": 158674262.0, + "step": 6126 + }, + { + "epoch": 0.6728530639139029, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.4455432891845703, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.6960914134979248, + "num_tokens": 158695067.0, + "step": 6127 + }, + { + "epoch": 0.6729628816165166, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.4675867557525635, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7314815521240234, + "num_tokens": 158716763.0, + "step": 6128 + }, + { + "epoch": 0.6730726993191303, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.106281042098999, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7038243412971497, + "num_tokens": 158743441.0, + "step": 6129 + }, + { + "epoch": 0.6731825170217439, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.2192208766937256, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7117788791656494, + "num_tokens": 158768551.0, + "step": 6130 + }, + { + "epoch": 0.6732923347243576, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.2650983333587646, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7024092674255371, + "num_tokens": 158793558.0, + "step": 6131 + }, + { + "epoch": 0.6734021524269712, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.1493422985076904, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7297536730766296, + "num_tokens": 158818430.0, + "step": 6132 + }, + { + "epoch": 0.6735119701295849, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.300083875656128, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6936429738998413, + "num_tokens": 158843138.0, + "step": 6133 + }, + { + "epoch": 0.6736217878321985, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0944955348968506, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.7049031257629395, + "num_tokens": 158874477.0, + "step": 6134 + }, + { + "epoch": 0.6737316055348123, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.225419521331787, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7102366089820862, + "num_tokens": 158899004.0, + "step": 6135 + }, + { + "epoch": 0.6738414232374259, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.191704034805298, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7044068574905396, + "num_tokens": 158924939.0, + "step": 6136 + }, + { + "epoch": 0.6739512409400396, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.1802704334259033, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7190137505531311, + "num_tokens": 158950178.0, + "step": 6137 + }, + { + "epoch": 0.6740610586426532, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0053505897521973, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7007858753204346, + "num_tokens": 158979755.0, + "step": 6138 + }, + { + "epoch": 0.6741708763452668, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.138854742050171, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7072204351425171, + "num_tokens": 159007944.0, + "step": 6139 + }, + { + "epoch": 0.6742806940478805, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.9704644680023193, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7192424535751343, + "num_tokens": 159037619.0, + "step": 6140 + }, + { + "epoch": 0.6743905117504941, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.1371564865112305, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7142804861068726, + "num_tokens": 159065207.0, + "step": 6141 + }, + { + "epoch": 0.6745003294531079, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.149770736694336, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.6989767551422119, + "num_tokens": 159091637.0, + "step": 6142 + }, + { + "epoch": 0.6746101471557215, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.3798153400421143, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7089192867279053, + "num_tokens": 159114372.0, + "step": 6143 + }, + { + "epoch": 0.6747199648583352, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.152137279510498, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.694544792175293, + "num_tokens": 159141680.0, + "step": 6144 + }, + { + "epoch": 0.6748297825609488, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.472564458847046, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7156018018722534, + "num_tokens": 159162816.0, + "step": 6145 + }, + { + "epoch": 0.6749396002635625, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.1624794006347656, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.6977235078811646, + "num_tokens": 159188403.0, + "step": 6146 + }, + { + "epoch": 0.6750494179661761, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0256049633026123, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6964684724807739, + "num_tokens": 159218711.0, + "step": 6147 + }, + { + "epoch": 0.6751592356687898, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.1403841972351074, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6881946325302124, + "num_tokens": 159246877.0, + "step": 6148 + }, + { + "epoch": 0.6752690533714034, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.079664945602417, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7113271355628967, + "num_tokens": 159275042.0, + "step": 6149 + }, + { + "epoch": 0.6753788710740172, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.85099458694458, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7154809236526489, + "num_tokens": 159307958.0, + "step": 6150 + }, + { + "epoch": 0.6754886887766308, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.4302616119384766, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7045437097549438, + "num_tokens": 159330280.0, + "step": 6151 + }, + { + "epoch": 0.6755985064792445, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.0695576667785645, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.7015023231506348, + "num_tokens": 159358692.0, + "step": 6152 + }, + { + "epoch": 0.6757083241818581, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.9788565635681152, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7080284357070923, + "num_tokens": 159387486.0, + "step": 6153 + }, + { + "epoch": 0.6758181418844718, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.2709929943084717, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7096554040908813, + "num_tokens": 159411417.0, + "step": 6154 + }, + { + "epoch": 0.6759279595870854, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.9905308485031128, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.711310863494873, + "num_tokens": 159442444.0, + "step": 6155 + }, + { + "epoch": 0.676037777289699, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.2251081466674805, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6978020071983337, + "num_tokens": 159466545.0, + "step": 6156 + }, + { + "epoch": 0.6761475949923128, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 2.2523512840270996, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6985450983047485, + "num_tokens": 159491951.0, + "step": 6157 + }, + { + "epoch": 0.6762574126949265, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.9651447534561157, + "learning_rate": 1e-06, + "loss": 1.1088, + "mean_token_accuracy": 0.6735261678695679, + "num_tokens": 159524107.0, + "step": 6158 + }, + { + "epoch": 0.6763672303975401, + "ewc_loss": 5.602836608886719e-05, + "grad_norm": 1.7664538621902466, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7134178876876831, + "num_tokens": 159560268.0, + "step": 6159 + }, + { + "epoch": 0.6764770481001537, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.2697129249572754, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7409592866897583, + "num_tokens": 159582247.0, + "step": 6160 + }, + { + "epoch": 0.6765868658027674, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.203679323196411, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7141430377960205, + "num_tokens": 159608502.0, + "step": 6161 + }, + { + "epoch": 0.676696683505381, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.449537515640259, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7225879430770874, + "num_tokens": 159629373.0, + "step": 6162 + }, + { + "epoch": 0.6768065012079947, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.2849040031433105, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6923478245735168, + "num_tokens": 159654888.0, + "step": 6163 + }, + { + "epoch": 0.6769163189106084, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.2304461002349854, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7136012315750122, + "num_tokens": 159678909.0, + "step": 6164 + }, + { + "epoch": 0.6770261366132221, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.5740063190460205, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7167357206344604, + "num_tokens": 159699296.0, + "step": 6165 + }, + { + "epoch": 0.6771359543158357, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.053345203399658, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.692831814289093, + "num_tokens": 159728051.0, + "step": 6166 + }, + { + "epoch": 0.6772457720184494, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.187288761138916, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7166226506233215, + "num_tokens": 159752609.0, + "step": 6167 + }, + { + "epoch": 0.677355589721063, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.596083164215088, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.6988790035247803, + "num_tokens": 159775222.0, + "step": 6168 + }, + { + "epoch": 0.6774654074236767, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.4735262393951416, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6958099603652954, + "num_tokens": 159795594.0, + "step": 6169 + }, + { + "epoch": 0.6775752251262903, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.20188045501709, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.7002501487731934, + "num_tokens": 159821063.0, + "step": 6170 + }, + { + "epoch": 0.6776850428289041, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 1.9750717878341675, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.71112060546875, + "num_tokens": 159850148.0, + "step": 6171 + }, + { + "epoch": 0.6777948605315177, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.257800579071045, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7137618064880371, + "num_tokens": 159875384.0, + "step": 6172 + }, + { + "epoch": 0.6779046782341314, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.2134876251220703, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7101947069168091, + "num_tokens": 159898946.0, + "step": 6173 + }, + { + "epoch": 0.678014495936745, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.261383533477783, + "learning_rate": 1e-06, + "loss": 1.0544, + "mean_token_accuracy": 0.6827841997146606, + "num_tokens": 159925065.0, + "step": 6174 + }, + { + "epoch": 0.6781243136393587, + "ewc_loss": 5.626678466796875e-05, + "grad_norm": 2.190277099609375, + "learning_rate": 1e-06, + "loss": 1.0429, + "mean_token_accuracy": 0.6826556921005249, + "num_tokens": 159950968.0, + "step": 6175 + }, + { + "epoch": 0.6782341313419723, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.404315710067749, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.695813775062561, + "num_tokens": 159972968.0, + "step": 6176 + }, + { + "epoch": 0.678343949044586, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.007847309112549, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7416518926620483, + "num_tokens": 160000824.0, + "step": 6177 + }, + { + "epoch": 0.6784537667471996, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2908666133880615, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6833782196044922, + "num_tokens": 160026215.0, + "step": 6178 + }, + { + "epoch": 0.6785635844498134, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2841544151306152, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.704075813293457, + "num_tokens": 160048907.0, + "step": 6179 + }, + { + "epoch": 0.678673402152427, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.0395500659942627, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.696401834487915, + "num_tokens": 160078943.0, + "step": 6180 + }, + { + "epoch": 0.6787832198550406, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.1827244758605957, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7107669115066528, + "num_tokens": 160105582.0, + "step": 6181 + }, + { + "epoch": 0.6788930375576543, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2748403549194336, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7064511775970459, + "num_tokens": 160128859.0, + "step": 6182 + }, + { + "epoch": 0.6790028552602679, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 1.9624875783920288, + "learning_rate": 1e-06, + "loss": 1.1051, + "mean_token_accuracy": 0.672039270401001, + "num_tokens": 160161771.0, + "step": 6183 + }, + { + "epoch": 0.6791126729628816, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.3833060264587402, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.711270809173584, + "num_tokens": 160185315.0, + "step": 6184 + }, + { + "epoch": 0.6792224906654952, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.111650228500366, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.733873724937439, + "num_tokens": 160210093.0, + "step": 6185 + }, + { + "epoch": 0.679332308368109, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2699382305145264, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.699546217918396, + "num_tokens": 160235947.0, + "step": 6186 + }, + { + "epoch": 0.6794421260707226, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.107245922088623, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7150195240974426, + "num_tokens": 160262839.0, + "step": 6187 + }, + { + "epoch": 0.6795519437733363, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.1074767112731934, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7036944627761841, + "num_tokens": 160291066.0, + "step": 6188 + }, + { + "epoch": 0.6796617614759499, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2264034748077393, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6974521279335022, + "num_tokens": 160316101.0, + "step": 6189 + }, + { + "epoch": 0.6797715791785636, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.0572431087493896, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7210092544555664, + "num_tokens": 160342531.0, + "step": 6190 + }, + { + "epoch": 0.6798813968811772, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.536311626434326, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7090522646903992, + "num_tokens": 160361907.0, + "step": 6191 + }, + { + "epoch": 0.6799912145837909, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.222541332244873, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.728703498840332, + "num_tokens": 160385012.0, + "step": 6192 + }, + { + "epoch": 0.6801010322864046, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 1.9579038619995117, + "learning_rate": 1e-06, + "loss": 1.0795, + "mean_token_accuracy": 0.677009105682373, + "num_tokens": 160418417.0, + "step": 6193 + }, + { + "epoch": 0.6802108499890183, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.630688428878784, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7203297019004822, + "num_tokens": 160436585.0, + "step": 6194 + }, + { + "epoch": 0.6803206676916319, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.0346908569335938, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6977609395980835, + "num_tokens": 160465543.0, + "step": 6195 + }, + { + "epoch": 0.6804304853942456, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.1792232990264893, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6906524896621704, + "num_tokens": 160494434.0, + "step": 6196 + }, + { + "epoch": 0.6805403030968592, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.0095932483673096, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6984100341796875, + "num_tokens": 160522554.0, + "step": 6197 + }, + { + "epoch": 0.6806501207994728, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.3477673530578613, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7042319774627686, + "num_tokens": 160545157.0, + "step": 6198 + }, + { + "epoch": 0.6807599385020865, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.0065927505493164, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6966000199317932, + "num_tokens": 160573827.0, + "step": 6199 + }, + { + "epoch": 0.6808697562047002, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.581867218017578, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.72316575050354, + "num_tokens": 160592697.0, + "step": 6200 + }, + { + "epoch": 0.6809795739073139, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.054490089416504, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.706323504447937, + "num_tokens": 160621152.0, + "step": 6201 + }, + { + "epoch": 0.6810893916099275, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.039717197418213, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7064915895462036, + "num_tokens": 160649670.0, + "step": 6202 + }, + { + "epoch": 0.6811992093125412, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.4192922115325928, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7043091654777527, + "num_tokens": 160672004.0, + "step": 6203 + }, + { + "epoch": 0.6813090270151548, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.038660764694214, + "learning_rate": 1e-06, + "loss": 1.0251, + "mean_token_accuracy": 0.6916647553443909, + "num_tokens": 160701671.0, + "step": 6204 + }, + { + "epoch": 0.6814188447177685, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.355191946029663, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.692434549331665, + "num_tokens": 160724240.0, + "step": 6205 + }, + { + "epoch": 0.6815286624203821, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.1822526454925537, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.6945551037788391, + "num_tokens": 160750043.0, + "step": 6206 + }, + { + "epoch": 0.6816384801229959, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.289430856704712, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6973061561584473, + "num_tokens": 160775094.0, + "step": 6207 + }, + { + "epoch": 0.6817482978256095, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 1.9932715892791748, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7326929569244385, + "num_tokens": 160804642.0, + "step": 6208 + }, + { + "epoch": 0.6818581155282232, + "ewc_loss": 5.650520324707031e-05, + "grad_norm": 2.2838706970214844, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7147096395492554, + "num_tokens": 160828357.0, + "step": 6209 + }, + { + "epoch": 0.6819679332308368, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.2538154125213623, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7281598448753357, + "num_tokens": 160852946.0, + "step": 6210 + }, + { + "epoch": 0.6820777509334505, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.288963556289673, + "learning_rate": 1e-06, + "loss": 1.0466, + "mean_token_accuracy": 0.6880316734313965, + "num_tokens": 160880009.0, + "step": 6211 + }, + { + "epoch": 0.6821875686360641, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 1.9220857620239258, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7094351053237915, + "num_tokens": 160910735.0, + "step": 6212 + }, + { + "epoch": 0.6822973863386778, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.0889997482299805, + "learning_rate": 1e-06, + "loss": 1.0727, + "mean_token_accuracy": 0.6791768670082092, + "num_tokens": 160940187.0, + "step": 6213 + }, + { + "epoch": 0.6824072040412914, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.0673680305480957, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7174547910690308, + "num_tokens": 160967733.0, + "step": 6214 + }, + { + "epoch": 0.6825170217439052, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.314962148666382, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.715161919593811, + "num_tokens": 160991485.0, + "step": 6215 + }, + { + "epoch": 0.6826268394465188, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.097648859024048, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6960502862930298, + "num_tokens": 161022484.0, + "step": 6216 + }, + { + "epoch": 0.6827366571491325, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.648902177810669, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7085991501808167, + "num_tokens": 161041677.0, + "step": 6217 + }, + { + "epoch": 0.6828464748517461, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 1.9028472900390625, + "learning_rate": 1e-06, + "loss": 1.0985, + "mean_token_accuracy": 0.6804753541946411, + "num_tokens": 161078083.0, + "step": 6218 + }, + { + "epoch": 0.6829562925543597, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.0979835987091064, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.709689736366272, + "num_tokens": 161105164.0, + "step": 6219 + }, + { + "epoch": 0.6830661102569734, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.0891170501708984, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6997893452644348, + "num_tokens": 161134797.0, + "step": 6220 + }, + { + "epoch": 0.683175927959587, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 1.9567925930023193, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7131927013397217, + "num_tokens": 161163605.0, + "step": 6221 + }, + { + "epoch": 0.6832857456622008, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.179612398147583, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7082209587097168, + "num_tokens": 161189308.0, + "step": 6222 + }, + { + "epoch": 0.6833955633648144, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.3015217781066895, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7064002752304077, + "num_tokens": 161212337.0, + "step": 6223 + }, + { + "epoch": 0.6835053810674281, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.3350679874420166, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7097584009170532, + "num_tokens": 161234858.0, + "step": 6224 + }, + { + "epoch": 0.6836151987700417, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.0871219635009766, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7110333442687988, + "num_tokens": 161261948.0, + "step": 6225 + }, + { + "epoch": 0.6837250164726554, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 2.200435161590576, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7105154395103455, + "num_tokens": 161285655.0, + "step": 6226 + }, + { + "epoch": 0.683834834175269, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.088730573654175, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7081038355827332, + "num_tokens": 161313042.0, + "step": 6227 + }, + { + "epoch": 0.6839446518778827, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.314415216445923, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6997255682945251, + "num_tokens": 161339421.0, + "step": 6228 + }, + { + "epoch": 0.6840544695804964, + "ewc_loss": 5.698204040527344e-05, + "grad_norm": 1.8846378326416016, + "learning_rate": 1e-06, + "loss": 1.0644, + "mean_token_accuracy": 0.6867140531539917, + "num_tokens": 161374361.0, + "step": 6229 + }, + { + "epoch": 0.6841642872831101, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.151434898376465, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.705675482749939, + "num_tokens": 161400878.0, + "step": 6230 + }, + { + "epoch": 0.6842741049857237, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.109710693359375, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7066116333007812, + "num_tokens": 161428464.0, + "step": 6231 + }, + { + "epoch": 0.6843839226883374, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.318222761154175, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7329031229019165, + "num_tokens": 161450688.0, + "step": 6232 + }, + { + "epoch": 0.684493740390951, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3462300300598145, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7054345607757568, + "num_tokens": 161472646.0, + "step": 6233 + }, + { + "epoch": 0.6846035580935647, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.365116834640503, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6970398426055908, + "num_tokens": 161495999.0, + "step": 6234 + }, + { + "epoch": 0.6847133757961783, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.2463386058807373, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7200475931167603, + "num_tokens": 161518854.0, + "step": 6235 + }, + { + "epoch": 0.6848231934987921, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3931567668914795, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7003101110458374, + "num_tokens": 161541927.0, + "step": 6236 + }, + { + "epoch": 0.6849330112014057, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.175584316253662, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.70833420753479, + "num_tokens": 161568560.0, + "step": 6237 + }, + { + "epoch": 0.6850428289040194, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3967068195343018, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7055773735046387, + "num_tokens": 161592384.0, + "step": 6238 + }, + { + "epoch": 0.685152646606633, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.120697498321533, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6915873289108276, + "num_tokens": 161620897.0, + "step": 6239 + }, + { + "epoch": 0.6852624643092466, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.328228712081909, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7163017988204956, + "num_tokens": 161644189.0, + "step": 6240 + }, + { + "epoch": 0.6853722820118603, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.215958595275879, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6917699575424194, + "num_tokens": 161668350.0, + "step": 6241 + }, + { + "epoch": 0.6854820997144739, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3984060287475586, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7215292453765869, + "num_tokens": 161691163.0, + "step": 6242 + }, + { + "epoch": 0.6855919174170876, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.1781697273254395, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7206434011459351, + "num_tokens": 161715801.0, + "step": 6243 + }, + { + "epoch": 0.6857017351197013, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.194887638092041, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.69180828332901, + "num_tokens": 161743236.0, + "step": 6244 + }, + { + "epoch": 0.685811552822315, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.0223348140716553, + "learning_rate": 1e-06, + "loss": 1.0915, + "mean_token_accuracy": 0.6761994361877441, + "num_tokens": 161772576.0, + "step": 6245 + }, + { + "epoch": 0.6859213705249286, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.336463212966919, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7087982892990112, + "num_tokens": 161798925.0, + "step": 6246 + }, + { + "epoch": 0.6860311882275423, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.512662172317505, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7121862173080444, + "num_tokens": 161820941.0, + "step": 6247 + }, + { + "epoch": 0.6861410059301559, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.232147693634033, + "learning_rate": 1e-06, + "loss": 1.0359, + "mean_token_accuracy": 0.6971158385276794, + "num_tokens": 161846081.0, + "step": 6248 + }, + { + "epoch": 0.6862508236327696, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.090501308441162, + "learning_rate": 1e-06, + "loss": 1.1049, + "mean_token_accuracy": 0.6713172793388367, + "num_tokens": 161875326.0, + "step": 6249 + }, + { + "epoch": 0.6863606413353832, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.1728382110595703, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7093450427055359, + "num_tokens": 161902576.0, + "step": 6250 + }, + { + "epoch": 0.686470459037997, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.6318202018737793, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7191475629806519, + "num_tokens": 161920422.0, + "step": 6251 + }, + { + "epoch": 0.6865802767406106, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.379910469055176, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7169731259346008, + "num_tokens": 161943729.0, + "step": 6252 + }, + { + "epoch": 0.6866900944432243, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.173964738845825, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6954616904258728, + "num_tokens": 161970782.0, + "step": 6253 + }, + { + "epoch": 0.6867999121458379, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 1.8436009883880615, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.6906263828277588, + "num_tokens": 162007033.0, + "step": 6254 + }, + { + "epoch": 0.6869097298484516, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.470834732055664, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.713096022605896, + "num_tokens": 162027881.0, + "step": 6255 + }, + { + "epoch": 0.6870195475510652, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.0270822048187256, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.701439380645752, + "num_tokens": 162058289.0, + "step": 6256 + }, + { + "epoch": 0.6871293652536788, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3983585834503174, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7272560596466064, + "num_tokens": 162079047.0, + "step": 6257 + }, + { + "epoch": 0.6872391829562926, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3912477493286133, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7056863307952881, + "num_tokens": 162100921.0, + "step": 6258 + }, + { + "epoch": 0.6873490006589063, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.253763198852539, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7002028226852417, + "num_tokens": 162125491.0, + "step": 6259 + }, + { + "epoch": 0.6874588183615199, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.221637487411499, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6964797973632812, + "num_tokens": 162150785.0, + "step": 6260 + }, + { + "epoch": 0.6875686360641335, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.3253824710845947, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6952993869781494, + "num_tokens": 162173188.0, + "step": 6261 + }, + { + "epoch": 0.6876784537667472, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 1.9694756269454956, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7227212190628052, + "num_tokens": 162203116.0, + "step": 6262 + }, + { + "epoch": 0.6877882714693608, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.065770149230957, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.7004818320274353, + "num_tokens": 162230643.0, + "step": 6263 + }, + { + "epoch": 0.6878980891719745, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.163417100906372, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.6936421394348145, + "num_tokens": 162259943.0, + "step": 6264 + }, + { + "epoch": 0.6880079068745882, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.2181572914123535, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7056916952133179, + "num_tokens": 162284553.0, + "step": 6265 + }, + { + "epoch": 0.6881177245772019, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.306835889816284, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6952922344207764, + "num_tokens": 162309858.0, + "step": 6266 + }, + { + "epoch": 0.6882275422798155, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.1229448318481445, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7040008902549744, + "num_tokens": 162337622.0, + "step": 6267 + }, + { + "epoch": 0.6883373599824292, + "ewc_loss": 5.7220458984375e-05, + "grad_norm": 2.0637521743774414, + "learning_rate": 1e-06, + "loss": 1.0393, + "mean_token_accuracy": 0.6971243619918823, + "num_tokens": 162366547.0, + "step": 6268 + }, + { + "epoch": 0.6884471776850428, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.008181571960449, + "learning_rate": 1e-06, + "loss": 1.0486, + "mean_token_accuracy": 0.6824113130569458, + "num_tokens": 162396141.0, + "step": 6269 + }, + { + "epoch": 0.6885569953876565, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.2986812591552734, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7184344530105591, + "num_tokens": 162420169.0, + "step": 6270 + }, + { + "epoch": 0.6886668130902701, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.214405059814453, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6924433708190918, + "num_tokens": 162445872.0, + "step": 6271 + }, + { + "epoch": 0.6887766307928838, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.030728578567505, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.704102635383606, + "num_tokens": 162473104.0, + "step": 6272 + }, + { + "epoch": 0.6888864484954975, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.200035810470581, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7212697267532349, + "num_tokens": 162498431.0, + "step": 6273 + }, + { + "epoch": 0.6889962661981112, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 1.9723219871520996, + "learning_rate": 1e-06, + "loss": 1.0681, + "mean_token_accuracy": 0.6799841523170471, + "num_tokens": 162528782.0, + "step": 6274 + }, + { + "epoch": 0.6891060839007248, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.2002718448638916, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7018054723739624, + "num_tokens": 162553110.0, + "step": 6275 + }, + { + "epoch": 0.6892159016033385, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.0913219451904297, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.6906133890151978, + "num_tokens": 162581838.0, + "step": 6276 + }, + { + "epoch": 0.6893257193059521, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.281154155731201, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6902978420257568, + "num_tokens": 162604965.0, + "step": 6277 + }, + { + "epoch": 0.6894355370085657, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.0141441822052, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7129992246627808, + "num_tokens": 162634386.0, + "step": 6278 + }, + { + "epoch": 0.6895453547111794, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.3102970123291016, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6824434399604797, + "num_tokens": 162660520.0, + "step": 6279 + }, + { + "epoch": 0.6896551724137931, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.4042582511901855, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7212635278701782, + "num_tokens": 162683482.0, + "step": 6280 + }, + { + "epoch": 0.6897649901164068, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.062682628631592, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7071144580841064, + "num_tokens": 162712341.0, + "step": 6281 + }, + { + "epoch": 0.6898748078190204, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.2465639114379883, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7199268341064453, + "num_tokens": 162735854.0, + "step": 6282 + }, + { + "epoch": 0.6899846255216341, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 1.9851405620574951, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7034800052642822, + "num_tokens": 162767452.0, + "step": 6283 + }, + { + "epoch": 0.6900944432242477, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.458127737045288, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7219370603561401, + "num_tokens": 162788127.0, + "step": 6284 + }, + { + "epoch": 0.6902042609268614, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.079572916030884, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7061941623687744, + "num_tokens": 162817399.0, + "step": 6285 + }, + { + "epoch": 0.690314078629475, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.3121378421783447, + "learning_rate": 1e-06, + "loss": 1.0384, + "mean_token_accuracy": 0.6848578453063965, + "num_tokens": 162842073.0, + "step": 6286 + }, + { + "epoch": 0.6904238963320888, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.229124069213867, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7323542833328247, + "num_tokens": 162865066.0, + "step": 6287 + }, + { + "epoch": 0.6905337140347024, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.2496283054351807, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7111563086509705, + "num_tokens": 162889656.0, + "step": 6288 + }, + { + "epoch": 0.6906435317373161, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.224924325942993, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7137256264686584, + "num_tokens": 162913759.0, + "step": 6289 + }, + { + "epoch": 0.6907533494399297, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.1561336517333984, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7119206190109253, + "num_tokens": 162939871.0, + "step": 6290 + }, + { + "epoch": 0.6908631671425434, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.105295419692993, + "learning_rate": 1e-06, + "loss": 1.0375, + "mean_token_accuracy": 0.7053105235099792, + "num_tokens": 162968388.0, + "step": 6291 + }, + { + "epoch": 0.690972984845157, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.3599390983581543, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7241732478141785, + "num_tokens": 162991016.0, + "step": 6292 + }, + { + "epoch": 0.6910828025477707, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.230814218521118, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7355782985687256, + "num_tokens": 163014896.0, + "step": 6293 + }, + { + "epoch": 0.6911926202503844, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.5128018856048584, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7208859920501709, + "num_tokens": 163035705.0, + "step": 6294 + }, + { + "epoch": 0.6913024379529981, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.137035608291626, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6817771196365356, + "num_tokens": 163064708.0, + "step": 6295 + }, + { + "epoch": 0.6914122556556117, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.112140655517578, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6797865629196167, + "num_tokens": 163092567.0, + "step": 6296 + }, + { + "epoch": 0.6915220733582254, + "ewc_loss": 5.745887756347656e-05, + "grad_norm": 2.1646904945373535, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7161754369735718, + "num_tokens": 163117372.0, + "step": 6297 + }, + { + "epoch": 0.691631891060839, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1354193687438965, + "learning_rate": 1e-06, + "loss": 1.0793, + "mean_token_accuracy": 0.6773669719696045, + "num_tokens": 163142418.0, + "step": 6298 + }, + { + "epoch": 0.6917417087634526, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.040485382080078, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6939156651496887, + "num_tokens": 163172080.0, + "step": 6299 + }, + { + "epoch": 0.6918515264660663, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1434383392333984, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7006610631942749, + "num_tokens": 163197984.0, + "step": 6300 + }, + { + "epoch": 0.6919613441686799, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.094304323196411, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6954164505004883, + "num_tokens": 163226464.0, + "step": 6301 + }, + { + "epoch": 0.6920711618712937, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.09456467628479, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7111090421676636, + "num_tokens": 163256158.0, + "step": 6302 + }, + { + "epoch": 0.6921809795739073, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.483736515045166, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6988463997840881, + "num_tokens": 163277834.0, + "step": 6303 + }, + { + "epoch": 0.692290797276521, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.390251874923706, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7160401344299316, + "num_tokens": 163299028.0, + "step": 6304 + }, + { + "epoch": 0.6924006149791346, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.9697668552398682, + "learning_rate": 1e-06, + "loss": 1.0232, + "mean_token_accuracy": 0.6884719133377075, + "num_tokens": 163329095.0, + "step": 6305 + }, + { + "epoch": 0.6925104326817483, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1151785850524902, + "learning_rate": 1e-06, + "loss": 1.0627, + "mean_token_accuracy": 0.6818076372146606, + "num_tokens": 163357278.0, + "step": 6306 + }, + { + "epoch": 0.6926202503843619, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.233396053314209, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7103137373924255, + "num_tokens": 163381779.0, + "step": 6307 + }, + { + "epoch": 0.6927300680869756, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.8910008668899536, + "learning_rate": 1e-06, + "loss": 1.0918, + "mean_token_accuracy": 0.6762090921401978, + "num_tokens": 163415146.0, + "step": 6308 + }, + { + "epoch": 0.6928398857895893, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.2250711917877197, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6946672201156616, + "num_tokens": 163440947.0, + "step": 6309 + }, + { + "epoch": 0.692949703492203, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.4421164989471436, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7161476612091064, + "num_tokens": 163460802.0, + "step": 6310 + }, + { + "epoch": 0.6930595211948166, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.8323419094085693, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7005693912506104, + "num_tokens": 163496352.0, + "step": 6311 + }, + { + "epoch": 0.6931693388974303, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.9906079769134521, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.7099392414093018, + "num_tokens": 163526044.0, + "step": 6312 + }, + { + "epoch": 0.6932791566000439, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.3853542804718018, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6910642981529236, + "num_tokens": 163550244.0, + "step": 6313 + }, + { + "epoch": 0.6933889743026576, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.2667155265808105, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.6922752857208252, + "num_tokens": 163576341.0, + "step": 6314 + }, + { + "epoch": 0.6934987920052712, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.0854926109313965, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.7047297954559326, + "num_tokens": 163604671.0, + "step": 6315 + }, + { + "epoch": 0.693608609707885, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.294394016265869, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6975221633911133, + "num_tokens": 163629666.0, + "step": 6316 + }, + { + "epoch": 0.6937184274104986, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.9799880981445312, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7102301120758057, + "num_tokens": 163658034.0, + "step": 6317 + }, + { + "epoch": 0.6938282451131123, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.2048025131225586, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7127238512039185, + "num_tokens": 163683740.0, + "step": 6318 + }, + { + "epoch": 0.6939380628157259, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.2407724857330322, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7277536392211914, + "num_tokens": 163706240.0, + "step": 6319 + }, + { + "epoch": 0.6940478805183395, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.4806911945343018, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7221957445144653, + "num_tokens": 163726391.0, + "step": 6320 + }, + { + "epoch": 0.6941576982209532, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1340646743774414, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7142080664634705, + "num_tokens": 163754046.0, + "step": 6321 + }, + { + "epoch": 0.6942675159235668, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.275616407394409, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.70524001121521, + "num_tokens": 163779277.0, + "step": 6322 + }, + { + "epoch": 0.6943773336261806, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.0138113498687744, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6879132986068726, + "num_tokens": 163811487.0, + "step": 6323 + }, + { + "epoch": 0.6944871513287942, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1169116497039795, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6970378160476685, + "num_tokens": 163837897.0, + "step": 6324 + }, + { + "epoch": 0.6945969690314079, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.092289447784424, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6924457550048828, + "num_tokens": 163866188.0, + "step": 6325 + }, + { + "epoch": 0.6947067867340215, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.0437445640563965, + "learning_rate": 1e-06, + "loss": 1.093, + "mean_token_accuracy": 0.6818177700042725, + "num_tokens": 163897427.0, + "step": 6326 + }, + { + "epoch": 0.6948166044366352, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.0272462368011475, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7190361022949219, + "num_tokens": 163926436.0, + "step": 6327 + }, + { + "epoch": 0.6949264221392488, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1003174781799316, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6898088455200195, + "num_tokens": 163956225.0, + "step": 6328 + }, + { + "epoch": 0.6950362398418625, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.244795560836792, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.7077491879463196, + "num_tokens": 163982151.0, + "step": 6329 + }, + { + "epoch": 0.6951460575444761, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.2427821159362793, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.7000371813774109, + "num_tokens": 164007634.0, + "step": 6330 + }, + { + "epoch": 0.6952558752470899, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 1.9710108041763306, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7126678228378296, + "num_tokens": 164035131.0, + "step": 6331 + }, + { + "epoch": 0.6953656929497035, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.013000965118408, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7005546689033508, + "num_tokens": 164063187.0, + "step": 6332 + }, + { + "epoch": 0.6954755106523172, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.4050302505493164, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7222377061843872, + "num_tokens": 164085148.0, + "step": 6333 + }, + { + "epoch": 0.6955853283549308, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.335303544998169, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6910122036933899, + "num_tokens": 164108696.0, + "step": 6334 + }, + { + "epoch": 0.6956951460575445, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.3197402954101562, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7164761424064636, + "num_tokens": 164132496.0, + "step": 6335 + }, + { + "epoch": 0.6958049637601581, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.200226306915283, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7076038122177124, + "num_tokens": 164158020.0, + "step": 6336 + }, + { + "epoch": 0.6959147814627717, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.4382858276367188, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7185758352279663, + "num_tokens": 164179998.0, + "step": 6337 + }, + { + "epoch": 0.6960245991653855, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.7581119537353516, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7274023294448853, + "num_tokens": 164197247.0, + "step": 6338 + }, + { + "epoch": 0.6961344168679992, + "ewc_loss": 5.7697296142578125e-05, + "grad_norm": 2.1453969478607178, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7078317403793335, + "num_tokens": 164222013.0, + "step": 6339 + }, + { + "epoch": 0.6962442345706128, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.2660512924194336, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7213125824928284, + "num_tokens": 164244311.0, + "step": 6340 + }, + { + "epoch": 0.6963540522732264, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.0398142337799072, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7043178677558899, + "num_tokens": 164272798.0, + "step": 6341 + }, + { + "epoch": 0.6964638699758401, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.159919261932373, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7078356742858887, + "num_tokens": 164298367.0, + "step": 6342 + }, + { + "epoch": 0.6965736876784537, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.104736566543579, + "learning_rate": 1e-06, + "loss": 1.0443, + "mean_token_accuracy": 0.6857361793518066, + "num_tokens": 164328980.0, + "step": 6343 + }, + { + "epoch": 0.6966835053810674, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.257251262664795, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7023425102233887, + "num_tokens": 164354971.0, + "step": 6344 + }, + { + "epoch": 0.6967933230836811, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 1.9969731569290161, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7101996541023254, + "num_tokens": 164382484.0, + "step": 6345 + }, + { + "epoch": 0.6969031407862948, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.3882038593292236, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7129055857658386, + "num_tokens": 164406329.0, + "step": 6346 + }, + { + "epoch": 0.6970129584889084, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 1.9463711977005005, + "learning_rate": 1e-06, + "loss": 1.1012, + "mean_token_accuracy": 0.6782001852989197, + "num_tokens": 164438435.0, + "step": 6347 + }, + { + "epoch": 0.6971227761915221, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.5297744274139404, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7012964487075806, + "num_tokens": 164458132.0, + "step": 6348 + }, + { + "epoch": 0.6972325938941357, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.2205843925476074, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7187157273292542, + "num_tokens": 164483148.0, + "step": 6349 + }, + { + "epoch": 0.6973424115967494, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.24569034576416, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7187121510505676, + "num_tokens": 164507414.0, + "step": 6350 + }, + { + "epoch": 0.697452229299363, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.037383794784546, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7082400918006897, + "num_tokens": 164535855.0, + "step": 6351 + }, + { + "epoch": 0.6975620470019768, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.333366870880127, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7144515514373779, + "num_tokens": 164558183.0, + "step": 6352 + }, + { + "epoch": 0.6976718647045904, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.108203411102295, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7091379165649414, + "num_tokens": 164584290.0, + "step": 6353 + }, + { + "epoch": 0.6977816824072041, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.6724209785461426, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7148522138595581, + "num_tokens": 164601322.0, + "step": 6354 + }, + { + "epoch": 0.6978915001098177, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.839185953140259, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7316220998764038, + "num_tokens": 164618813.0, + "step": 6355 + }, + { + "epoch": 0.6980013178124314, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.1986043453216553, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7175905704498291, + "num_tokens": 164642036.0, + "step": 6356 + }, + { + "epoch": 0.698111135515045, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 1.922452688217163, + "learning_rate": 1e-06, + "loss": 0.9908, + "mean_token_accuracy": 0.7119832038879395, + "num_tokens": 164674367.0, + "step": 6357 + }, + { + "epoch": 0.6982209532176586, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.2692675590515137, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7123650312423706, + "num_tokens": 164696694.0, + "step": 6358 + }, + { + "epoch": 0.6983307709202724, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.127711534500122, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7305333018302917, + "num_tokens": 164722792.0, + "step": 6359 + }, + { + "epoch": 0.698440588622886, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.2226593494415283, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7121750116348267, + "num_tokens": 164748014.0, + "step": 6360 + }, + { + "epoch": 0.6985504063254997, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.233086109161377, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6962241530418396, + "num_tokens": 164772019.0, + "step": 6361 + }, + { + "epoch": 0.6986602240281133, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.0724740028381348, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.6973082423210144, + "num_tokens": 164800535.0, + "step": 6362 + }, + { + "epoch": 0.698770041730727, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.1744487285614014, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6853275895118713, + "num_tokens": 164827307.0, + "step": 6363 + }, + { + "epoch": 0.6988798594333406, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.4698472023010254, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7185152769088745, + "num_tokens": 164850156.0, + "step": 6364 + }, + { + "epoch": 0.6989896771359543, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.208737850189209, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7089400887489319, + "num_tokens": 164876107.0, + "step": 6365 + }, + { + "epoch": 0.6990994948385679, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.003075122833252, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7023929953575134, + "num_tokens": 164905536.0, + "step": 6366 + }, + { + "epoch": 0.6992093125411817, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.245248794555664, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7197849750518799, + "num_tokens": 164930711.0, + "step": 6367 + }, + { + "epoch": 0.6993191302437953, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 1.931682825088501, + "learning_rate": 1e-06, + "loss": 1.0729, + "mean_token_accuracy": 0.6785391569137573, + "num_tokens": 164965215.0, + "step": 6368 + }, + { + "epoch": 0.699428947946409, + "ewc_loss": 5.793571472167969e-05, + "grad_norm": 2.428344964981079, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7023188471794128, + "num_tokens": 164987382.0, + "step": 6369 + }, + { + "epoch": 0.6995387656490226, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.196347713470459, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6885824203491211, + "num_tokens": 165016576.0, + "step": 6370 + }, + { + "epoch": 0.6996485833516363, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1657137870788574, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6971651911735535, + "num_tokens": 165043318.0, + "step": 6371 + }, + { + "epoch": 0.6997584010542499, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.5174553394317627, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7038697004318237, + "num_tokens": 165063912.0, + "step": 6372 + }, + { + "epoch": 0.6998682187568636, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.090010166168213, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.7002630233764648, + "num_tokens": 165091183.0, + "step": 6373 + }, + { + "epoch": 0.6999780364594773, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1220498085021973, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7217718362808228, + "num_tokens": 165117875.0, + "step": 6374 + }, + { + "epoch": 0.700087854162091, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.466031312942505, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7221620082855225, + "num_tokens": 165139314.0, + "step": 6375 + }, + { + "epoch": 0.7001976718647046, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.281001091003418, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7155196666717529, + "num_tokens": 165164950.0, + "step": 6376 + }, + { + "epoch": 0.7003074895673183, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.245356321334839, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7108150720596313, + "num_tokens": 165190399.0, + "step": 6377 + }, + { + "epoch": 0.7004173072699319, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.4962399005889893, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.7079777121543884, + "num_tokens": 165213424.0, + "step": 6378 + }, + { + "epoch": 0.7005271249725455, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1701502799987793, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7133477926254272, + "num_tokens": 165239322.0, + "step": 6379 + }, + { + "epoch": 0.7006369426751592, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 1.8239812850952148, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6778650283813477, + "num_tokens": 165276415.0, + "step": 6380 + }, + { + "epoch": 0.700746760377773, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1569035053253174, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6867740154266357, + "num_tokens": 165304440.0, + "step": 6381 + }, + { + "epoch": 0.7008565780803866, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.257219076156616, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.716856837272644, + "num_tokens": 165329692.0, + "step": 6382 + }, + { + "epoch": 0.7009663957830002, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 1.9822694063186646, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7088545560836792, + "num_tokens": 165361978.0, + "step": 6383 + }, + { + "epoch": 0.7010762134856139, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.03532075881958, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7156428098678589, + "num_tokens": 165391661.0, + "step": 6384 + }, + { + "epoch": 0.7011860311882275, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.4484119415283203, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7069483995437622, + "num_tokens": 165412857.0, + "step": 6385 + }, + { + "epoch": 0.7012958488908412, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.4483773708343506, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7285804748535156, + "num_tokens": 165434553.0, + "step": 6386 + }, + { + "epoch": 0.7014056665934548, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.2193968296051025, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7229031920433044, + "num_tokens": 165458779.0, + "step": 6387 + }, + { + "epoch": 0.7015154842960686, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.0319297313690186, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7271926403045654, + "num_tokens": 165484387.0, + "step": 6388 + }, + { + "epoch": 0.7016253019986822, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.238691806793213, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7117030620574951, + "num_tokens": 165508058.0, + "step": 6389 + }, + { + "epoch": 0.7017351197012959, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.138617515563965, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6973392963409424, + "num_tokens": 165534481.0, + "step": 6390 + }, + { + "epoch": 0.7018449374039095, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.0510711669921875, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7060095071792603, + "num_tokens": 165562167.0, + "step": 6391 + }, + { + "epoch": 0.7019547551065232, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.6315109729766846, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7286553978919983, + "num_tokens": 165580369.0, + "step": 6392 + }, + { + "epoch": 0.7020645728091368, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.151778221130371, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7015731334686279, + "num_tokens": 165606836.0, + "step": 6393 + }, + { + "epoch": 0.7021743905117505, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.4251937866210938, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.702863872051239, + "num_tokens": 165629360.0, + "step": 6394 + }, + { + "epoch": 0.7022842082143641, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.307229518890381, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6942035555839539, + "num_tokens": 165653418.0, + "step": 6395 + }, + { + "epoch": 0.7023940259169779, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1468281745910645, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7324045896530151, + "num_tokens": 165678775.0, + "step": 6396 + }, + { + "epoch": 0.7025038436195915, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.163576364517212, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.711034893989563, + "num_tokens": 165704460.0, + "step": 6397 + }, + { + "epoch": 0.7026136613222052, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1897778511047363, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7181457281112671, + "num_tokens": 165731850.0, + "step": 6398 + }, + { + "epoch": 0.7027234790248188, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2111544609069824, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6989049315452576, + "num_tokens": 165757577.0, + "step": 6399 + }, + { + "epoch": 0.7028332967274324, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.088347911834717, + "learning_rate": 1e-06, + "loss": 1.0797, + "mean_token_accuracy": 0.677595317363739, + "num_tokens": 165787726.0, + "step": 6400 + }, + { + "epoch": 0.7029431144300461, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.411766529083252, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7061201333999634, + "num_tokens": 165810931.0, + "step": 6401 + }, + { + "epoch": 0.7030529321326597, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2387406826019287, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6930408477783203, + "num_tokens": 165837238.0, + "step": 6402 + }, + { + "epoch": 0.7031627498352735, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.9771984815597534, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7074775099754333, + "num_tokens": 165867065.0, + "step": 6403 + }, + { + "epoch": 0.7032725675378871, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1115105152130127, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.6945560574531555, + "num_tokens": 165895370.0, + "step": 6404 + }, + { + "epoch": 0.7033823852405008, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.9591431617736816, + "learning_rate": 1e-06, + "loss": 1.0355, + "mean_token_accuracy": 0.6935535073280334, + "num_tokens": 165928520.0, + "step": 6405 + }, + { + "epoch": 0.7034922029431144, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2935421466827393, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6912570595741272, + "num_tokens": 165952076.0, + "step": 6406 + }, + { + "epoch": 0.7036020206457281, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1153616905212402, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.705933153629303, + "num_tokens": 165978496.0, + "step": 6407 + }, + { + "epoch": 0.7037118383483417, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2878692150115967, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7046545743942261, + "num_tokens": 166000883.0, + "step": 6408 + }, + { + "epoch": 0.7038216560509554, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.04530930519104, + "learning_rate": 1e-06, + "loss": 1.0621, + "mean_token_accuracy": 0.6829365491867065, + "num_tokens": 166029905.0, + "step": 6409 + }, + { + "epoch": 0.7039314737535691, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2712199687957764, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.703447699546814, + "num_tokens": 166052445.0, + "step": 6410 + }, + { + "epoch": 0.7040412914561828, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1703107357025146, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.7010589241981506, + "num_tokens": 166080139.0, + "step": 6411 + }, + { + "epoch": 0.7041511091587964, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1403918266296387, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7069268822669983, + "num_tokens": 166109391.0, + "step": 6412 + }, + { + "epoch": 0.7042609268614101, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.0938491821289062, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6796569228172302, + "num_tokens": 166138729.0, + "step": 6413 + }, + { + "epoch": 0.7043707445640237, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.328162908554077, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6856757998466492, + "num_tokens": 166164711.0, + "step": 6414 + }, + { + "epoch": 0.7044805622666374, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.744023561477661, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7211707234382629, + "num_tokens": 166182875.0, + "step": 6415 + }, + { + "epoch": 0.704590379969251, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.227198839187622, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7063237428665161, + "num_tokens": 166209426.0, + "step": 6416 + }, + { + "epoch": 0.7047001976718648, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2276718616485596, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6842058300971985, + "num_tokens": 166234077.0, + "step": 6417 + }, + { + "epoch": 0.7048100153744784, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.055736780166626, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.703373908996582, + "num_tokens": 166263591.0, + "step": 6418 + }, + { + "epoch": 0.704919833077092, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1972150802612305, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7109584808349609, + "num_tokens": 166287278.0, + "step": 6419 + }, + { + "epoch": 0.7050296507797057, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3498454093933105, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7333576679229736, + "num_tokens": 166309647.0, + "step": 6420 + }, + { + "epoch": 0.7051394684823193, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1414549350738525, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.6876842975616455, + "num_tokens": 166336724.0, + "step": 6421 + }, + { + "epoch": 0.705249286184933, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3836236000061035, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7057569622993469, + "num_tokens": 166361211.0, + "step": 6422 + }, + { + "epoch": 0.7053591038875466, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.07863450050354, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7293835878372192, + "num_tokens": 166385822.0, + "step": 6423 + }, + { + "epoch": 0.7054689215901603, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2812795639038086, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7456594109535217, + "num_tokens": 166407055.0, + "step": 6424 + }, + { + "epoch": 0.705578739292774, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2610225677490234, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6984968185424805, + "num_tokens": 166430708.0, + "step": 6425 + }, + { + "epoch": 0.7056885569953877, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.1441352367401123, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7045985460281372, + "num_tokens": 166456779.0, + "step": 6426 + }, + { + "epoch": 0.7057983746980013, + "ewc_loss": 5.817413330078125e-05, + "grad_norm": 2.152174472808838, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7105188369750977, + "num_tokens": 166481233.0, + "step": 6427 + }, + { + "epoch": 0.705908192400615, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.514315605163574, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.72023606300354, + "num_tokens": 166501202.0, + "step": 6428 + }, + { + "epoch": 0.7060180101032286, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.949358344078064, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7022774815559387, + "num_tokens": 166531152.0, + "step": 6429 + }, + { + "epoch": 0.7061278278058423, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3306984901428223, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.717121958732605, + "num_tokens": 166553589.0, + "step": 6430 + }, + { + "epoch": 0.7062376455084559, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.208240032196045, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7118018865585327, + "num_tokens": 166578512.0, + "step": 6431 + }, + { + "epoch": 0.7063474632110697, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.464970827102661, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7144264578819275, + "num_tokens": 166599150.0, + "step": 6432 + }, + { + "epoch": 0.7064572809136833, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2127504348754883, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6883345246315002, + "num_tokens": 166626449.0, + "step": 6433 + }, + { + "epoch": 0.706567098616297, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.418065071105957, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7240879535675049, + "num_tokens": 166648587.0, + "step": 6434 + }, + { + "epoch": 0.7066769163189106, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3323028087615967, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7220669984817505, + "num_tokens": 166670674.0, + "step": 6435 + }, + { + "epoch": 0.7067867340215243, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3317711353302, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7074435353279114, + "num_tokens": 166695813.0, + "step": 6436 + }, + { + "epoch": 0.7068965517241379, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.4616472721099854, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7164779901504517, + "num_tokens": 166716119.0, + "step": 6437 + }, + { + "epoch": 0.7070063694267515, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.9410045146942139, + "learning_rate": 1e-06, + "loss": 1.0905, + "mean_token_accuracy": 0.6738440990447998, + "num_tokens": 166750666.0, + "step": 6438 + }, + { + "epoch": 0.7071161871293653, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.391284227371216, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7072693109512329, + "num_tokens": 166772775.0, + "step": 6439 + }, + { + "epoch": 0.707226004831979, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.41626238822937, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7231618165969849, + "num_tokens": 166792522.0, + "step": 6440 + }, + { + "epoch": 0.7073358225345926, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.0824649333953857, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7019188404083252, + "num_tokens": 166821634.0, + "step": 6441 + }, + { + "epoch": 0.7074456402372062, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.266159772872925, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6943943500518799, + "num_tokens": 166847977.0, + "step": 6442 + }, + { + "epoch": 0.7075554579398199, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2974138259887695, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.7023117542266846, + "num_tokens": 166871906.0, + "step": 6443 + }, + { + "epoch": 0.7076652756424335, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.8729852437973022, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7048414349555969, + "num_tokens": 166903174.0, + "step": 6444 + }, + { + "epoch": 0.7077750933450472, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1189420223236084, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7259660363197327, + "num_tokens": 166927652.0, + "step": 6445 + }, + { + "epoch": 0.7078849110476609, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2081429958343506, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7091164588928223, + "num_tokens": 166951923.0, + "step": 6446 + }, + { + "epoch": 0.7079947287502746, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.484462261199951, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7111859321594238, + "num_tokens": 166973547.0, + "step": 6447 + }, + { + "epoch": 0.7081045464528882, + "ewc_loss": 5.8650970458984375e-05, + "grad_norm": 2.5330381393432617, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7279335856437683, + "num_tokens": 166992928.0, + "step": 6448 + }, + { + "epoch": 0.7082143641555019, + "ewc_loss": 5.8650970458984375e-05, + "grad_norm": 2.3448903560638428, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7184803485870361, + "num_tokens": 167014443.0, + "step": 6449 + }, + { + "epoch": 0.7083241818581155, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.09926176071167, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.7015535831451416, + "num_tokens": 167040164.0, + "step": 6450 + }, + { + "epoch": 0.7084339995607292, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1087846755981445, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.6990379095077515, + "num_tokens": 167065933.0, + "step": 6451 + }, + { + "epoch": 0.7085438172633428, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.6412150859832764, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7137253880500793, + "num_tokens": 167083429.0, + "step": 6452 + }, + { + "epoch": 0.7086536349659565, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2108662128448486, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7104119062423706, + "num_tokens": 167108697.0, + "step": 6453 + }, + { + "epoch": 0.7087634526685702, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 1.9784027338027954, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7171535491943359, + "num_tokens": 167138333.0, + "step": 6454 + }, + { + "epoch": 0.7088732703711839, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.070678234100342, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7025492191314697, + "num_tokens": 167166755.0, + "step": 6455 + }, + { + "epoch": 0.7089830880737975, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.3603317737579346, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7160897254943848, + "num_tokens": 167189785.0, + "step": 6456 + }, + { + "epoch": 0.7090929057764112, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.5371787548065186, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7206845283508301, + "num_tokens": 167211804.0, + "step": 6457 + }, + { + "epoch": 0.7092027234790248, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.4128856658935547, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.6982265114784241, + "num_tokens": 167233908.0, + "step": 6458 + }, + { + "epoch": 0.7093125411816384, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2732248306274414, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7096723318099976, + "num_tokens": 167258245.0, + "step": 6459 + }, + { + "epoch": 0.7094223588842521, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.1279232501983643, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.6987042427062988, + "num_tokens": 167284185.0, + "step": 6460 + }, + { + "epoch": 0.7095321765868658, + "ewc_loss": 5.841255187988281e-05, + "grad_norm": 2.2369384765625, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7101356983184814, + "num_tokens": 167309018.0, + "step": 6461 + }, + { + "epoch": 0.7096419942894795, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.9763973951339722, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7118297815322876, + "num_tokens": 167337963.0, + "step": 6462 + }, + { + "epoch": 0.7097518119920931, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.9422600269317627, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.685874342918396, + "num_tokens": 167368817.0, + "step": 6463 + }, + { + "epoch": 0.7098616296947068, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.1028497219085693, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7068969011306763, + "num_tokens": 167395048.0, + "step": 6464 + }, + { + "epoch": 0.7099714473973204, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.9505345821380615, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7102743983268738, + "num_tokens": 167424637.0, + "step": 6465 + }, + { + "epoch": 0.7100812650999341, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.841563105583191, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6835331916809082, + "num_tokens": 167459751.0, + "step": 6466 + }, + { + "epoch": 0.7101910828025477, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.1358189582824707, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7192094922065735, + "num_tokens": 167484847.0, + "step": 6467 + }, + { + "epoch": 0.7103009005051615, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2215735912323, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7141628265380859, + "num_tokens": 167507780.0, + "step": 6468 + }, + { + "epoch": 0.7104107182077751, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2878715991973877, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7205904722213745, + "num_tokens": 167530304.0, + "step": 6469 + }, + { + "epoch": 0.7105205359103888, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.4441492557525635, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6970008611679077, + "num_tokens": 167551363.0, + "step": 6470 + }, + { + "epoch": 0.7106303536130024, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.0377867221832275, + "learning_rate": 1e-06, + "loss": 1.0394, + "mean_token_accuracy": 0.6845657229423523, + "num_tokens": 167582669.0, + "step": 6471 + }, + { + "epoch": 0.7107401713156161, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.91107976436615, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6903937458992004, + "num_tokens": 167613850.0, + "step": 6472 + }, + { + "epoch": 0.7108499890182297, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.9035104513168335, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7058767080307007, + "num_tokens": 167646709.0, + "step": 6473 + }, + { + "epoch": 0.7109598067208434, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.032832384109497, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.7126251459121704, + "num_tokens": 167675359.0, + "step": 6474 + }, + { + "epoch": 0.7110696244234571, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.330569267272949, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7024185061454773, + "num_tokens": 167698157.0, + "step": 6475 + }, + { + "epoch": 0.7111794421260708, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.11922287940979, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6974148750305176, + "num_tokens": 167725839.0, + "step": 6476 + }, + { + "epoch": 0.7112892598286844, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2359397411346436, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7255640029907227, + "num_tokens": 167749767.0, + "step": 6477 + }, + { + "epoch": 0.711399077531298, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.016862154006958, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6958762407302856, + "num_tokens": 167777055.0, + "step": 6478 + }, + { + "epoch": 0.7115088952339117, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.3064045906066895, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7271429300308228, + "num_tokens": 167799343.0, + "step": 6479 + }, + { + "epoch": 0.7116187129365253, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2439911365509033, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7225338220596313, + "num_tokens": 167822339.0, + "step": 6480 + }, + { + "epoch": 0.711728530639139, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.139960527420044, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.727448582649231, + "num_tokens": 167847874.0, + "step": 6481 + }, + { + "epoch": 0.7118383483417526, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2812728881835938, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.6951860189437866, + "num_tokens": 167871087.0, + "step": 6482 + }, + { + "epoch": 0.7119481660443664, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.1234707832336426, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6931546330451965, + "num_tokens": 167898170.0, + "step": 6483 + }, + { + "epoch": 0.71205798374698, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.159636974334717, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7073958516120911, + "num_tokens": 167923656.0, + "step": 6484 + }, + { + "epoch": 0.7121678014495937, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.174173355102539, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6871829628944397, + "num_tokens": 167949266.0, + "step": 6485 + }, + { + "epoch": 0.7122776191522073, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2054450511932373, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7095912098884583, + "num_tokens": 167972994.0, + "step": 6486 + }, + { + "epoch": 0.712387436854821, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2701964378356934, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7021011710166931, + "num_tokens": 167997392.0, + "step": 6487 + }, + { + "epoch": 0.7124972545574346, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.330185651779175, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6964439153671265, + "num_tokens": 168022751.0, + "step": 6488 + }, + { + "epoch": 0.7126070722600483, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.0837252140045166, + "learning_rate": 1e-06, + "loss": 1.0661, + "mean_token_accuracy": 0.6867128014564514, + "num_tokens": 168052042.0, + "step": 6489 + }, + { + "epoch": 0.712716889962662, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.618709087371826, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7080996036529541, + "num_tokens": 168071391.0, + "step": 6490 + }, + { + "epoch": 0.7128267076652757, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.1116979122161865, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.698257565498352, + "num_tokens": 168102465.0, + "step": 6491 + }, + { + "epoch": 0.7129365253678893, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.6593000888824463, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7006393671035767, + "num_tokens": 168122278.0, + "step": 6492 + }, + { + "epoch": 0.713046343070503, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.0636162757873535, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.7090588808059692, + "num_tokens": 168151102.0, + "step": 6493 + }, + { + "epoch": 0.7131561607731166, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2664244174957275, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7334995269775391, + "num_tokens": 168175128.0, + "step": 6494 + }, + { + "epoch": 0.7132659784757303, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 1.9917855262756348, + "learning_rate": 1e-06, + "loss": 1.0584, + "mean_token_accuracy": 0.676398515701294, + "num_tokens": 168207639.0, + "step": 6495 + }, + { + "epoch": 0.7133757961783439, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.234480857849121, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7063828706741333, + "num_tokens": 168233243.0, + "step": 6496 + }, + { + "epoch": 0.7134856138809577, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.22633957862854, + "learning_rate": 1e-06, + "loss": 1.0807, + "mean_token_accuracy": 0.6787077784538269, + "num_tokens": 168260065.0, + "step": 6497 + }, + { + "epoch": 0.7135954315835713, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2654049396514893, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7222483158111572, + "num_tokens": 168284638.0, + "step": 6498 + }, + { + "epoch": 0.713705249286185, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.279202699661255, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7064920663833618, + "num_tokens": 168310957.0, + "step": 6499 + }, + { + "epoch": 0.7138150669887986, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.2873001098632812, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.684646725654602, + "num_tokens": 168336000.0, + "step": 6500 + }, + { + "epoch": 0.7139248846914122, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.4365170001983643, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7114847898483276, + "num_tokens": 168358893.0, + "step": 6501 + }, + { + "epoch": 0.7140347023940259, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.241532802581787, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7268206477165222, + "num_tokens": 168382930.0, + "step": 6502 + }, + { + "epoch": 0.7141445200966395, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.1355652809143066, + "learning_rate": 1e-06, + "loss": 1.0277, + "mean_token_accuracy": 0.6919653415679932, + "num_tokens": 168410606.0, + "step": 6503 + }, + { + "epoch": 0.7142543377992533, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.0226731300354004, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7081009149551392, + "num_tokens": 168441184.0, + "step": 6504 + }, + { + "epoch": 0.7143641555018669, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.4422812461853027, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7173738479614258, + "num_tokens": 168464355.0, + "step": 6505 + }, + { + "epoch": 0.7144739732044806, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.0541834831237793, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7102570533752441, + "num_tokens": 168491266.0, + "step": 6506 + }, + { + "epoch": 0.7145837909070942, + "ewc_loss": 5.888938903808594e-05, + "grad_norm": 2.3900649547576904, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6938956379890442, + "num_tokens": 168516585.0, + "step": 6507 + }, + { + "epoch": 0.7146936086097079, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.777489423751831, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6897944211959839, + "num_tokens": 168535196.0, + "step": 6508 + }, + { + "epoch": 0.7148034263123215, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.124732732772827, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7106962203979492, + "num_tokens": 168561111.0, + "step": 6509 + }, + { + "epoch": 0.7149132440149352, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.7443909645080566, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.725983738899231, + "num_tokens": 168578428.0, + "step": 6510 + }, + { + "epoch": 0.7150230617175488, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.0485410690307617, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.708767294883728, + "num_tokens": 168607648.0, + "step": 6511 + }, + { + "epoch": 0.7151328794201626, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.0761125087738037, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6957421898841858, + "num_tokens": 168636005.0, + "step": 6512 + }, + { + "epoch": 0.7152426971227762, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.0883731842041016, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6966426968574524, + "num_tokens": 168664419.0, + "step": 6513 + }, + { + "epoch": 0.7153525148253899, + "ewc_loss": 5.91278076171875e-05, + "grad_norm": 2.2763845920562744, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7274169921875, + "num_tokens": 168687639.0, + "step": 6514 + }, + { + "epoch": 0.7154623325280035, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2711634635925293, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7057031393051147, + "num_tokens": 168712456.0, + "step": 6515 + }, + { + "epoch": 0.7155721502306172, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.054596424102783, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7218904495239258, + "num_tokens": 168740946.0, + "step": 6516 + }, + { + "epoch": 0.7156819679332308, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.0832715034484863, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.688983142375946, + "num_tokens": 168769603.0, + "step": 6517 + }, + { + "epoch": 0.7157917856358444, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.168121814727783, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7269275188446045, + "num_tokens": 168795783.0, + "step": 6518 + }, + { + "epoch": 0.7159016033384582, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.7390427589416504, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7321230173110962, + "num_tokens": 168813454.0, + "step": 6519 + }, + { + "epoch": 0.7160114210410718, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.4406745433807373, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7205003499984741, + "num_tokens": 168834759.0, + "step": 6520 + }, + { + "epoch": 0.7161212387436855, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1766626834869385, + "learning_rate": 1e-06, + "loss": 1.0509, + "mean_token_accuracy": 0.6874921917915344, + "num_tokens": 168861563.0, + "step": 6521 + }, + { + "epoch": 0.7162310564462991, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1165952682495117, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7065747976303101, + "num_tokens": 168888884.0, + "step": 6522 + }, + { + "epoch": 0.7163408741489128, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.033005952835083, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.70769202709198, + "num_tokens": 168917774.0, + "step": 6523 + }, + { + "epoch": 0.7164506918515264, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.592294454574585, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7137901186943054, + "num_tokens": 168936506.0, + "step": 6524 + }, + { + "epoch": 0.7165605095541401, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.151921272277832, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.697097659111023, + "num_tokens": 168963308.0, + "step": 6525 + }, + { + "epoch": 0.7166703272567538, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.477396249771118, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.727769136428833, + "num_tokens": 168985502.0, + "step": 6526 + }, + { + "epoch": 0.7167801449593675, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.376573324203491, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.737899661064148, + "num_tokens": 169006486.0, + "step": 6527 + }, + { + "epoch": 0.7168899626619811, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.261932373046875, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7012753486633301, + "num_tokens": 169030895.0, + "step": 6528 + }, + { + "epoch": 0.7169997803645948, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.348393440246582, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7033050060272217, + "num_tokens": 169054078.0, + "step": 6529 + }, + { + "epoch": 0.7171095980672084, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.54252290725708, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7195436954498291, + "num_tokens": 169072668.0, + "step": 6530 + }, + { + "epoch": 0.7172194157698221, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1648900508880615, + "learning_rate": 1e-06, + "loss": 1.0365, + "mean_token_accuracy": 0.6944707036018372, + "num_tokens": 169099329.0, + "step": 6531 + }, + { + "epoch": 0.7173292334724357, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.449883222579956, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7126666307449341, + "num_tokens": 169120903.0, + "step": 6532 + }, + { + "epoch": 0.7174390511750495, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.112720251083374, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6889860033988953, + "num_tokens": 169149640.0, + "step": 6533 + }, + { + "epoch": 0.7175488688776631, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.4485421180725098, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7126296758651733, + "num_tokens": 169170780.0, + "step": 6534 + }, + { + "epoch": 0.7176586865802768, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.075361967086792, + "learning_rate": 1e-06, + "loss": 1.0485, + "mean_token_accuracy": 0.6922774314880371, + "num_tokens": 169200303.0, + "step": 6535 + }, + { + "epoch": 0.7177685042828904, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.9677270650863647, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7260399460792542, + "num_tokens": 169229281.0, + "step": 6536 + }, + { + "epoch": 0.717878321985504, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.4194393157958984, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7233307361602783, + "num_tokens": 169252296.0, + "step": 6537 + }, + { + "epoch": 0.7179881396881177, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.267132043838501, + "learning_rate": 1e-06, + "loss": 1.0426, + "mean_token_accuracy": 0.6897217631340027, + "num_tokens": 169279224.0, + "step": 6538 + }, + { + "epoch": 0.7180979573907313, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.128558874130249, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6999363899230957, + "num_tokens": 169307045.0, + "step": 6539 + }, + { + "epoch": 0.7182077750933451, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1988465785980225, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7058424353599548, + "num_tokens": 169333309.0, + "step": 6540 + }, + { + "epoch": 0.7183175927959587, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.117661714553833, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6906384229660034, + "num_tokens": 169362714.0, + "step": 6541 + }, + { + "epoch": 0.7184274104985724, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.222961187362671, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.6944476366043091, + "num_tokens": 169390450.0, + "step": 6542 + }, + { + "epoch": 0.718537228201186, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.0664896965026855, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.6959670186042786, + "num_tokens": 169419864.0, + "step": 6543 + }, + { + "epoch": 0.7186470459037997, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.3077728748321533, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7371290326118469, + "num_tokens": 169441594.0, + "step": 6544 + }, + { + "epoch": 0.7187568636064133, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.170957088470459, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7037311792373657, + "num_tokens": 169467608.0, + "step": 6545 + }, + { + "epoch": 0.718866681309027, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1518070697784424, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6984618902206421, + "num_tokens": 169494768.0, + "step": 6546 + }, + { + "epoch": 0.7189764990116406, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.069122791290283, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.701218843460083, + "num_tokens": 169521616.0, + "step": 6547 + }, + { + "epoch": 0.7190863167142544, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.4479565620422363, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7216600179672241, + "num_tokens": 169542364.0, + "step": 6548 + }, + { + "epoch": 0.719196134416868, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.062634229660034, + "learning_rate": 1e-06, + "loss": 1.0652, + "mean_token_accuracy": 0.6781805753707886, + "num_tokens": 169574336.0, + "step": 6549 + }, + { + "epoch": 0.7193059521194817, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.9175523519515991, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7033287286758423, + "num_tokens": 169602383.0, + "step": 6550 + }, + { + "epoch": 0.7194157698220953, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2510015964508057, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7053052186965942, + "num_tokens": 169625680.0, + "step": 6551 + }, + { + "epoch": 0.719525587524709, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.234861135482788, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7142907381057739, + "num_tokens": 169649483.0, + "step": 6552 + }, + { + "epoch": 0.7196354052273226, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.128173351287842, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7167338728904724, + "num_tokens": 169676176.0, + "step": 6553 + }, + { + "epoch": 0.7197452229299363, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.364595651626587, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7043367624282837, + "num_tokens": 169700230.0, + "step": 6554 + }, + { + "epoch": 0.71985504063255, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2819533348083496, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6987793445587158, + "num_tokens": 169724249.0, + "step": 6555 + }, + { + "epoch": 0.7199648583351637, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.183107852935791, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.700198769569397, + "num_tokens": 169750728.0, + "step": 6556 + }, + { + "epoch": 0.7200746760377773, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.9624532461166382, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.7005718946456909, + "num_tokens": 169782386.0, + "step": 6557 + }, + { + "epoch": 0.720184493740391, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.0215909481048584, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7224123477935791, + "num_tokens": 169809532.0, + "step": 6558 + }, + { + "epoch": 0.7202943114430046, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.0879273414611816, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6985143423080444, + "num_tokens": 169838593.0, + "step": 6559 + }, + { + "epoch": 0.7204041291456182, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1300201416015625, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7121676802635193, + "num_tokens": 169865464.0, + "step": 6560 + }, + { + "epoch": 0.7205139468482319, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.008863925933838, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7245981693267822, + "num_tokens": 169894499.0, + "step": 6561 + }, + { + "epoch": 0.7206237645508456, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.328596830368042, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6953990459442139, + "num_tokens": 169922616.0, + "step": 6562 + }, + { + "epoch": 0.7207335822534593, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1076290607452393, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.6865248680114746, + "num_tokens": 169950672.0, + "step": 6563 + }, + { + "epoch": 0.7208433999560729, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2728168964385986, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.698462188243866, + "num_tokens": 169974535.0, + "step": 6564 + }, + { + "epoch": 0.7209532176586866, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.1661791801452637, + "learning_rate": 1e-06, + "loss": 1.0833, + "mean_token_accuracy": 0.6809070110321045, + "num_tokens": 170002421.0, + "step": 6565 + }, + { + "epoch": 0.7210630353613002, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.940555214881897, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7192962169647217, + "num_tokens": 170032139.0, + "step": 6566 + }, + { + "epoch": 0.7211728530639139, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.458804130554199, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7062090635299683, + "num_tokens": 170051814.0, + "step": 6567 + }, + { + "epoch": 0.7212826707665275, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.9908181428909302, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7138232588768005, + "num_tokens": 170080831.0, + "step": 6568 + }, + { + "epoch": 0.7213924884691413, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.011845350265503, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.701896607875824, + "num_tokens": 170108149.0, + "step": 6569 + }, + { + "epoch": 0.7215023061717549, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.111621141433716, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6858271360397339, + "num_tokens": 170137356.0, + "step": 6570 + }, + { + "epoch": 0.7216121238743686, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.282119035720825, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7113335132598877, + "num_tokens": 170161300.0, + "step": 6571 + }, + { + "epoch": 0.7217219415769822, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.21309232711792, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7273802757263184, + "num_tokens": 170184667.0, + "step": 6572 + }, + { + "epoch": 0.7218317592795959, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 1.9753313064575195, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6925032138824463, + "num_tokens": 170215863.0, + "step": 6573 + }, + { + "epoch": 0.7219415769822095, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2659924030303955, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7132250666618347, + "num_tokens": 170239781.0, + "step": 6574 + }, + { + "epoch": 0.7220513946848232, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.0930821895599365, + "learning_rate": 1e-06, + "loss": 1.0041, + "mean_token_accuracy": 0.6935439705848694, + "num_tokens": 170266272.0, + "step": 6575 + }, + { + "epoch": 0.7221612123874368, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2889199256896973, + "learning_rate": 1e-06, + "loss": 1.0766, + "mean_token_accuracy": 0.6745453476905823, + "num_tokens": 170293997.0, + "step": 6576 + }, + { + "epoch": 0.7222710300900506, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2550861835479736, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7370494604110718, + "num_tokens": 170316894.0, + "step": 6577 + }, + { + "epoch": 0.7223808477926642, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.17844295501709, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7225353717803955, + "num_tokens": 170343035.0, + "step": 6578 + }, + { + "epoch": 0.7224906654952779, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.2555313110351562, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6983209252357483, + "num_tokens": 170368423.0, + "step": 6579 + }, + { + "epoch": 0.7226004831978915, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.157871723175049, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7242923974990845, + "num_tokens": 170393599.0, + "step": 6580 + }, + { + "epoch": 0.7227103009005051, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.16395902633667, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7266043424606323, + "num_tokens": 170419722.0, + "step": 6581 + }, + { + "epoch": 0.7228201186031188, + "ewc_loss": 5.936622619628906e-05, + "grad_norm": 2.290574789047241, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7119219303131104, + "num_tokens": 170443024.0, + "step": 6582 + }, + { + "epoch": 0.7229299363057324, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 1.9884225130081177, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7110804319381714, + "num_tokens": 170471462.0, + "step": 6583 + }, + { + "epoch": 0.7230397540083462, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.0481183528900146, + "learning_rate": 1e-06, + "loss": 1.0622, + "mean_token_accuracy": 0.6841117143630981, + "num_tokens": 170502437.0, + "step": 6584 + }, + { + "epoch": 0.7231495717109598, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.305302619934082, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7193672060966492, + "num_tokens": 170525603.0, + "step": 6585 + }, + { + "epoch": 0.7232593894135735, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.03166127204895, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.7035859227180481, + "num_tokens": 170555958.0, + "step": 6586 + }, + { + "epoch": 0.7233692071161871, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.239743232727051, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7031466364860535, + "num_tokens": 170579904.0, + "step": 6587 + }, + { + "epoch": 0.7234790248188008, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.591583490371704, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7186890244483948, + "num_tokens": 170599073.0, + "step": 6588 + }, + { + "epoch": 0.7235888425214144, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.03537917137146, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7299754023551941, + "num_tokens": 170624812.0, + "step": 6589 + }, + { + "epoch": 0.7236986602240281, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.2959372997283936, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6968007683753967, + "num_tokens": 170649253.0, + "step": 6590 + }, + { + "epoch": 0.7238084779266418, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.3163106441497803, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6941463351249695, + "num_tokens": 170672443.0, + "step": 6591 + }, + { + "epoch": 0.7239182956292555, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.009320020675659, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6934044361114502, + "num_tokens": 170701394.0, + "step": 6592 + }, + { + "epoch": 0.7240281133318691, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.3356964588165283, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7091585993766785, + "num_tokens": 170725227.0, + "step": 6593 + }, + { + "epoch": 0.7241379310344828, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.4386022090911865, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7102741599082947, + "num_tokens": 170747080.0, + "step": 6594 + }, + { + "epoch": 0.7242477487370964, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.143080711364746, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.724617600440979, + "num_tokens": 170769740.0, + "step": 6595 + }, + { + "epoch": 0.72435756643971, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.251781702041626, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.691950798034668, + "num_tokens": 170795664.0, + "step": 6596 + }, + { + "epoch": 0.7244673841423237, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 1.9637254476547241, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7174314260482788, + "num_tokens": 170824411.0, + "step": 6597 + }, + { + "epoch": 0.7245772018449375, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.212252378463745, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6888713240623474, + "num_tokens": 170850123.0, + "step": 6598 + }, + { + "epoch": 0.7246870195475511, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.3598761558532715, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7031944394111633, + "num_tokens": 170873265.0, + "step": 6599 + }, + { + "epoch": 0.7247968372501647, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.1132054328918457, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6982856392860413, + "num_tokens": 170899152.0, + "step": 6600 + }, + { + "epoch": 0.7249066549527784, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.0658280849456787, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6989611387252808, + "num_tokens": 170929558.0, + "step": 6601 + }, + { + "epoch": 0.725016472655392, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.162769317626953, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7194239497184753, + "num_tokens": 170955891.0, + "step": 6602 + }, + { + "epoch": 0.7251262903580057, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.015015125274658, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.7156140208244324, + "num_tokens": 170984853.0, + "step": 6603 + }, + { + "epoch": 0.7252361080606193, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 1.8995351791381836, + "learning_rate": 1e-06, + "loss": 1.0938, + "mean_token_accuracy": 0.6773877739906311, + "num_tokens": 171017608.0, + "step": 6604 + }, + { + "epoch": 0.725345925763233, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.3060379028320312, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.6979550123214722, + "num_tokens": 171043941.0, + "step": 6605 + }, + { + "epoch": 0.7254557434658467, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.3335843086242676, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6947065591812134, + "num_tokens": 171067231.0, + "step": 6606 + }, + { + "epoch": 0.7255655611684604, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.0314104557037354, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6990489959716797, + "num_tokens": 171094751.0, + "step": 6607 + }, + { + "epoch": 0.725675378871074, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.2129428386688232, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6921043992042542, + "num_tokens": 171121793.0, + "step": 6608 + }, + { + "epoch": 0.7257851965736877, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.1949360370635986, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.6970945596694946, + "num_tokens": 171147151.0, + "step": 6609 + }, + { + "epoch": 0.7258950142763013, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.126965045928955, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7009273767471313, + "num_tokens": 171173165.0, + "step": 6610 + }, + { + "epoch": 0.726004831978915, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.2616796493530273, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7240399718284607, + "num_tokens": 171195083.0, + "step": 6611 + }, + { + "epoch": 0.7261146496815286, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 1.920941948890686, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6892237663269043, + "num_tokens": 171227814.0, + "step": 6612 + }, + { + "epoch": 0.7262244673841424, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.1858813762664795, + "learning_rate": 1e-06, + "loss": 1.0041, + "mean_token_accuracy": 0.6919814348220825, + "num_tokens": 171255157.0, + "step": 6613 + }, + { + "epoch": 0.726334285086756, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.1506025791168213, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7095448970794678, + "num_tokens": 171279921.0, + "step": 6614 + }, + { + "epoch": 0.7264441027893697, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.264505386352539, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.690019965171814, + "num_tokens": 171306367.0, + "step": 6615 + }, + { + "epoch": 0.7265539204919833, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.0962018966674805, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.689788281917572, + "num_tokens": 171335470.0, + "step": 6616 + }, + { + "epoch": 0.726663738194597, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.274183511734009, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6961129307746887, + "num_tokens": 171360291.0, + "step": 6617 + }, + { + "epoch": 0.7267735558972106, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.1964681148529053, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.7089931964874268, + "num_tokens": 171386241.0, + "step": 6618 + }, + { + "epoch": 0.7268833735998242, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.1882336139678955, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.6893774271011353, + "num_tokens": 171412587.0, + "step": 6619 + }, + { + "epoch": 0.726993191302438, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.383781671524048, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7047010064125061, + "num_tokens": 171434516.0, + "step": 6620 + }, + { + "epoch": 0.7271030090050516, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.0405330657958984, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.6872997283935547, + "num_tokens": 171464916.0, + "step": 6621 + }, + { + "epoch": 0.7272128267076653, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.268704414367676, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6965951919555664, + "num_tokens": 171489467.0, + "step": 6622 + }, + { + "epoch": 0.7273226444102789, + "ewc_loss": 5.9604644775390625e-05, + "grad_norm": 2.652121067047119, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7392135858535767, + "num_tokens": 171506265.0, + "step": 6623 + }, + { + "epoch": 0.7274324621128926, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.3226354122161865, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7304816842079163, + "num_tokens": 171527160.0, + "step": 6624 + }, + { + "epoch": 0.7275422798155062, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.1437723636627197, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7126033306121826, + "num_tokens": 171552198.0, + "step": 6625 + }, + { + "epoch": 0.7276520975181199, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.459702491760254, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7099054455757141, + "num_tokens": 171573001.0, + "step": 6626 + }, + { + "epoch": 0.7277619152207336, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 1.9795114994049072, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6903089284896851, + "num_tokens": 171602431.0, + "step": 6627 + }, + { + "epoch": 0.7278717329233473, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.0919206142425537, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.7018133401870728, + "num_tokens": 171631809.0, + "step": 6628 + }, + { + "epoch": 0.7279815506259609, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.281121253967285, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7027654647827148, + "num_tokens": 171655463.0, + "step": 6629 + }, + { + "epoch": 0.7280913683285746, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.256458282470703, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7209001779556274, + "num_tokens": 171679233.0, + "step": 6630 + }, + { + "epoch": 0.7282011860311882, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.281916379928589, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7129701375961304, + "num_tokens": 171704875.0, + "step": 6631 + }, + { + "epoch": 0.7283110037338019, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 1.8638321161270142, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7154375314712524, + "num_tokens": 171736957.0, + "step": 6632 + }, + { + "epoch": 0.7284208214364155, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.0726735591888428, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6827985048294067, + "num_tokens": 171765231.0, + "step": 6633 + }, + { + "epoch": 0.7285306391390292, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.346020221710205, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6885390281677246, + "num_tokens": 171790280.0, + "step": 6634 + }, + { + "epoch": 0.7286404568416429, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.020601987838745, + "learning_rate": 1e-06, + "loss": 1.0747, + "mean_token_accuracy": 0.6848032474517822, + "num_tokens": 171823396.0, + "step": 6635 + }, + { + "epoch": 0.7287502745442566, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 1.9707435369491577, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7149648666381836, + "num_tokens": 171852932.0, + "step": 6636 + }, + { + "epoch": 0.7288600922468702, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.0290863513946533, + "learning_rate": 1e-06, + "loss": 1.0772, + "mean_token_accuracy": 0.6772092580795288, + "num_tokens": 171882645.0, + "step": 6637 + }, + { + "epoch": 0.7289699099494839, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.258183479309082, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6885442137718201, + "num_tokens": 171907999.0, + "step": 6638 + }, + { + "epoch": 0.7290797276520975, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.4748973846435547, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7190448045730591, + "num_tokens": 171929437.0, + "step": 6639 + }, + { + "epoch": 0.7291895453547111, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.258553981781006, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.719966471195221, + "num_tokens": 171952768.0, + "step": 6640 + }, + { + "epoch": 0.7292993630573248, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.0043742656707764, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6985515356063843, + "num_tokens": 171982817.0, + "step": 6641 + }, + { + "epoch": 0.7294091807599385, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.1915178298950195, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7048625349998474, + "num_tokens": 172008264.0, + "step": 6642 + }, + { + "epoch": 0.7295189984625522, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.01084566116333, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7001447677612305, + "num_tokens": 172037187.0, + "step": 6643 + }, + { + "epoch": 0.7296288161651658, + "ewc_loss": 5.984306335449219e-05, + "grad_norm": 2.30553936958313, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.70112144947052, + "num_tokens": 172061265.0, + "step": 6644 + }, + { + "epoch": 0.7297386338677795, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.211777687072754, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7118191719055176, + "num_tokens": 172088106.0, + "step": 6645 + }, + { + "epoch": 0.7298484515703931, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2299153804779053, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.7003500461578369, + "num_tokens": 172111278.0, + "step": 6646 + }, + { + "epoch": 0.7299582692730068, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1155498027801514, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6957360506057739, + "num_tokens": 172139161.0, + "step": 6647 + }, + { + "epoch": 0.7300680869756204, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.202859878540039, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6850610971450806, + "num_tokens": 172164497.0, + "step": 6648 + }, + { + "epoch": 0.7301779046782342, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 1.9343764781951904, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.6902076005935669, + "num_tokens": 172195776.0, + "step": 6649 + }, + { + "epoch": 0.7302877223808478, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1146156787872314, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.713749885559082, + "num_tokens": 172222583.0, + "step": 6650 + }, + { + "epoch": 0.7303975400834615, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.237349510192871, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6935076713562012, + "num_tokens": 172247462.0, + "step": 6651 + }, + { + "epoch": 0.7305073577860751, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.090785026550293, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.691824734210968, + "num_tokens": 172275377.0, + "step": 6652 + }, + { + "epoch": 0.7306171754886888, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2760069370269775, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7260497212409973, + "num_tokens": 172297610.0, + "step": 6653 + }, + { + "epoch": 0.7307269931913024, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.152637243270874, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7137001752853394, + "num_tokens": 172322826.0, + "step": 6654 + }, + { + "epoch": 0.7308368108939161, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2972073554992676, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6974348425865173, + "num_tokens": 172347222.0, + "step": 6655 + }, + { + "epoch": 0.7309466285965298, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.0143580436706543, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7272812128067017, + "num_tokens": 172373660.0, + "step": 6656 + }, + { + "epoch": 0.7310564462991435, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 1.9963092803955078, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6943375468254089, + "num_tokens": 172402897.0, + "step": 6657 + }, + { + "epoch": 0.7311662640017571, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.3332698345184326, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7148173451423645, + "num_tokens": 172424908.0, + "step": 6658 + }, + { + "epoch": 0.7312760817043708, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.146754264831543, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7025921940803528, + "num_tokens": 172452343.0, + "step": 6659 + }, + { + "epoch": 0.7313858994069844, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2020444869995117, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7333307266235352, + "num_tokens": 172474687.0, + "step": 6660 + }, + { + "epoch": 0.731495717109598, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.076463222503662, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7166079878807068, + "num_tokens": 172501083.0, + "step": 6661 + }, + { + "epoch": 0.7316055348122117, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.051227331161499, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6946488618850708, + "num_tokens": 172529837.0, + "step": 6662 + }, + { + "epoch": 0.7317153525148253, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 1.8758450746536255, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7009250521659851, + "num_tokens": 172565242.0, + "step": 6663 + }, + { + "epoch": 0.7318251702174391, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.117647886276245, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7241136431694031, + "num_tokens": 172590053.0, + "step": 6664 + }, + { + "epoch": 0.7319349879200527, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 1.9570139646530151, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7175339460372925, + "num_tokens": 172617737.0, + "step": 6665 + }, + { + "epoch": 0.7320448056226664, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1350951194763184, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7134735584259033, + "num_tokens": 172643719.0, + "step": 6666 + }, + { + "epoch": 0.73215462332528, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.0641181468963623, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7120516300201416, + "num_tokens": 172670802.0, + "step": 6667 + }, + { + "epoch": 0.7322644410278937, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 1.9327911138534546, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.7015254497528076, + "num_tokens": 172701329.0, + "step": 6668 + }, + { + "epoch": 0.7323742587305073, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.3571105003356934, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7010143995285034, + "num_tokens": 172723687.0, + "step": 6669 + }, + { + "epoch": 0.732484076433121, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.0013742446899414, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7005499601364136, + "num_tokens": 172752354.0, + "step": 6670 + }, + { + "epoch": 0.7325938941357347, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2863779067993164, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.7004737257957458, + "num_tokens": 172776323.0, + "step": 6671 + }, + { + "epoch": 0.7327037118383484, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2271878719329834, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7158557772636414, + "num_tokens": 172800904.0, + "step": 6672 + }, + { + "epoch": 0.732813529540962, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.2850921154022217, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.713862419128418, + "num_tokens": 172823370.0, + "step": 6673 + }, + { + "epoch": 0.7329233472435757, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.3723361492156982, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.714252769947052, + "num_tokens": 172846258.0, + "step": 6674 + }, + { + "epoch": 0.7330331649461893, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.4577574729919434, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.7052947282791138, + "num_tokens": 172867940.0, + "step": 6675 + }, + { + "epoch": 0.733142982648803, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.6538264751434326, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.704087495803833, + "num_tokens": 172886700.0, + "step": 6676 + }, + { + "epoch": 0.7332528003514166, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1792373657226562, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7048195600509644, + "num_tokens": 172914941.0, + "step": 6677 + }, + { + "epoch": 0.7333626180540304, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.158726692199707, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7189711332321167, + "num_tokens": 172939971.0, + "step": 6678 + }, + { + "epoch": 0.733472435756644, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.16780948638916, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7099648714065552, + "num_tokens": 172968136.0, + "step": 6679 + }, + { + "epoch": 0.7335822534592576, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1878468990325928, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.705647885799408, + "num_tokens": 172993683.0, + "step": 6680 + }, + { + "epoch": 0.7336920711618713, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.280656337738037, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7009936571121216, + "num_tokens": 173018448.0, + "step": 6681 + }, + { + "epoch": 0.7338018888644849, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1857075691223145, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.7020491361618042, + "num_tokens": 173044670.0, + "step": 6682 + }, + { + "epoch": 0.7339117065670986, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.1756460666656494, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.694326639175415, + "num_tokens": 173072158.0, + "step": 6683 + }, + { + "epoch": 0.7340215242697122, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.9714428186416626, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7113297581672668, + "num_tokens": 173101918.0, + "step": 6684 + }, + { + "epoch": 0.734131341972326, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.220452070236206, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6995922923088074, + "num_tokens": 173127160.0, + "step": 6685 + }, + { + "epoch": 0.7342411596749396, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.097073554992676, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6914348602294922, + "num_tokens": 173153811.0, + "step": 6686 + }, + { + "epoch": 0.7343509773775533, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.264603614807129, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.6871224045753479, + "num_tokens": 173178814.0, + "step": 6687 + }, + { + "epoch": 0.7344607950801669, + "ewc_loss": 6.008148193359375e-05, + "grad_norm": 2.3983051776885986, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.6933603286743164, + "num_tokens": 173200290.0, + "step": 6688 + }, + { + "epoch": 0.7345706127827806, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.115802764892578, + "learning_rate": 1e-06, + "loss": 1.0456, + "mean_token_accuracy": 0.6935648918151855, + "num_tokens": 173226366.0, + "step": 6689 + }, + { + "epoch": 0.7346804304853942, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.2730698585510254, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6976466178894043, + "num_tokens": 173253655.0, + "step": 6690 + }, + { + "epoch": 0.7347902481880079, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.5294480323791504, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7198070883750916, + "num_tokens": 173275181.0, + "step": 6691 + }, + { + "epoch": 0.7349000658906216, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1556639671325684, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7105148434638977, + "num_tokens": 173302924.0, + "step": 6692 + }, + { + "epoch": 0.7350098835932353, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.238217353820801, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6997346878051758, + "num_tokens": 173327534.0, + "step": 6693 + }, + { + "epoch": 0.7351197012958489, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.2328953742980957, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7211454510688782, + "num_tokens": 173352614.0, + "step": 6694 + }, + { + "epoch": 0.7352295189984626, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.4447853565216064, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7066525220870972, + "num_tokens": 173374986.0, + "step": 6695 + }, + { + "epoch": 0.7353393367010762, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.6560699939727783, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7043854594230652, + "num_tokens": 173394416.0, + "step": 6696 + }, + { + "epoch": 0.7354491544036899, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.305927038192749, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7183796167373657, + "num_tokens": 173417035.0, + "step": 6697 + }, + { + "epoch": 0.7355589721063035, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.964097023010254, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7164846062660217, + "num_tokens": 173447210.0, + "step": 6698 + }, + { + "epoch": 0.7356687898089171, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1581473350524902, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7042566537857056, + "num_tokens": 173473313.0, + "step": 6699 + }, + { + "epoch": 0.7357786075115309, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.2869417667388916, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7405209541320801, + "num_tokens": 173495252.0, + "step": 6700 + }, + { + "epoch": 0.7358884252141445, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1779494285583496, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7262122631072998, + "num_tokens": 173521442.0, + "step": 6701 + }, + { + "epoch": 0.7359982429167582, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1323461532592773, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7234891653060913, + "num_tokens": 173547300.0, + "step": 6702 + }, + { + "epoch": 0.7361080606193718, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.127318859100342, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7142329216003418, + "num_tokens": 173576514.0, + "step": 6703 + }, + { + "epoch": 0.7362178783219855, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.3708748817443848, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.7025834321975708, + "num_tokens": 173602159.0, + "step": 6704 + }, + { + "epoch": 0.7363276960245991, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.191849946975708, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7103526592254639, + "num_tokens": 173627353.0, + "step": 6705 + }, + { + "epoch": 0.7364375137272128, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.0677926540374756, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7249049544334412, + "num_tokens": 173655914.0, + "step": 6706 + }, + { + "epoch": 0.7365473314298265, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1873555183410645, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7015426158905029, + "num_tokens": 173682089.0, + "step": 6707 + }, + { + "epoch": 0.7366571491324402, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.073756456375122, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7211636304855347, + "num_tokens": 173710899.0, + "step": 6708 + }, + { + "epoch": 0.7367669668350538, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.3656740188598633, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.698724627494812, + "num_tokens": 173734020.0, + "step": 6709 + }, + { + "epoch": 0.7368767845376675, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.987831711769104, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7178685665130615, + "num_tokens": 173763730.0, + "step": 6710 + }, + { + "epoch": 0.7369866022402811, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.3007242679595947, + "learning_rate": 1e-06, + "loss": 1.0603, + "mean_token_accuracy": 0.6733354330062866, + "num_tokens": 173789322.0, + "step": 6711 + }, + { + "epoch": 0.7370964199428948, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.9085839986801147, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6909705400466919, + "num_tokens": 173821838.0, + "step": 6712 + }, + { + "epoch": 0.7372062376455084, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.108030319213867, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7181594967842102, + "num_tokens": 173848132.0, + "step": 6713 + }, + { + "epoch": 0.7373160553481222, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.189772844314575, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7309118509292603, + "num_tokens": 173872793.0, + "step": 6714 + }, + { + "epoch": 0.7374258730507358, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.208430528640747, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7182130813598633, + "num_tokens": 173897476.0, + "step": 6715 + }, + { + "epoch": 0.7375356907533495, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.2700836658477783, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7032546997070312, + "num_tokens": 173921544.0, + "step": 6716 + }, + { + "epoch": 0.7376455084559631, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.4462740421295166, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.718933641910553, + "num_tokens": 173942604.0, + "step": 6717 + }, + { + "epoch": 0.7377553261585768, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.238091230392456, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7069733142852783, + "num_tokens": 173967385.0, + "step": 6718 + }, + { + "epoch": 0.7378651438611904, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.327995777130127, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7163424491882324, + "num_tokens": 173989018.0, + "step": 6719 + }, + { + "epoch": 0.737974961563804, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.203943967819214, + "learning_rate": 1e-06, + "loss": 1.0615, + "mean_token_accuracy": 0.6881066560745239, + "num_tokens": 174014756.0, + "step": 6720 + }, + { + "epoch": 0.7380847792664178, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.52962327003479, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6942932605743408, + "num_tokens": 174036322.0, + "step": 6721 + }, + { + "epoch": 0.7381945969690314, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.098259449005127, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7312757968902588, + "num_tokens": 174061966.0, + "step": 6722 + }, + { + "epoch": 0.7383044146716451, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.5388834476470947, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7009731531143188, + "num_tokens": 174082133.0, + "step": 6723 + }, + { + "epoch": 0.7384142323742587, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.9247468709945679, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7029446363449097, + "num_tokens": 174114587.0, + "step": 6724 + }, + { + "epoch": 0.7385240500768724, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 1.9976496696472168, + "learning_rate": 1e-06, + "loss": 1.0461, + "mean_token_accuracy": 0.6855080723762512, + "num_tokens": 174144910.0, + "step": 6725 + }, + { + "epoch": 0.738633867779486, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.1677446365356445, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7074782848358154, + "num_tokens": 174170044.0, + "step": 6726 + }, + { + "epoch": 0.7387436854820997, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.1495954990386963, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7295993566513062, + "num_tokens": 174194128.0, + "step": 6727 + }, + { + "epoch": 0.7388535031847133, + "ewc_loss": 6.031990051269531e-05, + "grad_norm": 2.4252617359161377, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7309879064559937, + "num_tokens": 174213613.0, + "step": 6728 + }, + { + "epoch": 0.7389633208873271, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.2171573638916016, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7205737829208374, + "num_tokens": 174239072.0, + "step": 6729 + }, + { + "epoch": 0.7390731385899407, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.0834615230560303, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6944701671600342, + "num_tokens": 174266710.0, + "step": 6730 + }, + { + "epoch": 0.7391829562925544, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3274848461151123, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7074925303459167, + "num_tokens": 174289578.0, + "step": 6731 + }, + { + "epoch": 0.739292773995168, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3144443035125732, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.708710789680481, + "num_tokens": 174312357.0, + "step": 6732 + }, + { + "epoch": 0.7394025916977817, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.5075793266296387, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7114884257316589, + "num_tokens": 174333509.0, + "step": 6733 + }, + { + "epoch": 0.7395124094003953, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3562123775482178, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7259449362754822, + "num_tokens": 174354178.0, + "step": 6734 + }, + { + "epoch": 0.739622227103009, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3423335552215576, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7038285136222839, + "num_tokens": 174377254.0, + "step": 6735 + }, + { + "epoch": 0.7397320448056227, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.2624895572662354, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7142000198364258, + "num_tokens": 174400581.0, + "step": 6736 + }, + { + "epoch": 0.7398418625082364, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 1.9383976459503174, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7001166343688965, + "num_tokens": 174431436.0, + "step": 6737 + }, + { + "epoch": 0.73995168021085, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3051979541778564, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.7062886953353882, + "num_tokens": 174456476.0, + "step": 6738 + }, + { + "epoch": 0.7400614979134637, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.3616044521331787, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6890087127685547, + "num_tokens": 174479634.0, + "step": 6739 + }, + { + "epoch": 0.7401713156160773, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.2493200302124023, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7053998708724976, + "num_tokens": 174504223.0, + "step": 6740 + }, + { + "epoch": 0.7402811333186909, + "ewc_loss": 6.079673767089844e-05, + "grad_norm": 2.0368008613586426, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7259597182273865, + "num_tokens": 174532251.0, + "step": 6741 + }, + { + "epoch": 0.7403909510213046, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.578950881958008, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7219057083129883, + "num_tokens": 174551363.0, + "step": 6742 + }, + { + "epoch": 0.7405007687239183, + "ewc_loss": 6.103515625e-05, + "grad_norm": 1.8872658014297485, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7129092812538147, + "num_tokens": 174585552.0, + "step": 6743 + }, + { + "epoch": 0.740610586426532, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.245772361755371, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.694311261177063, + "num_tokens": 174612006.0, + "step": 6744 + }, + { + "epoch": 0.7407204041291456, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3294975757598877, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6993681788444519, + "num_tokens": 174635438.0, + "step": 6745 + }, + { + "epoch": 0.7408302218317593, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.0165014266967773, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7000953555107117, + "num_tokens": 174663764.0, + "step": 6746 + }, + { + "epoch": 0.7409400395343729, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.323258638381958, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7106931805610657, + "num_tokens": 174686855.0, + "step": 6747 + }, + { + "epoch": 0.7410498572369866, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.2671492099761963, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6943516135215759, + "num_tokens": 174712408.0, + "step": 6748 + }, + { + "epoch": 0.7411596749396002, + "ewc_loss": 6.103515625e-05, + "grad_norm": 1.9640922546386719, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.6999092102050781, + "num_tokens": 174746301.0, + "step": 6749 + }, + { + "epoch": 0.741269492642214, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.161168098449707, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7073651552200317, + "num_tokens": 174772707.0, + "step": 6750 + }, + { + "epoch": 0.7413793103448276, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.018681526184082, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7022093534469604, + "num_tokens": 174800887.0, + "step": 6751 + }, + { + "epoch": 0.7414891280474413, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1784610748291016, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7221944332122803, + "num_tokens": 174825484.0, + "step": 6752 + }, + { + "epoch": 0.7415989457500549, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.029409170150757, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6984418630599976, + "num_tokens": 174856788.0, + "step": 6753 + }, + { + "epoch": 0.7417087634526686, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.088057518005371, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7188469171524048, + "num_tokens": 174883870.0, + "step": 6754 + }, + { + "epoch": 0.7418185811552822, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.151838541030884, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7141547799110413, + "num_tokens": 174911277.0, + "step": 6755 + }, + { + "epoch": 0.7419283988578959, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.154325485229492, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7190130949020386, + "num_tokens": 174937395.0, + "step": 6756 + }, + { + "epoch": 0.7420382165605095, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.2218780517578125, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6874431371688843, + "num_tokens": 174962830.0, + "step": 6757 + }, + { + "epoch": 0.7421480342631233, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1182003021240234, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7019001245498657, + "num_tokens": 174989345.0, + "step": 6758 + }, + { + "epoch": 0.7422578519657369, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.28667950630188, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7195612788200378, + "num_tokens": 175012986.0, + "step": 6759 + }, + { + "epoch": 0.7423676696683505, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.6058719158172607, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7115198373794556, + "num_tokens": 175032096.0, + "step": 6760 + }, + { + "epoch": 0.7424774873709642, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.5267367362976074, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7326873540878296, + "num_tokens": 175051955.0, + "step": 6761 + }, + { + "epoch": 0.7425873050735778, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.210446357727051, + "learning_rate": 1e-06, + "loss": 1.0742, + "mean_token_accuracy": 0.6796071529388428, + "num_tokens": 175078938.0, + "step": 6762 + }, + { + "epoch": 0.7426971227761915, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1028504371643066, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7222093939781189, + "num_tokens": 175104952.0, + "step": 6763 + }, + { + "epoch": 0.7428069404788051, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.243250846862793, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.6997871994972229, + "num_tokens": 175129499.0, + "step": 6764 + }, + { + "epoch": 0.7429167581814189, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3049676418304443, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6787676811218262, + "num_tokens": 175155643.0, + "step": 6765 + }, + { + "epoch": 0.7430265758840325, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3200936317443848, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6997045278549194, + "num_tokens": 175179215.0, + "step": 6766 + }, + { + "epoch": 0.7431363935866462, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.0540733337402344, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6975723505020142, + "num_tokens": 175208654.0, + "step": 6767 + }, + { + "epoch": 0.7432462112892598, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.158888816833496, + "learning_rate": 1e-06, + "loss": 1.0121, + "mean_token_accuracy": 0.7035910487174988, + "num_tokens": 175235570.0, + "step": 6768 + }, + { + "epoch": 0.7433560289918735, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1885764598846436, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.6956309080123901, + "num_tokens": 175260521.0, + "step": 6769 + }, + { + "epoch": 0.7434658466944871, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1233832836151123, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7206674814224243, + "num_tokens": 175287083.0, + "step": 6770 + }, + { + "epoch": 0.7435756643971008, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3912084102630615, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7174131870269775, + "num_tokens": 175309073.0, + "step": 6771 + }, + { + "epoch": 0.7436854820997145, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.5876259803771973, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.708940863609314, + "num_tokens": 175327506.0, + "step": 6772 + }, + { + "epoch": 0.7437952998023282, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.4769115447998047, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7112501859664917, + "num_tokens": 175350031.0, + "step": 6773 + }, + { + "epoch": 0.7439051175049418, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.317436695098877, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7117946147918701, + "num_tokens": 175371750.0, + "step": 6774 + }, + { + "epoch": 0.7440149352075555, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.064999580383301, + "learning_rate": 1e-06, + "loss": 1.0531, + "mean_token_accuracy": 0.680625319480896, + "num_tokens": 175398988.0, + "step": 6775 + }, + { + "epoch": 0.7441247529101691, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.0279347896575928, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.710292637348175, + "num_tokens": 175427102.0, + "step": 6776 + }, + { + "epoch": 0.7442345706127828, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.136441707611084, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7065953612327576, + "num_tokens": 175454098.0, + "step": 6777 + }, + { + "epoch": 0.7443443883153964, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.0797390937805176, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6913424730300903, + "num_tokens": 175481485.0, + "step": 6778 + }, + { + "epoch": 0.7444542060180102, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.266300678253174, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7069231867790222, + "num_tokens": 175505092.0, + "step": 6779 + }, + { + "epoch": 0.7445640237206238, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.278688669204712, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6984404921531677, + "num_tokens": 175529187.0, + "step": 6780 + }, + { + "epoch": 0.7446738414232374, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3511221408843994, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6903328895568848, + "num_tokens": 175554714.0, + "step": 6781 + }, + { + "epoch": 0.7447836591258511, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.1204562187194824, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6986677646636963, + "num_tokens": 175582796.0, + "step": 6782 + }, + { + "epoch": 0.7448934768284647, + "ewc_loss": 6.103515625e-05, + "grad_norm": 2.3146252632141113, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7030467987060547, + "num_tokens": 175606081.0, + "step": 6783 + }, + { + "epoch": 0.7450032945310784, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.1326255798339844, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7225993871688843, + "num_tokens": 175632000.0, + "step": 6784 + }, + { + "epoch": 0.745113112233692, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.1006386280059814, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7178686857223511, + "num_tokens": 175658364.0, + "step": 6785 + }, + { + "epoch": 0.7452229299363057, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.3139545917510986, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7123832106590271, + "num_tokens": 175683050.0, + "step": 6786 + }, + { + "epoch": 0.7453327476389194, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.265207529067993, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7188236117362976, + "num_tokens": 175706060.0, + "step": 6787 + }, + { + "epoch": 0.7454425653415331, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.0444421768188477, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7162133455276489, + "num_tokens": 175734301.0, + "step": 6788 + }, + { + "epoch": 0.7455523830441467, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.2349300384521484, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7236473560333252, + "num_tokens": 175757178.0, + "step": 6789 + }, + { + "epoch": 0.7456622007467604, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.4786536693573, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6902912855148315, + "num_tokens": 175778863.0, + "step": 6790 + }, + { + "epoch": 0.745772018449374, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.2048449516296387, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7117359638214111, + "num_tokens": 175806023.0, + "step": 6791 + }, + { + "epoch": 0.7458818361519877, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.392271041870117, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7066760659217834, + "num_tokens": 175828049.0, + "step": 6792 + }, + { + "epoch": 0.7459916538546013, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.5596814155578613, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7099388837814331, + "num_tokens": 175848135.0, + "step": 6793 + }, + { + "epoch": 0.7461014715572151, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.079038619995117, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7035447955131531, + "num_tokens": 175876498.0, + "step": 6794 + }, + { + "epoch": 0.7462112892598287, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.1941304206848145, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6999945640563965, + "num_tokens": 175902733.0, + "step": 6795 + }, + { + "epoch": 0.7463211069624424, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.380021333694458, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7093110680580139, + "num_tokens": 175925566.0, + "step": 6796 + }, + { + "epoch": 0.746430924665056, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.0492868423461914, + "learning_rate": 1e-06, + "loss": 1.0307, + "mean_token_accuracy": 0.6887807846069336, + "num_tokens": 175955080.0, + "step": 6797 + }, + { + "epoch": 0.7465407423676697, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.345853328704834, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7215708494186401, + "num_tokens": 175977085.0, + "step": 6798 + }, + { + "epoch": 0.7466505600702833, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.240433692932129, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7112667560577393, + "num_tokens": 176003377.0, + "step": 6799 + }, + { + "epoch": 0.7467603777728969, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 1.9844841957092285, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7083520293235779, + "num_tokens": 176030427.0, + "step": 6800 + }, + { + "epoch": 0.7468701954755107, + "ewc_loss": 6.151199340820312e-05, + "grad_norm": 2.0342419147491455, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.687701940536499, + "num_tokens": 176058425.0, + "step": 6801 + }, + { + "epoch": 0.7469800131781243, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.1982710361480713, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7341171503067017, + "num_tokens": 176081969.0, + "step": 6802 + }, + { + "epoch": 0.747089830880738, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.234898090362549, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7194227576255798, + "num_tokens": 176106404.0, + "step": 6803 + }, + { + "epoch": 0.7471996485833516, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.0739495754241943, + "learning_rate": 1e-06, + "loss": 1.0678, + "mean_token_accuracy": 0.6731127500534058, + "num_tokens": 176137354.0, + "step": 6804 + }, + { + "epoch": 0.7473094662859653, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.198758840560913, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.7039872407913208, + "num_tokens": 176162288.0, + "step": 6805 + }, + { + "epoch": 0.7474192839885789, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.1646060943603516, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.7025868892669678, + "num_tokens": 176189523.0, + "step": 6806 + }, + { + "epoch": 0.7475291016911926, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.3258655071258545, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7069743275642395, + "num_tokens": 176213616.0, + "step": 6807 + }, + { + "epoch": 0.7476389193938063, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.1812527179718018, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7230740785598755, + "num_tokens": 176236767.0, + "step": 6808 + }, + { + "epoch": 0.74774873709642, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.089510679244995, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7251968383789062, + "num_tokens": 176263344.0, + "step": 6809 + }, + { + "epoch": 0.7478585547990336, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.2051444053649902, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7057476043701172, + "num_tokens": 176288047.0, + "step": 6810 + }, + { + "epoch": 0.7479683725016473, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.0657317638397217, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.7012457847595215, + "num_tokens": 176317781.0, + "step": 6811 + }, + { + "epoch": 0.7480781902042609, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.2198781967163086, + "learning_rate": 1e-06, + "loss": 1.0566, + "mean_token_accuracy": 0.6821866035461426, + "num_tokens": 176347624.0, + "step": 6812 + }, + { + "epoch": 0.7481880079068746, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.5453343391418457, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7177373766899109, + "num_tokens": 176367772.0, + "step": 6813 + }, + { + "epoch": 0.7482978256094882, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 1.9870457649230957, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7119324803352356, + "num_tokens": 176400767.0, + "step": 6814 + }, + { + "epoch": 0.7484076433121019, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.219170331954956, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6872826218605042, + "num_tokens": 176425357.0, + "step": 6815 + }, + { + "epoch": 0.7485174610147156, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.489912748336792, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.694599986076355, + "num_tokens": 176445214.0, + "step": 6816 + }, + { + "epoch": 0.7486272787173293, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.16510009765625, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7061705589294434, + "num_tokens": 176472967.0, + "step": 6817 + }, + { + "epoch": 0.7487370964199429, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9681037664413452, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.720396101474762, + "num_tokens": 176502370.0, + "step": 6818 + }, + { + "epoch": 0.7488469141225566, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2874062061309814, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7062687873840332, + "num_tokens": 176525646.0, + "step": 6819 + }, + { + "epoch": 0.7489567318251702, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1058788299560547, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6963198781013489, + "num_tokens": 176552922.0, + "step": 6820 + }, + { + "epoch": 0.7490665495277838, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.0896785259246826, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7097101807594299, + "num_tokens": 176580090.0, + "step": 6821 + }, + { + "epoch": 0.7491763672303975, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.037320852279663, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7080322504043579, + "num_tokens": 176607392.0, + "step": 6822 + }, + { + "epoch": 0.7492861849330112, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.3919029235839844, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7205168604850769, + "num_tokens": 176631020.0, + "step": 6823 + }, + { + "epoch": 0.7493960026356249, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1342389583587646, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7261499166488647, + "num_tokens": 176659765.0, + "step": 6824 + }, + { + "epoch": 0.7495058203382385, + "ewc_loss": 6.198883056640625e-05, + "grad_norm": 2.044804334640503, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7269852757453918, + "num_tokens": 176688642.0, + "step": 6825 + }, + { + "epoch": 0.7496156380408522, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.306370258331299, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7328840494155884, + "num_tokens": 176712306.0, + "step": 6826 + }, + { + "epoch": 0.7497254557434658, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.391751289367676, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7307940125465393, + "num_tokens": 176731830.0, + "step": 6827 + }, + { + "epoch": 0.7498352734460795, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.027811050415039, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7012416124343872, + "num_tokens": 176762912.0, + "step": 6828 + }, + { + "epoch": 0.7499450911486931, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.192854404449463, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7019466161727905, + "num_tokens": 176787808.0, + "step": 6829 + }, + { + "epoch": 0.7500549088513069, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.39231538772583, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7136917114257812, + "num_tokens": 176809620.0, + "step": 6830 + }, + { + "epoch": 0.7501647265539205, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9895602464675903, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7016965746879578, + "num_tokens": 176838777.0, + "step": 6831 + }, + { + "epoch": 0.7502745442565342, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.443103313446045, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.6852881908416748, + "num_tokens": 176863801.0, + "step": 6832 + }, + { + "epoch": 0.7503843619591478, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1038663387298584, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7100706100463867, + "num_tokens": 176892098.0, + "step": 6833 + }, + { + "epoch": 0.7504941796617615, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0456769466400146, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7199454307556152, + "num_tokens": 176919806.0, + "step": 6834 + }, + { + "epoch": 0.7506039973643751, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.214850425720215, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6860018372535706, + "num_tokens": 176944799.0, + "step": 6835 + }, + { + "epoch": 0.7507138150669888, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.379852294921875, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.6964459419250488, + "num_tokens": 176968499.0, + "step": 6836 + }, + { + "epoch": 0.7508236327696025, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2243473529815674, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7124010920524597, + "num_tokens": 176992760.0, + "step": 6837 + }, + { + "epoch": 0.7509334504722162, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.107692241668701, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6961790323257446, + "num_tokens": 177019679.0, + "step": 6838 + }, + { + "epoch": 0.7510432681748298, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.213735342025757, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.6949532628059387, + "num_tokens": 177048441.0, + "step": 6839 + }, + { + "epoch": 0.7511530858774434, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0858070850372314, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7061849236488342, + "num_tokens": 177076460.0, + "step": 6840 + }, + { + "epoch": 0.7512629035800571, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3521695137023926, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.703477144241333, + "num_tokens": 177099730.0, + "step": 6841 + }, + { + "epoch": 0.7513727212826707, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3510303497314453, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7046843767166138, + "num_tokens": 177122154.0, + "step": 6842 + }, + { + "epoch": 0.7514825389852844, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.4663100242614746, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.715053915977478, + "num_tokens": 177142945.0, + "step": 6843 + }, + { + "epoch": 0.7515923566878981, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1146485805511475, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.7030574679374695, + "num_tokens": 177170514.0, + "step": 6844 + }, + { + "epoch": 0.7517021743905118, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.211503267288208, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.6995277404785156, + "num_tokens": 177196408.0, + "step": 6845 + }, + { + "epoch": 0.7518119920931254, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3648743629455566, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7236680388450623, + "num_tokens": 177220300.0, + "step": 6846 + }, + { + "epoch": 0.7519218097957391, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.006182909011841, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.6969105005264282, + "num_tokens": 177251674.0, + "step": 6847 + }, + { + "epoch": 0.7520316274983527, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.947737693786621, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6971486806869507, + "num_tokens": 177284386.0, + "step": 6848 + }, + { + "epoch": 0.7521414452009664, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9518903493881226, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6989971399307251, + "num_tokens": 177314575.0, + "step": 6849 + }, + { + "epoch": 0.75225126290358, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1663782596588135, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6935328841209412, + "num_tokens": 177340530.0, + "step": 6850 + }, + { + "epoch": 0.7523610806061937, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3043036460876465, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7031168341636658, + "num_tokens": 177363114.0, + "step": 6851 + }, + { + "epoch": 0.7524708983088074, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.027951240539551, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7046148777008057, + "num_tokens": 177392813.0, + "step": 6852 + }, + { + "epoch": 0.7525807160114211, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3867173194885254, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.7088831663131714, + "num_tokens": 177416349.0, + "step": 6853 + }, + { + "epoch": 0.7526905337140347, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2679219245910645, + "learning_rate": 1e-06, + "loss": 1.0273, + "mean_token_accuracy": 0.6877206563949585, + "num_tokens": 177441313.0, + "step": 6854 + }, + { + "epoch": 0.7528003514166484, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3979291915893555, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7453442215919495, + "num_tokens": 177462066.0, + "step": 6855 + }, + { + "epoch": 0.752910169119262, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2402615547180176, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7069467306137085, + "num_tokens": 177488623.0, + "step": 6856 + }, + { + "epoch": 0.7530199868218757, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0447211265563965, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7099429965019226, + "num_tokens": 177517433.0, + "step": 6857 + }, + { + "epoch": 0.7531298045244893, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.114902973175049, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7025240659713745, + "num_tokens": 177546971.0, + "step": 6858 + }, + { + "epoch": 0.7532396222271031, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3626246452331543, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6901251673698425, + "num_tokens": 177568348.0, + "step": 6859 + }, + { + "epoch": 0.7533494399297167, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.211780309677124, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.711556613445282, + "num_tokens": 177595015.0, + "step": 6860 + }, + { + "epoch": 0.7534592576323303, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1493473052978516, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7111442685127258, + "num_tokens": 177621428.0, + "step": 6861 + }, + { + "epoch": 0.753569075334944, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.144664764404297, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7117832899093628, + "num_tokens": 177648253.0, + "step": 6862 + }, + { + "epoch": 0.7536788930375576, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.651362657546997, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7350777387619019, + "num_tokens": 177666539.0, + "step": 6863 + }, + { + "epoch": 0.7537887107401713, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3276379108428955, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7182668447494507, + "num_tokens": 177691319.0, + "step": 6864 + }, + { + "epoch": 0.7538985284427849, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1057705879211426, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7012231349945068, + "num_tokens": 177721356.0, + "step": 6865 + }, + { + "epoch": 0.7540083461453987, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9804452657699585, + "learning_rate": 1e-06, + "loss": 1.0509, + "mean_token_accuracy": 0.6839597821235657, + "num_tokens": 177753114.0, + "step": 6866 + }, + { + "epoch": 0.7541181638480123, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.153547525405884, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7259339094161987, + "num_tokens": 177777685.0, + "step": 6867 + }, + { + "epoch": 0.754227981550626, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0616061687469482, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7080938816070557, + "num_tokens": 177805844.0, + "step": 6868 + }, + { + "epoch": 0.7543377992532396, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.116950273513794, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6958894729614258, + "num_tokens": 177835153.0, + "step": 6869 + }, + { + "epoch": 0.7544476169558533, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1154415607452393, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7120414972305298, + "num_tokens": 177861316.0, + "step": 6870 + }, + { + "epoch": 0.7545574346584669, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.193105459213257, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.71174156665802, + "num_tokens": 177888119.0, + "step": 6871 + }, + { + "epoch": 0.7546672523610806, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.020338296890259, + "learning_rate": 1e-06, + "loss": 1.0437, + "mean_token_accuracy": 0.6833833456039429, + "num_tokens": 177920001.0, + "step": 6872 + }, + { + "epoch": 0.7547770700636943, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.660552978515625, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7107164859771729, + "num_tokens": 177938096.0, + "step": 6873 + }, + { + "epoch": 0.754886887766308, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2262980937957764, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7016483545303345, + "num_tokens": 177961331.0, + "step": 6874 + }, + { + "epoch": 0.7549967054689216, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1721394062042236, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7284988760948181, + "num_tokens": 177985902.0, + "step": 6875 + }, + { + "epoch": 0.7551065231715353, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.318690538406372, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6882779598236084, + "num_tokens": 178011077.0, + "step": 6876 + }, + { + "epoch": 0.7552163408741489, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.4098269939422607, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7197535037994385, + "num_tokens": 178033075.0, + "step": 6877 + }, + { + "epoch": 0.7553261585767626, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.04245662689209, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7057847380638123, + "num_tokens": 178060752.0, + "step": 6878 + }, + { + "epoch": 0.7554359762793762, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9987338781356812, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7262589931488037, + "num_tokens": 178089622.0, + "step": 6879 + }, + { + "epoch": 0.7555457939819898, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9948914051055908, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6799459457397461, + "num_tokens": 178120948.0, + "step": 6880 + }, + { + "epoch": 0.7556556116846036, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.6638355255126953, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6929728984832764, + "num_tokens": 178143058.0, + "step": 6881 + }, + { + "epoch": 0.7557654293872172, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.247030735015869, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7036010026931763, + "num_tokens": 178166367.0, + "step": 6882 + }, + { + "epoch": 0.7558752470898309, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.5489463806152344, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7149558663368225, + "num_tokens": 178185632.0, + "step": 6883 + }, + { + "epoch": 0.7559850647924445, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.318124771118164, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6951348781585693, + "num_tokens": 178210509.0, + "step": 6884 + }, + { + "epoch": 0.7560948824950582, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.8894003629684448, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6941940784454346, + "num_tokens": 178246098.0, + "step": 6885 + }, + { + "epoch": 0.7562047001976718, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2656078338623047, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7043660879135132, + "num_tokens": 178269653.0, + "step": 6886 + }, + { + "epoch": 0.7563145179002855, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1691298484802246, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7064911127090454, + "num_tokens": 178297507.0, + "step": 6887 + }, + { + "epoch": 0.7564243356028992, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2255334854125977, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7120822668075562, + "num_tokens": 178322444.0, + "step": 6888 + }, + { + "epoch": 0.7565341533055129, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.4089670181274414, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6911777257919312, + "num_tokens": 178345073.0, + "step": 6889 + }, + { + "epoch": 0.7566439710081265, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.033069372177124, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6954058408737183, + "num_tokens": 178373921.0, + "step": 6890 + }, + { + "epoch": 0.7567537887107402, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.062105417251587, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7075663805007935, + "num_tokens": 178401628.0, + "step": 6891 + }, + { + "epoch": 0.7568636064133538, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2568540573120117, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7318127155303955, + "num_tokens": 178424561.0, + "step": 6892 + }, + { + "epoch": 0.7569734241159675, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.136631965637207, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7312049865722656, + "num_tokens": 178447980.0, + "step": 6893 + }, + { + "epoch": 0.7570832418185811, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1564536094665527, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7189716100692749, + "num_tokens": 178471718.0, + "step": 6894 + }, + { + "epoch": 0.7571930595211949, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.212419271469116, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7043426632881165, + "num_tokens": 178496847.0, + "step": 6895 + }, + { + "epoch": 0.7573028772238085, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.009871244430542, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6888368129730225, + "num_tokens": 178527000.0, + "step": 6896 + }, + { + "epoch": 0.7574126949264222, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.349179744720459, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.7062622308731079, + "num_tokens": 178550133.0, + "step": 6897 + }, + { + "epoch": 0.7575225126290358, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.271287679672241, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7014085054397583, + "num_tokens": 178577050.0, + "step": 6898 + }, + { + "epoch": 0.7576323303316495, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.132167339324951, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7246739864349365, + "num_tokens": 178602725.0, + "step": 6899 + }, + { + "epoch": 0.7577421480342631, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9507337808609009, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.713937520980835, + "num_tokens": 178633785.0, + "step": 6900 + }, + { + "epoch": 0.7578519657368767, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.5124351978302, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7070226669311523, + "num_tokens": 178654144.0, + "step": 6901 + }, + { + "epoch": 0.7579617834394905, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.155818462371826, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7015013694763184, + "num_tokens": 178680025.0, + "step": 6902 + }, + { + "epoch": 0.7580716011421041, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3964903354644775, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.704724907875061, + "num_tokens": 178702119.0, + "step": 6903 + }, + { + "epoch": 0.7581814188447178, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0015246868133545, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.698447048664093, + "num_tokens": 178731672.0, + "step": 6904 + }, + { + "epoch": 0.7582912365473314, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.090585231781006, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7263749837875366, + "num_tokens": 178758582.0, + "step": 6905 + }, + { + "epoch": 0.7584010542499451, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.967629313468933, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6902746558189392, + "num_tokens": 178787672.0, + "step": 6906 + }, + { + "epoch": 0.7585108719525587, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.135591983795166, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6882498860359192, + "num_tokens": 178814095.0, + "step": 6907 + }, + { + "epoch": 0.7586206896551724, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0344388484954834, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6929861307144165, + "num_tokens": 178843461.0, + "step": 6908 + }, + { + "epoch": 0.758730507357786, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.092639446258545, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.7019858360290527, + "num_tokens": 178870250.0, + "step": 6909 + }, + { + "epoch": 0.7588403250603998, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.225382089614868, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.71982342004776, + "num_tokens": 178893417.0, + "step": 6910 + }, + { + "epoch": 0.7589501427630134, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.327392578125, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7033601999282837, + "num_tokens": 178915244.0, + "step": 6911 + }, + { + "epoch": 0.7590599604656271, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3075287342071533, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7019819617271423, + "num_tokens": 178938120.0, + "step": 6912 + }, + { + "epoch": 0.7591697781682407, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.127777576446533, + "learning_rate": 1e-06, + "loss": 1.0589, + "mean_token_accuracy": 0.6839752197265625, + "num_tokens": 178965558.0, + "step": 6913 + }, + { + "epoch": 0.7592795958708544, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.060670852661133, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7179521918296814, + "num_tokens": 178994111.0, + "step": 6914 + }, + { + "epoch": 0.759389413573468, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3403096199035645, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7053757309913635, + "num_tokens": 179017532.0, + "step": 6915 + }, + { + "epoch": 0.7594992312760817, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.6387720108032227, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7039317488670349, + "num_tokens": 179036660.0, + "step": 6916 + }, + { + "epoch": 0.7596090489786954, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0316994190216064, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7057000994682312, + "num_tokens": 179064753.0, + "step": 6917 + }, + { + "epoch": 0.7597188666813091, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2723772525787354, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7189682722091675, + "num_tokens": 179087979.0, + "step": 6918 + }, + { + "epoch": 0.7598286843839227, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.288588762283325, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.7010871767997742, + "num_tokens": 179113525.0, + "step": 6919 + }, + { + "epoch": 0.7599385020865363, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1048929691314697, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7165380716323853, + "num_tokens": 179140003.0, + "step": 6920 + }, + { + "epoch": 0.76004831978915, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.210390090942383, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7221531867980957, + "num_tokens": 179164191.0, + "step": 6921 + }, + { + "epoch": 0.7601581374917636, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.026779890060425, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6839861869812012, + "num_tokens": 179194456.0, + "step": 6922 + }, + { + "epoch": 0.7602679551943773, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1322782039642334, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7310065031051636, + "num_tokens": 179220806.0, + "step": 6923 + }, + { + "epoch": 0.760377772896991, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.402869939804077, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7036317586898804, + "num_tokens": 179243326.0, + "step": 6924 + }, + { + "epoch": 0.7604875905996047, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1264755725860596, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7089439630508423, + "num_tokens": 179270466.0, + "step": 6925 + }, + { + "epoch": 0.7605974083022183, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9940301179885864, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7196367979049683, + "num_tokens": 179300590.0, + "step": 6926 + }, + { + "epoch": 0.760707226004832, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2474918365478516, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6896179914474487, + "num_tokens": 179324837.0, + "step": 6927 + }, + { + "epoch": 0.7608170437074456, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.98058021068573, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6983519792556763, + "num_tokens": 179356085.0, + "step": 6928 + }, + { + "epoch": 0.7609268614100593, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.259246587753296, + "learning_rate": 1e-06, + "loss": 1.0665, + "mean_token_accuracy": 0.6809115409851074, + "num_tokens": 179380943.0, + "step": 6929 + }, + { + "epoch": 0.7610366791126729, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1688058376312256, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.6995924711227417, + "num_tokens": 179406636.0, + "step": 6930 + }, + { + "epoch": 0.7611464968152867, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.175419330596924, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6978527307510376, + "num_tokens": 179432128.0, + "step": 6931 + }, + { + "epoch": 0.7612563145179003, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.2413787841796875, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.7034274339675903, + "num_tokens": 179457591.0, + "step": 6932 + }, + { + "epoch": 0.761366132220514, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.258518934249878, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7125465869903564, + "num_tokens": 179483646.0, + "step": 6933 + }, + { + "epoch": 0.7614759499231276, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0741465091705322, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7130885124206543, + "num_tokens": 179511204.0, + "step": 6934 + }, + { + "epoch": 0.7615857676257413, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1539528369903564, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.706511378288269, + "num_tokens": 179536067.0, + "step": 6935 + }, + { + "epoch": 0.7616955853283549, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.0159411430358887, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7380337715148926, + "num_tokens": 179563707.0, + "step": 6936 + }, + { + "epoch": 0.7618054030309686, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.201913595199585, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7233157157897949, + "num_tokens": 179587191.0, + "step": 6937 + }, + { + "epoch": 0.7619152207335822, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.183558464050293, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.733555257320404, + "num_tokens": 179610040.0, + "step": 6938 + }, + { + "epoch": 0.762025038436196, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.262216091156006, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7025465965270996, + "num_tokens": 179635594.0, + "step": 6939 + }, + { + "epoch": 0.7621348561388096, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.329479694366455, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7139352560043335, + "num_tokens": 179657892.0, + "step": 6940 + }, + { + "epoch": 0.7622446738414232, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.099459171295166, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7292729616165161, + "num_tokens": 179683394.0, + "step": 6941 + }, + { + "epoch": 0.7623544915440369, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0297346115112305, + "learning_rate": 1e-06, + "loss": 1.0507, + "mean_token_accuracy": 0.6856258511543274, + "num_tokens": 179711403.0, + "step": 6942 + }, + { + "epoch": 0.7624643092466505, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.5768847465515137, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7203883528709412, + "num_tokens": 179730406.0, + "step": 6943 + }, + { + "epoch": 0.7625741269492642, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0649609565734863, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6931003928184509, + "num_tokens": 179761665.0, + "step": 6944 + }, + { + "epoch": 0.7626839446518778, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2485461235046387, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7088471055030823, + "num_tokens": 179784254.0, + "step": 6945 + }, + { + "epoch": 0.7627937623544916, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.487867593765259, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7039594054222107, + "num_tokens": 179807077.0, + "step": 6946 + }, + { + "epoch": 0.7629035800571052, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8776978254318237, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7174001932144165, + "num_tokens": 179838235.0, + "step": 6947 + }, + { + "epoch": 0.7630133977597189, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.09596586227417, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.705900251865387, + "num_tokens": 179866103.0, + "step": 6948 + }, + { + "epoch": 0.7631232154623325, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3784642219543457, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.721645176410675, + "num_tokens": 179888382.0, + "step": 6949 + }, + { + "epoch": 0.7632330331649462, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.407958507537842, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7104479074478149, + "num_tokens": 179908263.0, + "step": 6950 + }, + { + "epoch": 0.7633428508675598, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.155256509780884, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7165112495422363, + "num_tokens": 179935433.0, + "step": 6951 + }, + { + "epoch": 0.7634526685701735, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.146362781524658, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7010180950164795, + "num_tokens": 179961446.0, + "step": 6952 + }, + { + "epoch": 0.7635624862727872, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1090502738952637, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7190856337547302, + "num_tokens": 179987906.0, + "step": 6953 + }, + { + "epoch": 0.7636723039754009, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2819783687591553, + "learning_rate": 1e-06, + "loss": 1.0607, + "mean_token_accuracy": 0.6754634380340576, + "num_tokens": 180011551.0, + "step": 6954 + }, + { + "epoch": 0.7637821216780145, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.159252882003784, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7151311635971069, + "num_tokens": 180037809.0, + "step": 6955 + }, + { + "epoch": 0.7638919393806282, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1780664920806885, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7129522562026978, + "num_tokens": 180063827.0, + "step": 6956 + }, + { + "epoch": 0.7640017570832418, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1663570404052734, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6941230297088623, + "num_tokens": 180089028.0, + "step": 6957 + }, + { + "epoch": 0.7641115747858555, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.320758819580078, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7124855518341064, + "num_tokens": 180113074.0, + "step": 6958 + }, + { + "epoch": 0.7642213924884691, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1732373237609863, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6911333799362183, + "num_tokens": 180140288.0, + "step": 6959 + }, + { + "epoch": 0.7643312101910829, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1107027530670166, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6919673681259155, + "num_tokens": 180165848.0, + "step": 6960 + }, + { + "epoch": 0.7644410278936965, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.078439235687256, + "learning_rate": 1e-06, + "loss": 1.0655, + "mean_token_accuracy": 0.6834700107574463, + "num_tokens": 180193944.0, + "step": 6961 + }, + { + "epoch": 0.7645508455963101, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.3347816467285156, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7029706239700317, + "num_tokens": 180217722.0, + "step": 6962 + }, + { + "epoch": 0.7646606632989238, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 2.1451497077941895, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7247693538665771, + "num_tokens": 180242650.0, + "step": 6963 + }, + { + "epoch": 0.7647704810015374, + "ewc_loss": 6.246566772460938e-05, + "grad_norm": 1.9072763919830322, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7353296279907227, + "num_tokens": 180272453.0, + "step": 6964 + }, + { + "epoch": 0.7648802987041511, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.030439615249634, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.717745304107666, + "num_tokens": 180300422.0, + "step": 6965 + }, + { + "epoch": 0.7649901164067647, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4046859741210938, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7303795218467712, + "num_tokens": 180322956.0, + "step": 6966 + }, + { + "epoch": 0.7650999341093784, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2057461738586426, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7027299404144287, + "num_tokens": 180348589.0, + "step": 6967 + }, + { + "epoch": 0.7652097518119921, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5112216472625732, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7062101364135742, + "num_tokens": 180369574.0, + "step": 6968 + }, + { + "epoch": 0.7653195695146058, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3178954124450684, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7040514945983887, + "num_tokens": 180391533.0, + "step": 6969 + }, + { + "epoch": 0.7654293872172194, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0866150856018066, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6882862448692322, + "num_tokens": 180419677.0, + "step": 6970 + }, + { + "epoch": 0.7655392049198331, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0794713497161865, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6882239580154419, + "num_tokens": 180449792.0, + "step": 6971 + }, + { + "epoch": 0.7656490226224467, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.511212110519409, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7082819938659668, + "num_tokens": 180470960.0, + "step": 6972 + }, + { + "epoch": 0.7657588403250604, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.020798683166504, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6815093755722046, + "num_tokens": 180501704.0, + "step": 6973 + }, + { + "epoch": 0.765868658027674, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4683940410614014, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7047696709632874, + "num_tokens": 180522239.0, + "step": 6974 + }, + { + "epoch": 0.7659784757302878, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3225648403167725, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.7068653106689453, + "num_tokens": 180545659.0, + "step": 6975 + }, + { + "epoch": 0.7660882934329014, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1015443801879883, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7257543802261353, + "num_tokens": 180570244.0, + "step": 6976 + }, + { + "epoch": 0.7661981111355151, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0247581005096436, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7221740484237671, + "num_tokens": 180597361.0, + "step": 6977 + }, + { + "epoch": 0.7663079288381287, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.221285820007324, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7013208270072937, + "num_tokens": 180622400.0, + "step": 6978 + }, + { + "epoch": 0.7664177465407424, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.073394298553467, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7103385925292969, + "num_tokens": 180649632.0, + "step": 6979 + }, + { + "epoch": 0.766527564243356, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0326778888702393, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7162445187568665, + "num_tokens": 180678695.0, + "step": 6980 + }, + { + "epoch": 0.7666373819459696, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4520583152770996, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7225919961929321, + "num_tokens": 180700718.0, + "step": 6981 + }, + { + "epoch": 0.7667471996485834, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2221009731292725, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7304415702819824, + "num_tokens": 180723553.0, + "step": 6982 + }, + { + "epoch": 0.766857017351197, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.254617929458618, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7137492895126343, + "num_tokens": 180748214.0, + "step": 6983 + }, + { + "epoch": 0.7669668350538107, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1875557899475098, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7055686116218567, + "num_tokens": 180773817.0, + "step": 6984 + }, + { + "epoch": 0.7670766527564243, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3787574768066406, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7103238105773926, + "num_tokens": 180794784.0, + "step": 6985 + }, + { + "epoch": 0.767186470459038, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4131455421447754, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7034441232681274, + "num_tokens": 180815988.0, + "step": 6986 + }, + { + "epoch": 0.7672962881616516, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0625908374786377, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6993966102600098, + "num_tokens": 180842777.0, + "step": 6987 + }, + { + "epoch": 0.7674061058642653, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1998636722564697, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7149451971054077, + "num_tokens": 180868011.0, + "step": 6988 + }, + { + "epoch": 0.767515923566879, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0072388648986816, + "learning_rate": 1e-06, + "loss": 1.0292, + "mean_token_accuracy": 0.6916658878326416, + "num_tokens": 180899163.0, + "step": 6989 + }, + { + "epoch": 0.7676257412694927, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1818203926086426, + "learning_rate": 1e-06, + "loss": 1.0879, + "mean_token_accuracy": 0.6733177304267883, + "num_tokens": 180926392.0, + "step": 6990 + }, + { + "epoch": 0.7677355589721063, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1748268604278564, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6967278718948364, + "num_tokens": 180951585.0, + "step": 6991 + }, + { + "epoch": 0.76784537667472, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.260204315185547, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7064398527145386, + "num_tokens": 180975363.0, + "step": 6992 + }, + { + "epoch": 0.7679551943773336, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1588258743286133, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6902700066566467, + "num_tokens": 181003116.0, + "step": 6993 + }, + { + "epoch": 0.7680650120799473, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1831769943237305, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.691999077796936, + "num_tokens": 181029285.0, + "step": 6994 + }, + { + "epoch": 0.7681748297825609, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.201080799102783, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7118290662765503, + "num_tokens": 181053653.0, + "step": 6995 + }, + { + "epoch": 0.7682846474851747, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1623146533966064, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7066133618354797, + "num_tokens": 181079899.0, + "step": 6996 + }, + { + "epoch": 0.7683944651877883, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2536182403564453, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6924817562103271, + "num_tokens": 181105803.0, + "step": 6997 + }, + { + "epoch": 0.768504282890402, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1434085369110107, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7074829339981079, + "num_tokens": 181132456.0, + "step": 6998 + }, + { + "epoch": 0.7686141005930156, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4015328884124756, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6942812204360962, + "num_tokens": 181155100.0, + "step": 6999 + }, + { + "epoch": 0.7687239182956292, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.381747245788574, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7079352140426636, + "num_tokens": 181176842.0, + "step": 7000 + }, + { + "epoch": 0.7688337359982429, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.258955717086792, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7020975351333618, + "num_tokens": 181200229.0, + "step": 7001 + }, + { + "epoch": 0.7689435537008565, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.087024450302124, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.7042679786682129, + "num_tokens": 181229968.0, + "step": 7002 + }, + { + "epoch": 0.7690533714034702, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1599907875061035, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6968446969985962, + "num_tokens": 181256302.0, + "step": 7003 + }, + { + "epoch": 0.7691631891060839, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9148772954940796, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7331998348236084, + "num_tokens": 181286191.0, + "step": 7004 + }, + { + "epoch": 0.7692730068086976, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4264132976531982, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7028478980064392, + "num_tokens": 181307029.0, + "step": 7005 + }, + { + "epoch": 0.7693828245113112, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1708009243011475, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7059915661811829, + "num_tokens": 181332716.0, + "step": 7006 + }, + { + "epoch": 0.7694926422139249, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.058070182800293, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7082775235176086, + "num_tokens": 181361598.0, + "step": 7007 + }, + { + "epoch": 0.7696024599165385, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2452144622802734, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6913565397262573, + "num_tokens": 181385434.0, + "step": 7008 + }, + { + "epoch": 0.7697122776191522, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.228504180908203, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6977682113647461, + "num_tokens": 181411830.0, + "step": 7009 + }, + { + "epoch": 0.7698220953217658, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9516892433166504, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7019097805023193, + "num_tokens": 181442415.0, + "step": 7010 + }, + { + "epoch": 0.7699319130243796, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.282001495361328, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7027972936630249, + "num_tokens": 181466967.0, + "step": 7011 + }, + { + "epoch": 0.7700417307269932, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5449883937835693, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7215732336044312, + "num_tokens": 181487298.0, + "step": 7012 + }, + { + "epoch": 0.7701515484296069, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4331893920898438, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7390519976615906, + "num_tokens": 181507506.0, + "step": 7013 + }, + { + "epoch": 0.7702613661322205, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.02897310256958, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7061455249786377, + "num_tokens": 181536470.0, + "step": 7014 + }, + { + "epoch": 0.7703711838348342, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2215418815612793, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.7069994211196899, + "num_tokens": 181562596.0, + "step": 7015 + }, + { + "epoch": 0.7704810015374478, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.16857647895813, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6897454857826233, + "num_tokens": 181591441.0, + "step": 7016 + }, + { + "epoch": 0.7705908192400615, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.161271333694458, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6985605955123901, + "num_tokens": 181617441.0, + "step": 7017 + }, + { + "epoch": 0.7707006369426752, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1683566570281982, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7122336626052856, + "num_tokens": 181644105.0, + "step": 7018 + }, + { + "epoch": 0.7708104546452889, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1154136657714844, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7147639393806458, + "num_tokens": 181669368.0, + "step": 7019 + }, + { + "epoch": 0.7709202723479025, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.438068151473999, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7052595615386963, + "num_tokens": 181690185.0, + "step": 7020 + }, + { + "epoch": 0.7710300900505161, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0713465213775635, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7178728580474854, + "num_tokens": 181717010.0, + "step": 7021 + }, + { + "epoch": 0.7711399077531298, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1615867614746094, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6924090385437012, + "num_tokens": 181742561.0, + "step": 7022 + }, + { + "epoch": 0.7712497254557434, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2969393730163574, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6914098262786865, + "num_tokens": 181766667.0, + "step": 7023 + }, + { + "epoch": 0.7713595431583571, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.214707851409912, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7302143573760986, + "num_tokens": 181792346.0, + "step": 7024 + }, + { + "epoch": 0.7714693608609708, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.230192184448242, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6842798590660095, + "num_tokens": 181817478.0, + "step": 7025 + }, + { + "epoch": 0.7715791785635845, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.28214430809021, + "learning_rate": 1e-06, + "loss": 1.042, + "mean_token_accuracy": 0.6959208250045776, + "num_tokens": 181841823.0, + "step": 7026 + }, + { + "epoch": 0.7716889962661981, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.354499340057373, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.6989385485649109, + "num_tokens": 181863086.0, + "step": 7027 + }, + { + "epoch": 0.7717988139688118, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2187416553497314, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.685014009475708, + "num_tokens": 181888119.0, + "step": 7028 + }, + { + "epoch": 0.7719086316714254, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0640697479248047, + "learning_rate": 1e-06, + "loss": 1.0256, + "mean_token_accuracy": 0.6937443017959595, + "num_tokens": 181914649.0, + "step": 7029 + }, + { + "epoch": 0.7720184493740391, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.220890522003174, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.698574423789978, + "num_tokens": 181940137.0, + "step": 7030 + }, + { + "epoch": 0.7721282670766527, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2459821701049805, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7276513576507568, + "num_tokens": 181963552.0, + "step": 7031 + }, + { + "epoch": 0.7722380847792664, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.449357032775879, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7385621070861816, + "num_tokens": 181983514.0, + "step": 7032 + }, + { + "epoch": 0.7723479024818801, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0215375423431396, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7087589502334595, + "num_tokens": 182012525.0, + "step": 7033 + }, + { + "epoch": 0.7724577201844938, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2581686973571777, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6899367570877075, + "num_tokens": 182036340.0, + "step": 7034 + }, + { + "epoch": 0.7725675378871074, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.544802665710449, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7154126763343811, + "num_tokens": 182055692.0, + "step": 7035 + }, + { + "epoch": 0.7726773555897211, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.023739814758301, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7099220752716064, + "num_tokens": 182084359.0, + "step": 7036 + }, + { + "epoch": 0.7727871732923347, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.358438014984131, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6892434358596802, + "num_tokens": 182108290.0, + "step": 7037 + }, + { + "epoch": 0.7728969909949484, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.04017972946167, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.6885324120521545, + "num_tokens": 182137242.0, + "step": 7038 + }, + { + "epoch": 0.773006808697562, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.044727325439453, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6935901641845703, + "num_tokens": 182166149.0, + "step": 7039 + }, + { + "epoch": 0.7731166264001758, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.189223527908325, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7159918546676636, + "num_tokens": 182189087.0, + "step": 7040 + }, + { + "epoch": 0.7732264441027894, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3174824714660645, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6948468089103699, + "num_tokens": 182213522.0, + "step": 7041 + }, + { + "epoch": 0.773336261805403, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3445773124694824, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6878790259361267, + "num_tokens": 182236627.0, + "step": 7042 + }, + { + "epoch": 0.7734460795080167, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4415199756622314, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7316008806228638, + "num_tokens": 182259748.0, + "step": 7043 + }, + { + "epoch": 0.7735558972106303, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.130976676940918, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7059574127197266, + "num_tokens": 182288135.0, + "step": 7044 + }, + { + "epoch": 0.773665714913244, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0909366607666016, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6949065923690796, + "num_tokens": 182313404.0, + "step": 7045 + }, + { + "epoch": 0.7737755326158576, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.010664701461792, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7126216888427734, + "num_tokens": 182342822.0, + "step": 7046 + }, + { + "epoch": 0.7738853503184714, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5968267917633057, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7282079458236694, + "num_tokens": 182360758.0, + "step": 7047 + }, + { + "epoch": 0.773995168021085, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8870514631271362, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6937283873558044, + "num_tokens": 182393280.0, + "step": 7048 + }, + { + "epoch": 0.7741049857236987, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2017033100128174, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7047731876373291, + "num_tokens": 182417912.0, + "step": 7049 + }, + { + "epoch": 0.7742148034263123, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2895758152008057, + "learning_rate": 1e-06, + "loss": 1.036, + "mean_token_accuracy": 0.6797452569007874, + "num_tokens": 182442234.0, + "step": 7050 + }, + { + "epoch": 0.774324621128926, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.215374708175659, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7183845043182373, + "num_tokens": 182466737.0, + "step": 7051 + }, + { + "epoch": 0.7744344388315396, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.060534715652466, + "learning_rate": 1e-06, + "loss": 1.0442, + "mean_token_accuracy": 0.6888701915740967, + "num_tokens": 182494786.0, + "step": 7052 + }, + { + "epoch": 0.7745442565341533, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.336902618408203, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7162794470787048, + "num_tokens": 182517728.0, + "step": 7053 + }, + { + "epoch": 0.774654074236767, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.368473768234253, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7263603210449219, + "num_tokens": 182539508.0, + "step": 7054 + }, + { + "epoch": 0.7747638919393807, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1217849254608154, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6928750276565552, + "num_tokens": 182565504.0, + "step": 7055 + }, + { + "epoch": 0.7748737096419943, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.089590072631836, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6952582001686096, + "num_tokens": 182593815.0, + "step": 7056 + }, + { + "epoch": 0.774983527344608, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0516128540039062, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7014926671981812, + "num_tokens": 182621249.0, + "step": 7057 + }, + { + "epoch": 0.7750933450472216, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.211965799331665, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7240313291549683, + "num_tokens": 182645287.0, + "step": 7058 + }, + { + "epoch": 0.7752031627498353, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1588165760040283, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6976059675216675, + "num_tokens": 182670630.0, + "step": 7059 + }, + { + "epoch": 0.7753129804524489, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2094671726226807, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7270134091377258, + "num_tokens": 182693031.0, + "step": 7060 + }, + { + "epoch": 0.7754227981550625, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.110219717025757, + "learning_rate": 1e-06, + "loss": 1.0714, + "mean_token_accuracy": 0.6782684922218323, + "num_tokens": 182720336.0, + "step": 7061 + }, + { + "epoch": 0.7755326158576763, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.200869083404541, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7066474556922913, + "num_tokens": 182744926.0, + "step": 7062 + }, + { + "epoch": 0.77564243356029, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.237457036972046, + "learning_rate": 1e-06, + "loss": 1.0585, + "mean_token_accuracy": 0.6848521828651428, + "num_tokens": 182771478.0, + "step": 7063 + }, + { + "epoch": 0.7757522512629036, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3799262046813965, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.702155351638794, + "num_tokens": 182795271.0, + "step": 7064 + }, + { + "epoch": 0.7758620689655172, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1683859825134277, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.6956402063369751, + "num_tokens": 182820752.0, + "step": 7065 + }, + { + "epoch": 0.7759718866681309, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.034844160079956, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6911357641220093, + "num_tokens": 182850623.0, + "step": 7066 + }, + { + "epoch": 0.7760817043707445, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1676836013793945, + "learning_rate": 1e-06, + "loss": 1.0578, + "mean_token_accuracy": 0.6844490766525269, + "num_tokens": 182876812.0, + "step": 7067 + }, + { + "epoch": 0.7761915220733582, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.172534465789795, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7233701944351196, + "num_tokens": 182901447.0, + "step": 7068 + }, + { + "epoch": 0.7763013397759719, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5815091133117676, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.6993846297264099, + "num_tokens": 182922621.0, + "step": 7069 + }, + { + "epoch": 0.7764111574785856, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.296555995941162, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7207363843917847, + "num_tokens": 182945907.0, + "step": 7070 + }, + { + "epoch": 0.7765209751811992, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.440793991088867, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7062255144119263, + "num_tokens": 182967961.0, + "step": 7071 + }, + { + "epoch": 0.7766307928838129, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0760085582733154, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7013397216796875, + "num_tokens": 182997028.0, + "step": 7072 + }, + { + "epoch": 0.7767406105864265, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2038776874542236, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7171597480773926, + "num_tokens": 183020275.0, + "step": 7073 + }, + { + "epoch": 0.7768504282890402, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3003077507019043, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7074791789054871, + "num_tokens": 183042996.0, + "step": 7074 + }, + { + "epoch": 0.7769602459916538, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.309343099594116, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7052112817764282, + "num_tokens": 183069469.0, + "step": 7075 + }, + { + "epoch": 0.7770700636942676, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.241711139678955, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.7017174959182739, + "num_tokens": 183094259.0, + "step": 7076 + }, + { + "epoch": 0.7771798813968812, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.353424549102783, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7334144115447998, + "num_tokens": 183114980.0, + "step": 7077 + }, + { + "epoch": 0.7772896990994949, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.275561571121216, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7072957754135132, + "num_tokens": 183140267.0, + "step": 7078 + }, + { + "epoch": 0.7773995168021085, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.228093385696411, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7188257575035095, + "num_tokens": 183165152.0, + "step": 7079 + }, + { + "epoch": 0.7775093345047221, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.026195526123047, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7134568691253662, + "num_tokens": 183192657.0, + "step": 7080 + }, + { + "epoch": 0.7776191522073358, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.968429446220398, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7365738749504089, + "num_tokens": 183221786.0, + "step": 7081 + }, + { + "epoch": 0.7777289699099494, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0512824058532715, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7277998924255371, + "num_tokens": 183248432.0, + "step": 7082 + }, + { + "epoch": 0.7778387876125632, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.095733165740967, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7137043476104736, + "num_tokens": 183276389.0, + "step": 7083 + }, + { + "epoch": 0.7779486053151768, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1555821895599365, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7107222080230713, + "num_tokens": 183302609.0, + "step": 7084 + }, + { + "epoch": 0.7780584230177905, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8977620601654053, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6970176100730896, + "num_tokens": 183336102.0, + "step": 7085 + }, + { + "epoch": 0.7781682407204041, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9367846250534058, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.6984437704086304, + "num_tokens": 183365553.0, + "step": 7086 + }, + { + "epoch": 0.7782780584230178, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.133246898651123, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6974431872367859, + "num_tokens": 183391577.0, + "step": 7087 + }, + { + "epoch": 0.7783878761256314, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1509721279144287, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6962324976921082, + "num_tokens": 183416801.0, + "step": 7088 + }, + { + "epoch": 0.7784976938282451, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.535057544708252, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7350624799728394, + "num_tokens": 183435573.0, + "step": 7089 + }, + { + "epoch": 0.7786075115308587, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9780936241149902, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6909316182136536, + "num_tokens": 183467300.0, + "step": 7090 + }, + { + "epoch": 0.7787173292334725, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3177907466888428, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7078160047531128, + "num_tokens": 183490195.0, + "step": 7091 + }, + { + "epoch": 0.7788271469360861, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3885087966918945, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7023497819900513, + "num_tokens": 183514431.0, + "step": 7092 + }, + { + "epoch": 0.7789369646386998, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1889870166778564, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7419301271438599, + "num_tokens": 183539682.0, + "step": 7093 + }, + { + "epoch": 0.7790467823413134, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.33986759185791, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7180532217025757, + "num_tokens": 183562472.0, + "step": 7094 + }, + { + "epoch": 0.7791566000439271, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1279633045196533, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7076631188392639, + "num_tokens": 183589502.0, + "step": 7095 + }, + { + "epoch": 0.7792664177465407, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5072579383850098, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7006877660751343, + "num_tokens": 183615986.0, + "step": 7096 + }, + { + "epoch": 0.7793762354491544, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.102510452270508, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6934329271316528, + "num_tokens": 183642905.0, + "step": 7097 + }, + { + "epoch": 0.7794860531517681, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.499340295791626, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7169142961502075, + "num_tokens": 183664664.0, + "step": 7098 + }, + { + "epoch": 0.7795958708543818, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.264932155609131, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7293376922607422, + "num_tokens": 183688895.0, + "step": 7099 + }, + { + "epoch": 0.7797056885569954, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1419055461883545, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6986690759658813, + "num_tokens": 183716026.0, + "step": 7100 + }, + { + "epoch": 0.779815506259609, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.961021065711975, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7080919742584229, + "num_tokens": 183744988.0, + "step": 7101 + }, + { + "epoch": 0.7799253239622227, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1941230297088623, + "learning_rate": 1e-06, + "loss": 1.0641, + "mean_token_accuracy": 0.6779507398605347, + "num_tokens": 183771237.0, + "step": 7102 + }, + { + "epoch": 0.7800351416648363, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0894196033477783, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.7014739513397217, + "num_tokens": 183798604.0, + "step": 7103 + }, + { + "epoch": 0.78014495936745, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2535552978515625, + "learning_rate": 1e-06, + "loss": 1.0527, + "mean_token_accuracy": 0.6814207434654236, + "num_tokens": 183824031.0, + "step": 7104 + }, + { + "epoch": 0.7802547770700637, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3470306396484375, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.712332546710968, + "num_tokens": 183847258.0, + "step": 7105 + }, + { + "epoch": 0.7803645947726774, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1552858352661133, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7173281908035278, + "num_tokens": 183872247.0, + "step": 7106 + }, + { + "epoch": 0.780474412475291, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1358189582824707, + "learning_rate": 1e-06, + "loss": 1.0643, + "mean_token_accuracy": 0.6908111572265625, + "num_tokens": 183899771.0, + "step": 7107 + }, + { + "epoch": 0.7805842301779047, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0340774059295654, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7041141390800476, + "num_tokens": 183927184.0, + "step": 7108 + }, + { + "epoch": 0.7806940478805183, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9740126132965088, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7053749561309814, + "num_tokens": 183957129.0, + "step": 7109 + }, + { + "epoch": 0.780803865583132, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0396389961242676, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7058144211769104, + "num_tokens": 183985840.0, + "step": 7110 + }, + { + "epoch": 0.7809136832857456, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1969833374023438, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.6941375732421875, + "num_tokens": 184012366.0, + "step": 7111 + }, + { + "epoch": 0.7810235009883594, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.041130781173706, + "learning_rate": 1e-06, + "loss": 1.0311, + "mean_token_accuracy": 0.6896411180496216, + "num_tokens": 184040951.0, + "step": 7112 + }, + { + "epoch": 0.781133318690973, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1677355766296387, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7135709524154663, + "num_tokens": 184067588.0, + "step": 7113 + }, + { + "epoch": 0.7812431363935867, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3446195125579834, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7341494560241699, + "num_tokens": 184089857.0, + "step": 7114 + }, + { + "epoch": 0.7813529540962003, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3537673950195312, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7239135503768921, + "num_tokens": 184113498.0, + "step": 7115 + }, + { + "epoch": 0.781462771798814, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2957749366760254, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6926445364952087, + "num_tokens": 184137404.0, + "step": 7116 + }, + { + "epoch": 0.7815725895014276, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4265613555908203, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7001843452453613, + "num_tokens": 184159063.0, + "step": 7117 + }, + { + "epoch": 0.7816824072040413, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2340850830078125, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7146982550621033, + "num_tokens": 184181374.0, + "step": 7118 + }, + { + "epoch": 0.7817922249066549, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.248725175857544, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7186586260795593, + "num_tokens": 184204543.0, + "step": 7119 + }, + { + "epoch": 0.7819020426092687, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.350472927093506, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7129617929458618, + "num_tokens": 184227436.0, + "step": 7120 + }, + { + "epoch": 0.7820118603118823, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2283780574798584, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7164446115493774, + "num_tokens": 184252789.0, + "step": 7121 + }, + { + "epoch": 0.782121678014496, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.082313299179077, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7143228650093079, + "num_tokens": 184279574.0, + "step": 7122 + }, + { + "epoch": 0.7822314957171096, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9900238513946533, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.699881911277771, + "num_tokens": 184311071.0, + "step": 7123 + }, + { + "epoch": 0.7823413134197232, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.073251962661743, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6860291957855225, + "num_tokens": 184339486.0, + "step": 7124 + }, + { + "epoch": 0.7824511311223369, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.202366828918457, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7283520102500916, + "num_tokens": 184362691.0, + "step": 7125 + }, + { + "epoch": 0.7825609488249505, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9740391969680786, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7035001516342163, + "num_tokens": 184391019.0, + "step": 7126 + }, + { + "epoch": 0.7826707665275643, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2885892391204834, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7192615866661072, + "num_tokens": 184412939.0, + "step": 7127 + }, + { + "epoch": 0.7827805842301779, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.047607660293579, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7148370742797852, + "num_tokens": 184441180.0, + "step": 7128 + }, + { + "epoch": 0.7828904019327916, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9424564838409424, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7287644147872925, + "num_tokens": 184471774.0, + "step": 7129 + }, + { + "epoch": 0.7830002196354052, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.314145803451538, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7100446820259094, + "num_tokens": 184494838.0, + "step": 7130 + }, + { + "epoch": 0.7831100373380189, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0728020668029785, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7245270609855652, + "num_tokens": 184520574.0, + "step": 7131 + }, + { + "epoch": 0.7832198550406325, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4024527072906494, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7126074433326721, + "num_tokens": 184542020.0, + "step": 7132 + }, + { + "epoch": 0.7833296727432462, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2334632873535156, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6888972520828247, + "num_tokens": 184567106.0, + "step": 7133 + }, + { + "epoch": 0.7834394904458599, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0956883430480957, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7142501473426819, + "num_tokens": 184592293.0, + "step": 7134 + }, + { + "epoch": 0.7835493081484736, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.363375663757324, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6804466843605042, + "num_tokens": 184614675.0, + "step": 7135 + }, + { + "epoch": 0.7836591258510872, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.030747890472412, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6920269727706909, + "num_tokens": 184643311.0, + "step": 7136 + }, + { + "epoch": 0.7837689435537009, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0834434032440186, + "learning_rate": 1e-06, + "loss": 1.0511, + "mean_token_accuracy": 0.6883388757705688, + "num_tokens": 184671028.0, + "step": 7137 + }, + { + "epoch": 0.7838787612563145, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0825839042663574, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7043890953063965, + "num_tokens": 184697887.0, + "step": 7138 + }, + { + "epoch": 0.7839885789589282, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.006455898284912, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7122367024421692, + "num_tokens": 184728433.0, + "step": 7139 + }, + { + "epoch": 0.7840983966615418, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.302013635635376, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7093256711959839, + "num_tokens": 184754313.0, + "step": 7140 + }, + { + "epoch": 0.7842082143641556, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9706048965454102, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6974171996116638, + "num_tokens": 184785843.0, + "step": 7141 + }, + { + "epoch": 0.7843180320667692, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1316914558410645, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7171108722686768, + "num_tokens": 184811996.0, + "step": 7142 + }, + { + "epoch": 0.7844278497693828, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4260644912719727, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6911672353744507, + "num_tokens": 184834675.0, + "step": 7143 + }, + { + "epoch": 0.7845376674719965, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.021618366241455, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7089601755142212, + "num_tokens": 184863266.0, + "step": 7144 + }, + { + "epoch": 0.7846474851746101, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.293673276901245, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7248379588127136, + "num_tokens": 184886024.0, + "step": 7145 + }, + { + "epoch": 0.7847573028772238, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3746461868286133, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6991367340087891, + "num_tokens": 184909965.0, + "step": 7146 + }, + { + "epoch": 0.7848671205798374, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2977709770202637, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7201043367385864, + "num_tokens": 184931186.0, + "step": 7147 + }, + { + "epoch": 0.7849769382824512, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1641156673431396, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6894665360450745, + "num_tokens": 184959886.0, + "step": 7148 + }, + { + "epoch": 0.7850867559850648, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1717212200164795, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7338120937347412, + "num_tokens": 184983109.0, + "step": 7149 + }, + { + "epoch": 0.7851965736876785, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.163524627685547, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6886707544326782, + "num_tokens": 185012198.0, + "step": 7150 + }, + { + "epoch": 0.7853063913902921, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2771925926208496, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7382975816726685, + "num_tokens": 185033034.0, + "step": 7151 + }, + { + "epoch": 0.7854162090929058, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1204679012298584, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6993734836578369, + "num_tokens": 185059299.0, + "step": 7152 + }, + { + "epoch": 0.7855260267955194, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0840964317321777, + "learning_rate": 1e-06, + "loss": 1.085, + "mean_token_accuracy": 0.6670492887496948, + "num_tokens": 185087271.0, + "step": 7153 + }, + { + "epoch": 0.7856358444981331, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.093783378601074, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.7010514736175537, + "num_tokens": 185113305.0, + "step": 7154 + }, + { + "epoch": 0.7857456622007467, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4285056591033936, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7151390314102173, + "num_tokens": 185134830.0, + "step": 7155 + }, + { + "epoch": 0.7858554799033605, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3975207805633545, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7255401611328125, + "num_tokens": 185154646.0, + "step": 7156 + }, + { + "epoch": 0.7859652976059741, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9889558553695679, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.696457028388977, + "num_tokens": 185183577.0, + "step": 7157 + }, + { + "epoch": 0.7860751153085878, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4665145874023438, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7205963730812073, + "num_tokens": 185203013.0, + "step": 7158 + }, + { + "epoch": 0.7861849330112014, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.031135320663452, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.688494861125946, + "num_tokens": 185229869.0, + "step": 7159 + }, + { + "epoch": 0.786294750713815, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.185652017593384, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7003524303436279, + "num_tokens": 185255496.0, + "step": 7160 + }, + { + "epoch": 0.7864045684164287, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.02388334274292, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.693869411945343, + "num_tokens": 185284771.0, + "step": 7161 + }, + { + "epoch": 0.7865143861190423, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3970000743865967, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7238428592681885, + "num_tokens": 185306814.0, + "step": 7162 + }, + { + "epoch": 0.7866242038216561, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.293581962585449, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6907109022140503, + "num_tokens": 185331860.0, + "step": 7163 + }, + { + "epoch": 0.7867340215242697, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0524206161499023, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.7023556232452393, + "num_tokens": 185359863.0, + "step": 7164 + }, + { + "epoch": 0.7868438392268834, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1202993392944336, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6994307041168213, + "num_tokens": 185387084.0, + "step": 7165 + }, + { + "epoch": 0.786953656929497, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.146146059036255, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7278929352760315, + "num_tokens": 185413133.0, + "step": 7166 + }, + { + "epoch": 0.7870634746321107, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1345460414886475, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6950888633728027, + "num_tokens": 185437925.0, + "step": 7167 + }, + { + "epoch": 0.7871732923347243, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.146885395050049, + "learning_rate": 1e-06, + "loss": 1.0554, + "mean_token_accuracy": 0.6809617280960083, + "num_tokens": 185466086.0, + "step": 7168 + }, + { + "epoch": 0.787283110037338, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0841104984283447, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7121032476425171, + "num_tokens": 185493799.0, + "step": 7169 + }, + { + "epoch": 0.7873929277399517, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1711573600769043, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7024855613708496, + "num_tokens": 185522535.0, + "step": 7170 + }, + { + "epoch": 0.7875027454425654, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0474634170532227, + "learning_rate": 1e-06, + "loss": 1.037, + "mean_token_accuracy": 0.683417797088623, + "num_tokens": 185556122.0, + "step": 7171 + }, + { + "epoch": 0.787612563145179, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.223874807357788, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.717795729637146, + "num_tokens": 185582159.0, + "step": 7172 + }, + { + "epoch": 0.7877223808477927, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.25341796875, + "learning_rate": 1e-06, + "loss": 1.043, + "mean_token_accuracy": 0.6811540126800537, + "num_tokens": 185607770.0, + "step": 7173 + }, + { + "epoch": 0.7878321985504063, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1803677082061768, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7162542343139648, + "num_tokens": 185633153.0, + "step": 7174 + }, + { + "epoch": 0.78794201625302, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0478904247283936, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7131928205490112, + "num_tokens": 185662012.0, + "step": 7175 + }, + { + "epoch": 0.7880518339556336, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9656295776367188, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6882248520851135, + "num_tokens": 185691755.0, + "step": 7176 + }, + { + "epoch": 0.7881616516582474, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.231081247329712, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.703717052936554, + "num_tokens": 185715651.0, + "step": 7177 + }, + { + "epoch": 0.788271469360861, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.196957588195801, + "learning_rate": 1e-06, + "loss": 1.0636, + "mean_token_accuracy": 0.6792314648628235, + "num_tokens": 185741311.0, + "step": 7178 + }, + { + "epoch": 0.7883812870634747, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0396225452423096, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7231388092041016, + "num_tokens": 185768557.0, + "step": 7179 + }, + { + "epoch": 0.7884911047660883, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1969611644744873, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6916582584381104, + "num_tokens": 185795015.0, + "step": 7180 + }, + { + "epoch": 0.788600922468702, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.209615707397461, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.692802906036377, + "num_tokens": 185819116.0, + "step": 7181 + }, + { + "epoch": 0.7887107401713156, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3452718257904053, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7101699113845825, + "num_tokens": 185841426.0, + "step": 7182 + }, + { + "epoch": 0.7888205578739292, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1907074451446533, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.7035166025161743, + "num_tokens": 185866540.0, + "step": 7183 + }, + { + "epoch": 0.7889303755765429, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.193406343460083, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7040221095085144, + "num_tokens": 185892827.0, + "step": 7184 + }, + { + "epoch": 0.7890401932791566, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.450446367263794, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7290262579917908, + "num_tokens": 185913702.0, + "step": 7185 + }, + { + "epoch": 0.7891500109817703, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.170424699783325, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7147964239120483, + "num_tokens": 185939475.0, + "step": 7186 + }, + { + "epoch": 0.7892598286843839, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9361653327941895, + "learning_rate": 1e-06, + "loss": 1.03, + "mean_token_accuracy": 0.6883975267410278, + "num_tokens": 185972425.0, + "step": 7187 + }, + { + "epoch": 0.7893696463869976, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8712999820709229, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6982519626617432, + "num_tokens": 186007050.0, + "step": 7188 + }, + { + "epoch": 0.7894794640896112, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2592856884002686, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7174084782600403, + "num_tokens": 186030479.0, + "step": 7189 + }, + { + "epoch": 0.7895892817922249, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.000739336013794, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.701857328414917, + "num_tokens": 186061037.0, + "step": 7190 + }, + { + "epoch": 0.7896990994948385, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.170616626739502, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7062786817550659, + "num_tokens": 186086817.0, + "step": 7191 + }, + { + "epoch": 0.7898089171974523, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1631369590759277, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7127474546432495, + "num_tokens": 186114481.0, + "step": 7192 + }, + { + "epoch": 0.7899187349000659, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.157308340072632, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6891099214553833, + "num_tokens": 186142585.0, + "step": 7193 + }, + { + "epoch": 0.7900285526026796, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.125450611114502, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.70256507396698, + "num_tokens": 186170466.0, + "step": 7194 + }, + { + "epoch": 0.7901383703052932, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1458003520965576, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7083733081817627, + "num_tokens": 186197787.0, + "step": 7195 + }, + { + "epoch": 0.7902481880079069, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3220813274383545, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7185624241828918, + "num_tokens": 186220283.0, + "step": 7196 + }, + { + "epoch": 0.7903580057105205, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9801124334335327, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7040576934814453, + "num_tokens": 186250044.0, + "step": 7197 + }, + { + "epoch": 0.7904678234131342, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0046470165252686, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6904317140579224, + "num_tokens": 186279918.0, + "step": 7198 + }, + { + "epoch": 0.7905776411157479, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2251923084259033, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6966481804847717, + "num_tokens": 186306253.0, + "step": 7199 + }, + { + "epoch": 0.7906874588183616, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0549261569976807, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7112597227096558, + "num_tokens": 186334988.0, + "step": 7200 + }, + { + "epoch": 0.7907972765209752, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1940174102783203, + "learning_rate": 1e-06, + "loss": 1.0633, + "mean_token_accuracy": 0.6767392158508301, + "num_tokens": 186362393.0, + "step": 7201 + }, + { + "epoch": 0.7909070942235888, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9311976432800293, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7264407277107239, + "num_tokens": 186391206.0, + "step": 7202 + }, + { + "epoch": 0.7910169119262025, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9966987371444702, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7058165073394775, + "num_tokens": 186420976.0, + "step": 7203 + }, + { + "epoch": 0.7911267296288161, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1668810844421387, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7118666172027588, + "num_tokens": 186447364.0, + "step": 7204 + }, + { + "epoch": 0.7912365473314298, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5081865787506104, + "learning_rate": 1e-06, + "loss": 1.067, + "mean_token_accuracy": 0.6968916654586792, + "num_tokens": 186470996.0, + "step": 7205 + }, + { + "epoch": 0.7913463650340435, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1637184619903564, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7277472019195557, + "num_tokens": 186496877.0, + "step": 7206 + }, + { + "epoch": 0.7914561827366572, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2879865169525146, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7203563451766968, + "num_tokens": 186519265.0, + "step": 7207 + }, + { + "epoch": 0.7915660004392708, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.093268632888794, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.701522946357727, + "num_tokens": 186548132.0, + "step": 7208 + }, + { + "epoch": 0.7916758181418845, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0576136112213135, + "learning_rate": 1e-06, + "loss": 1.1181, + "mean_token_accuracy": 0.6647626161575317, + "num_tokens": 186577678.0, + "step": 7209 + }, + { + "epoch": 0.7917856358444981, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9176450967788696, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7109948396682739, + "num_tokens": 186608047.0, + "step": 7210 + }, + { + "epoch": 0.7918954535471118, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8689202070236206, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6989679336547852, + "num_tokens": 186642725.0, + "step": 7211 + }, + { + "epoch": 0.7920052712497254, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9085524082183838, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7073302865028381, + "num_tokens": 186673867.0, + "step": 7212 + }, + { + "epoch": 0.7921150889523391, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9344184398651123, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6921848058700562, + "num_tokens": 186704477.0, + "step": 7213 + }, + { + "epoch": 0.7922249066549528, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.19126033782959, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6989249587059021, + "num_tokens": 186730133.0, + "step": 7214 + }, + { + "epoch": 0.7923347243575665, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.6230077743530273, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7103691101074219, + "num_tokens": 186750994.0, + "step": 7215 + }, + { + "epoch": 0.7924445420601801, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.543334722518921, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7150192856788635, + "num_tokens": 186770534.0, + "step": 7216 + }, + { + "epoch": 0.7925543597627938, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0944883823394775, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6958983540534973, + "num_tokens": 186798520.0, + "step": 7217 + }, + { + "epoch": 0.7926641774654074, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.012413740158081, + "learning_rate": 1e-06, + "loss": 1.0848, + "mean_token_accuracy": 0.6736394762992859, + "num_tokens": 186828948.0, + "step": 7218 + }, + { + "epoch": 0.792773995168021, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1438138484954834, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6994085311889648, + "num_tokens": 186857082.0, + "step": 7219 + }, + { + "epoch": 0.7928838128706347, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4944205284118652, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6913213729858398, + "num_tokens": 186880348.0, + "step": 7220 + }, + { + "epoch": 0.7929936305732485, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5989203453063965, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7085188627243042, + "num_tokens": 186899871.0, + "step": 7221 + }, + { + "epoch": 0.7931034482758621, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1085522174835205, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6913329362869263, + "num_tokens": 186930246.0, + "step": 7222 + }, + { + "epoch": 0.7932132659784757, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1461994647979736, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7161771655082703, + "num_tokens": 186957223.0, + "step": 7223 + }, + { + "epoch": 0.7933230836810894, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.303255796432495, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7083287239074707, + "num_tokens": 186980450.0, + "step": 7224 + }, + { + "epoch": 0.793432901383703, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2599456310272217, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7139638066291809, + "num_tokens": 187004492.0, + "step": 7225 + }, + { + "epoch": 0.7935427190863167, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1507058143615723, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7178482413291931, + "num_tokens": 187030339.0, + "step": 7226 + }, + { + "epoch": 0.7936525367889303, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2331111431121826, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6902755498886108, + "num_tokens": 187057485.0, + "step": 7227 + }, + { + "epoch": 0.7937623544915441, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5030415058135986, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7090963125228882, + "num_tokens": 187077206.0, + "step": 7228 + }, + { + "epoch": 0.7938721721941577, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0932905673980713, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7116690278053284, + "num_tokens": 187104941.0, + "step": 7229 + }, + { + "epoch": 0.7939819898967714, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.392139434814453, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7230919599533081, + "num_tokens": 187127950.0, + "step": 7230 + }, + { + "epoch": 0.794091807599385, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.189703941345215, + "learning_rate": 1e-06, + "loss": 1.0384, + "mean_token_accuracy": 0.6930164098739624, + "num_tokens": 187153998.0, + "step": 7231 + }, + { + "epoch": 0.7942016253019987, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3305981159210205, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7251405715942383, + "num_tokens": 187174998.0, + "step": 7232 + }, + { + "epoch": 0.7943114430046123, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.113497734069824, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7097572088241577, + "num_tokens": 187201294.0, + "step": 7233 + }, + { + "epoch": 0.794421260707226, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0318894386291504, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.6996417045593262, + "num_tokens": 187229083.0, + "step": 7234 + }, + { + "epoch": 0.7945310784098397, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3422675132751465, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7130869626998901, + "num_tokens": 187250489.0, + "step": 7235 + }, + { + "epoch": 0.7946408961124534, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.092390298843384, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7201889753341675, + "num_tokens": 187277302.0, + "step": 7236 + }, + { + "epoch": 0.794750713815067, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.285799980163574, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7117724418640137, + "num_tokens": 187300062.0, + "step": 7237 + }, + { + "epoch": 0.7948605315176807, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.516493082046509, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7201575636863708, + "num_tokens": 187319118.0, + "step": 7238 + }, + { + "epoch": 0.7949703492202943, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.15830397605896, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7096750140190125, + "num_tokens": 187345374.0, + "step": 7239 + }, + { + "epoch": 0.795080166922908, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1213018894195557, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7282182574272156, + "num_tokens": 187370379.0, + "step": 7240 + }, + { + "epoch": 0.7951899846255216, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.165505886077881, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7203016877174377, + "num_tokens": 187395194.0, + "step": 7241 + }, + { + "epoch": 0.7952998023281352, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.311062812805176, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7042337656021118, + "num_tokens": 187418656.0, + "step": 7242 + }, + { + "epoch": 0.795409620030749, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1852333545684814, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7011868953704834, + "num_tokens": 187442710.0, + "step": 7243 + }, + { + "epoch": 0.7955194377333626, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.0816893577575684, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7129199504852295, + "num_tokens": 187470929.0, + "step": 7244 + }, + { + "epoch": 0.7956292554359763, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.995160460472107, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7179925441741943, + "num_tokens": 187499503.0, + "step": 7245 + }, + { + "epoch": 0.7957390731385899, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.229313373565674, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7235683798789978, + "num_tokens": 187523173.0, + "step": 7246 + }, + { + "epoch": 0.7958488908412036, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.035829782485962, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7095590829849243, + "num_tokens": 187551392.0, + "step": 7247 + }, + { + "epoch": 0.7959587085438172, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.8798127174377441, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6879318356513977, + "num_tokens": 187583865.0, + "step": 7248 + }, + { + "epoch": 0.7960685262464309, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3596031665802, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7154120802879333, + "num_tokens": 187605803.0, + "step": 7249 + }, + { + "epoch": 0.7961783439490446, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.6801509857177734, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7022585868835449, + "num_tokens": 187628523.0, + "step": 7250 + }, + { + "epoch": 0.7962881616516583, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5572853088378906, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.703527569770813, + "num_tokens": 187648844.0, + "step": 7251 + }, + { + "epoch": 0.7963979793542719, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.5872538089752197, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.6988569498062134, + "num_tokens": 187668170.0, + "step": 7252 + }, + { + "epoch": 0.7965077970568856, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.143618583679199, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7095388770103455, + "num_tokens": 187695231.0, + "step": 7253 + }, + { + "epoch": 0.7966176147594992, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.39452862739563, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7493857145309448, + "num_tokens": 187716433.0, + "step": 7254 + }, + { + "epoch": 0.7967274324621129, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3803634643554688, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7415024638175964, + "num_tokens": 187736921.0, + "step": 7255 + }, + { + "epoch": 0.7968372501647265, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.129711627960205, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.6994827389717102, + "num_tokens": 187761659.0, + "step": 7256 + }, + { + "epoch": 0.7969470678673403, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9475911855697632, + "learning_rate": 1e-06, + "loss": 1.0715, + "mean_token_accuracy": 0.6787710189819336, + "num_tokens": 187794860.0, + "step": 7257 + }, + { + "epoch": 0.7970568855699539, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.1037580966949463, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7038382291793823, + "num_tokens": 187822637.0, + "step": 7258 + }, + { + "epoch": 0.7971667032725676, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 1.9733725786209106, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7052599191665649, + "num_tokens": 187853797.0, + "step": 7259 + }, + { + "epoch": 0.7972765209751812, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.4622132778167725, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7169880867004395, + "num_tokens": 187874537.0, + "step": 7260 + }, + { + "epoch": 0.7973863386777948, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.128488302230835, + "learning_rate": 1e-06, + "loss": 1.0837, + "mean_token_accuracy": 0.6726246476173401, + "num_tokens": 187907296.0, + "step": 7261 + }, + { + "epoch": 0.7974961563804085, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.108991861343384, + "learning_rate": 1e-06, + "loss": 1.1164, + "mean_token_accuracy": 0.665515124797821, + "num_tokens": 187935828.0, + "step": 7262 + }, + { + "epoch": 0.7976059740830221, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.2429113388061523, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7260158061981201, + "num_tokens": 187958585.0, + "step": 7263 + }, + { + "epoch": 0.7977157917856359, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3738770484924316, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7190114259719849, + "num_tokens": 187979118.0, + "step": 7264 + }, + { + "epoch": 0.7978256094882495, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.23232102394104, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7099625468254089, + "num_tokens": 188005438.0, + "step": 7265 + }, + { + "epoch": 0.7979354271908632, + "ewc_loss": 6.29425048828125e-05, + "grad_norm": 2.3676345348358154, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7087832689285278, + "num_tokens": 188029090.0, + "step": 7266 + }, + { + "epoch": 0.7980452448934768, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1741509437561035, + "learning_rate": 1e-06, + "loss": 1.0644, + "mean_token_accuracy": 0.6798909902572632, + "num_tokens": 188053893.0, + "step": 7267 + }, + { + "epoch": 0.7981550625960905, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.175201892852783, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.7002403736114502, + "num_tokens": 188078874.0, + "step": 7268 + }, + { + "epoch": 0.7982648802987041, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1897308826446533, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7026674151420593, + "num_tokens": 188105436.0, + "step": 7269 + }, + { + "epoch": 0.7983746980013178, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.963591456413269, + "learning_rate": 1e-06, + "loss": 1.0713, + "mean_token_accuracy": 0.6886168718338013, + "num_tokens": 188139579.0, + "step": 7270 + }, + { + "epoch": 0.7984845157039314, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1843934059143066, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7190163731575012, + "num_tokens": 188165248.0, + "step": 7271 + }, + { + "epoch": 0.7985943334065452, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1060221195220947, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6955462098121643, + "num_tokens": 188192813.0, + "step": 7272 + }, + { + "epoch": 0.7987041511091588, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.508143663406372, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7168553471565247, + "num_tokens": 188212555.0, + "step": 7273 + }, + { + "epoch": 0.7988139688117725, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0978951454162598, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7041171789169312, + "num_tokens": 188240737.0, + "step": 7274 + }, + { + "epoch": 0.7989237865143861, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0011560916900635, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7071198225021362, + "num_tokens": 188268722.0, + "step": 7275 + }, + { + "epoch": 0.7990336042169998, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3004164695739746, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6852014660835266, + "num_tokens": 188295061.0, + "step": 7276 + }, + { + "epoch": 0.7991434219196134, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1356732845306396, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7181415557861328, + "num_tokens": 188322008.0, + "step": 7277 + }, + { + "epoch": 0.799253239622227, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0299949645996094, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6901183128356934, + "num_tokens": 188351369.0, + "step": 7278 + }, + { + "epoch": 0.7993630573248408, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3135769367218018, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7223048806190491, + "num_tokens": 188374342.0, + "step": 7279 + }, + { + "epoch": 0.7994728750274545, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.120002508163452, + "learning_rate": 1e-06, + "loss": 1.0724, + "mean_token_accuracy": 0.6832301020622253, + "num_tokens": 188403080.0, + "step": 7280 + }, + { + "epoch": 0.7995826927300681, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.8828712701797485, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6882239580154419, + "num_tokens": 188434808.0, + "step": 7281 + }, + { + "epoch": 0.7996925104326817, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0194437503814697, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.714005708694458, + "num_tokens": 188465375.0, + "step": 7282 + }, + { + "epoch": 0.7998023281352954, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.9432108402252197, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7101156711578369, + "num_tokens": 188495748.0, + "step": 7283 + }, + { + "epoch": 0.799912145837909, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.399739980697632, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7071113586425781, + "num_tokens": 188517345.0, + "step": 7284 + }, + { + "epoch": 0.8000219635405227, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.160578727722168, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7093806266784668, + "num_tokens": 188543347.0, + "step": 7285 + }, + { + "epoch": 0.8001317812431364, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.8503435850143433, + "learning_rate": 1e-06, + "loss": 1.0892, + "mean_token_accuracy": 0.6785150766372681, + "num_tokens": 188577493.0, + "step": 7286 + }, + { + "epoch": 0.8002415989457501, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.115976095199585, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7143178582191467, + "num_tokens": 188602905.0, + "step": 7287 + }, + { + "epoch": 0.8003514166483637, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.156635284423828, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.709560751914978, + "num_tokens": 188629470.0, + "step": 7288 + }, + { + "epoch": 0.8004612343509774, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1353654861450195, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6954988241195679, + "num_tokens": 188657305.0, + "step": 7289 + }, + { + "epoch": 0.800571052053591, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1123695373535156, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6817638278007507, + "num_tokens": 188685231.0, + "step": 7290 + }, + { + "epoch": 0.8006808697562047, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.9213998317718506, + "learning_rate": 1e-06, + "loss": 1.0632, + "mean_token_accuracy": 0.6839331388473511, + "num_tokens": 188719881.0, + "step": 7291 + }, + { + "epoch": 0.8007906874588183, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.4063596725463867, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7352643013000488, + "num_tokens": 188740271.0, + "step": 7292 + }, + { + "epoch": 0.8009005051614321, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.902117133140564, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7187265157699585, + "num_tokens": 188771459.0, + "step": 7293 + }, + { + "epoch": 0.8010103228640457, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2358524799346924, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.721319317817688, + "num_tokens": 188796163.0, + "step": 7294 + }, + { + "epoch": 0.8011201405666594, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.4650166034698486, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7104937434196472, + "num_tokens": 188818117.0, + "step": 7295 + }, + { + "epoch": 0.801229958269273, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.380352258682251, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7122718691825867, + "num_tokens": 188839835.0, + "step": 7296 + }, + { + "epoch": 0.8013397759718867, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.243553638458252, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7056448459625244, + "num_tokens": 188864548.0, + "step": 7297 + }, + { + "epoch": 0.8014495936745003, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0053839683532715, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6979627013206482, + "num_tokens": 188894610.0, + "step": 7298 + }, + { + "epoch": 0.801559411377114, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.461696147918701, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7245540618896484, + "num_tokens": 188915161.0, + "step": 7299 + }, + { + "epoch": 0.8016692290797276, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1691462993621826, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.6909558176994324, + "num_tokens": 188944775.0, + "step": 7300 + }, + { + "epoch": 0.8017790467823414, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0973782539367676, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7108384370803833, + "num_tokens": 188973964.0, + "step": 7301 + }, + { + "epoch": 0.801888864484955, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.206388473510742, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7126625776290894, + "num_tokens": 189000178.0, + "step": 7302 + }, + { + "epoch": 0.8019986821875686, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.024094820022583, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.7026344537734985, + "num_tokens": 189030193.0, + "step": 7303 + }, + { + "epoch": 0.8021084998901823, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.317128896713257, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6880981922149658, + "num_tokens": 189053274.0, + "step": 7304 + }, + { + "epoch": 0.8022183175927959, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.078099489212036, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6896380186080933, + "num_tokens": 189082363.0, + "step": 7305 + }, + { + "epoch": 0.8023281352954096, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1014182567596436, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6986266374588013, + "num_tokens": 189109955.0, + "step": 7306 + }, + { + "epoch": 0.8024379529980232, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.090628147125244, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7237948179244995, + "num_tokens": 189134395.0, + "step": 7307 + }, + { + "epoch": 0.802547770700637, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0954883098602295, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7049078941345215, + "num_tokens": 189162058.0, + "step": 7308 + }, + { + "epoch": 0.8026575884032506, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1429829597473145, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.7000969648361206, + "num_tokens": 189187227.0, + "step": 7309 + }, + { + "epoch": 0.8027674061058643, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.188403606414795, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7183615565299988, + "num_tokens": 189212907.0, + "step": 7310 + }, + { + "epoch": 0.8028772238084779, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1956698894500732, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.712146520614624, + "num_tokens": 189241018.0, + "step": 7311 + }, + { + "epoch": 0.8029870415110916, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3031766414642334, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7050765156745911, + "num_tokens": 189263735.0, + "step": 7312 + }, + { + "epoch": 0.8030968592137052, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.9848262071609497, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7244117259979248, + "num_tokens": 189293717.0, + "step": 7313 + }, + { + "epoch": 0.8032066769163189, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3325254917144775, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.7003979682922363, + "num_tokens": 189316183.0, + "step": 7314 + }, + { + "epoch": 0.8033164946189326, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1282782554626465, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7046489119529724, + "num_tokens": 189343617.0, + "step": 7315 + }, + { + "epoch": 0.8034263123215463, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1695809364318848, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6999713778495789, + "num_tokens": 189371355.0, + "step": 7316 + }, + { + "epoch": 0.8035361300241599, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.6075210571289062, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7085164785385132, + "num_tokens": 189391749.0, + "step": 7317 + }, + { + "epoch": 0.8036459477267736, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.040534734725952, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6915838718414307, + "num_tokens": 189421161.0, + "step": 7318 + }, + { + "epoch": 0.8037557654293872, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.5135409832000732, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7212883234024048, + "num_tokens": 189440703.0, + "step": 7319 + }, + { + "epoch": 0.8038655831320008, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.474229335784912, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7110064029693604, + "num_tokens": 189460939.0, + "step": 7320 + }, + { + "epoch": 0.8039754008346145, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.024137258529663, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.6964118480682373, + "num_tokens": 189492736.0, + "step": 7321 + }, + { + "epoch": 0.8040852185372283, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2602486610412598, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.703558087348938, + "num_tokens": 189517571.0, + "step": 7322 + }, + { + "epoch": 0.8041950362398419, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.098461866378784, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7138859629631042, + "num_tokens": 189544763.0, + "step": 7323 + }, + { + "epoch": 0.8043048539424555, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.097888469696045, + "learning_rate": 1e-06, + "loss": 1.0793, + "mean_token_accuracy": 0.6854169368743896, + "num_tokens": 189571734.0, + "step": 7324 + }, + { + "epoch": 0.8044146716450692, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0501277446746826, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7169288396835327, + "num_tokens": 189599544.0, + "step": 7325 + }, + { + "epoch": 0.8045244893476828, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.4849443435668945, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.695990264415741, + "num_tokens": 189620801.0, + "step": 7326 + }, + { + "epoch": 0.8046343070502965, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3124284744262695, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6958682537078857, + "num_tokens": 189642798.0, + "step": 7327 + }, + { + "epoch": 0.8047441247529101, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3731157779693604, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.714400053024292, + "num_tokens": 189667045.0, + "step": 7328 + }, + { + "epoch": 0.8048539424555239, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.196711778640747, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7172794342041016, + "num_tokens": 189693170.0, + "step": 7329 + }, + { + "epoch": 0.8049637601581375, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.431518077850342, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6965526938438416, + "num_tokens": 189717385.0, + "step": 7330 + }, + { + "epoch": 0.8050735778607512, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.244530439376831, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7067183256149292, + "num_tokens": 189743773.0, + "step": 7331 + }, + { + "epoch": 0.8051833955633648, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0417568683624268, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6844570636749268, + "num_tokens": 189775152.0, + "step": 7332 + }, + { + "epoch": 0.8052932132659785, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1047182083129883, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7380362749099731, + "num_tokens": 189801396.0, + "step": 7333 + }, + { + "epoch": 0.8054030309685921, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.388826847076416, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6921430826187134, + "num_tokens": 189824889.0, + "step": 7334 + }, + { + "epoch": 0.8055128486712058, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.5060670375823975, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7145925760269165, + "num_tokens": 189845178.0, + "step": 7335 + }, + { + "epoch": 0.8056226663738194, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.161027669906616, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7099515795707703, + "num_tokens": 189869934.0, + "step": 7336 + }, + { + "epoch": 0.8057324840764332, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1109814643859863, + "learning_rate": 1e-06, + "loss": 1.0286, + "mean_token_accuracy": 0.6909404993057251, + "num_tokens": 189897957.0, + "step": 7337 + }, + { + "epoch": 0.8058423017790468, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1054117679595947, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.6828101873397827, + "num_tokens": 189926967.0, + "step": 7338 + }, + { + "epoch": 0.8059521194816605, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.365945816040039, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7232593297958374, + "num_tokens": 189948959.0, + "step": 7339 + }, + { + "epoch": 0.8060619371842741, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.978892207145691, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7024723291397095, + "num_tokens": 189979548.0, + "step": 7340 + }, + { + "epoch": 0.8061717548868877, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3163528442382812, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6923145055770874, + "num_tokens": 190003615.0, + "step": 7341 + }, + { + "epoch": 0.8062815725895014, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.12265944480896, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7305691242218018, + "num_tokens": 190027255.0, + "step": 7342 + }, + { + "epoch": 0.806391390292115, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0690507888793945, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7070156931877136, + "num_tokens": 190054949.0, + "step": 7343 + }, + { + "epoch": 0.8065012079947288, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1486153602600098, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.73134845495224, + "num_tokens": 190077434.0, + "step": 7344 + }, + { + "epoch": 0.8066110256973424, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1606838703155518, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7098459005355835, + "num_tokens": 190105048.0, + "step": 7345 + }, + { + "epoch": 0.8067208433999561, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.9954386949539185, + "learning_rate": 1e-06, + "loss": 1.1301, + "mean_token_accuracy": 0.6666631102561951, + "num_tokens": 190135929.0, + "step": 7346 + }, + { + "epoch": 0.8068306611025697, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0130128860473633, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6927949786186218, + "num_tokens": 190164509.0, + "step": 7347 + }, + { + "epoch": 0.8069404788051834, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2665600776672363, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7235366106033325, + "num_tokens": 190186916.0, + "step": 7348 + }, + { + "epoch": 0.807050296507797, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.098344087600708, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.697415828704834, + "num_tokens": 190215352.0, + "step": 7349 + }, + { + "epoch": 0.8071601142104107, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2253496646881104, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7158059477806091, + "num_tokens": 190240627.0, + "step": 7350 + }, + { + "epoch": 0.8072699319130244, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1993658542633057, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7013998031616211, + "num_tokens": 190266976.0, + "step": 7351 + }, + { + "epoch": 0.8073797496156381, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3489911556243896, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7275876998901367, + "num_tokens": 190290414.0, + "step": 7352 + }, + { + "epoch": 0.8074895673182517, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.4525582790374756, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6930016279220581, + "num_tokens": 190314442.0, + "step": 7353 + }, + { + "epoch": 0.8075993850208654, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1036012172698975, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.728082537651062, + "num_tokens": 190341818.0, + "step": 7354 + }, + { + "epoch": 0.807709202723479, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0796215534210205, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7277101278305054, + "num_tokens": 190368152.0, + "step": 7355 + }, + { + "epoch": 0.8078190204260927, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.4230666160583496, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7079427242279053, + "num_tokens": 190387620.0, + "step": 7356 + }, + { + "epoch": 0.8079288381287063, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2246651649475098, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7087327837944031, + "num_tokens": 190411877.0, + "step": 7357 + }, + { + "epoch": 0.8080386558313201, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2367963790893555, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.716334342956543, + "num_tokens": 190436207.0, + "step": 7358 + }, + { + "epoch": 0.8081484735339337, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.8695862293243408, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7056913375854492, + "num_tokens": 190469819.0, + "step": 7359 + }, + { + "epoch": 0.8082582912365474, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0172653198242188, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.714207649230957, + "num_tokens": 190497336.0, + "step": 7360 + }, + { + "epoch": 0.808368108939161, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.065335988998413, + "learning_rate": 1e-06, + "loss": 1.0416, + "mean_token_accuracy": 0.6851770877838135, + "num_tokens": 190526504.0, + "step": 7361 + }, + { + "epoch": 0.8084779266417746, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.9780206680297852, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6980105638504028, + "num_tokens": 190559033.0, + "step": 7362 + }, + { + "epoch": 0.8085877443443883, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.648120641708374, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7299894094467163, + "num_tokens": 190576453.0, + "step": 7363 + }, + { + "epoch": 0.8086975620470019, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.3236796855926514, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.6990600228309631, + "num_tokens": 190601617.0, + "step": 7364 + }, + { + "epoch": 0.8088073797496156, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.463498115539551, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.6995123028755188, + "num_tokens": 190622769.0, + "step": 7365 + }, + { + "epoch": 0.8089171974522293, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 1.955384373664856, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.7052222490310669, + "num_tokens": 190654941.0, + "step": 7366 + }, + { + "epoch": 0.809027015154843, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2627720832824707, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6929997801780701, + "num_tokens": 190679722.0, + "step": 7367 + }, + { + "epoch": 0.8091368328574566, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.0148985385894775, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7018029093742371, + "num_tokens": 190708192.0, + "step": 7368 + }, + { + "epoch": 0.8092466505600703, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.296905755996704, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6984241008758545, + "num_tokens": 190731374.0, + "step": 7369 + }, + { + "epoch": 0.8093564682626839, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1582930088043213, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6985806226730347, + "num_tokens": 190759008.0, + "step": 7370 + }, + { + "epoch": 0.8094662859652976, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1413185596466064, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.707963764667511, + "num_tokens": 190784306.0, + "step": 7371 + }, + { + "epoch": 0.8095761036679112, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.2416930198669434, + "learning_rate": 1e-06, + "loss": 1.0722, + "mean_token_accuracy": 0.6817666292190552, + "num_tokens": 190812971.0, + "step": 7372 + }, + { + "epoch": 0.809685921370525, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.301734685897827, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7170354127883911, + "num_tokens": 190835592.0, + "step": 7373 + }, + { + "epoch": 0.8097957390731386, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.1445207595825195, + "learning_rate": 1e-06, + "loss": 1.0669, + "mean_token_accuracy": 0.6824142932891846, + "num_tokens": 190864363.0, + "step": 7374 + }, + { + "epoch": 0.8099055567757523, + "ewc_loss": 6.341934204101562e-05, + "grad_norm": 2.7169835567474365, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7073072195053101, + "num_tokens": 190882458.0, + "step": 7375 + }, + { + "epoch": 0.8100153744783659, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.1753218173980713, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6964207887649536, + "num_tokens": 190909073.0, + "step": 7376 + }, + { + "epoch": 0.8101251921809796, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.231776237487793, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7192029356956482, + "num_tokens": 190933418.0, + "step": 7377 + }, + { + "epoch": 0.8102350098835932, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.3200795650482178, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6920698881149292, + "num_tokens": 190957448.0, + "step": 7378 + }, + { + "epoch": 0.8103448275862069, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.059518575668335, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7209062576293945, + "num_tokens": 190984754.0, + "step": 7379 + }, + { + "epoch": 0.8104546452888206, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.504098415374756, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.6989601850509644, + "num_tokens": 191007254.0, + "step": 7380 + }, + { + "epoch": 0.8105644629914343, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.238408327102661, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7088824510574341, + "num_tokens": 191030203.0, + "step": 7381 + }, + { + "epoch": 0.8106742806940479, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.254254102706909, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7033295631408691, + "num_tokens": 191053705.0, + "step": 7382 + }, + { + "epoch": 0.8107840983966615, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.0902111530303955, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7110423445701599, + "num_tokens": 191081547.0, + "step": 7383 + }, + { + "epoch": 0.8108939160992752, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.754138469696045, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7264543175697327, + "num_tokens": 191100151.0, + "step": 7384 + }, + { + "epoch": 0.8110037338018888, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.012657642364502, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7168793082237244, + "num_tokens": 191128014.0, + "step": 7385 + }, + { + "epoch": 0.8111135515045025, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.118558645248413, + "learning_rate": 1e-06, + "loss": 1.0625, + "mean_token_accuracy": 0.6850008964538574, + "num_tokens": 191154703.0, + "step": 7386 + }, + { + "epoch": 0.8112233692071162, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.185154914855957, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7069640159606934, + "num_tokens": 191180285.0, + "step": 7387 + }, + { + "epoch": 0.8113331869097299, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.426378011703491, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.7033675909042358, + "num_tokens": 191201760.0, + "step": 7388 + }, + { + "epoch": 0.8114430046123435, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.094209671020508, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6978839039802551, + "num_tokens": 191231226.0, + "step": 7389 + }, + { + "epoch": 0.8115528223149572, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.350278377532959, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6846104264259338, + "num_tokens": 191254797.0, + "step": 7390 + }, + { + "epoch": 0.8116626400175708, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.3679113388061523, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7253499627113342, + "num_tokens": 191278029.0, + "step": 7391 + }, + { + "epoch": 0.8117724577201845, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 1.955973505973816, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7022574543952942, + "num_tokens": 191306205.0, + "step": 7392 + }, + { + "epoch": 0.8118822754227981, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.201829433441162, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6911027431488037, + "num_tokens": 191331158.0, + "step": 7393 + }, + { + "epoch": 0.8119920931254118, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.062446117401123, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6940780282020569, + "num_tokens": 191362487.0, + "step": 7394 + }, + { + "epoch": 0.8121019108280255, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.162602186203003, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.7015341520309448, + "num_tokens": 191389257.0, + "step": 7395 + }, + { + "epoch": 0.8122117285306392, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.2671215534210205, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7003593444824219, + "num_tokens": 191414774.0, + "step": 7396 + }, + { + "epoch": 0.8123215462332528, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.30405855178833, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7139883041381836, + "num_tokens": 191438055.0, + "step": 7397 + }, + { + "epoch": 0.8124313639358665, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.2604103088378906, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7028337121009827, + "num_tokens": 191461414.0, + "step": 7398 + }, + { + "epoch": 0.8125411816384801, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.280458688735962, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7207744717597961, + "num_tokens": 191485426.0, + "step": 7399 + }, + { + "epoch": 0.8126509993410937, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.13520884513855, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7049393057823181, + "num_tokens": 191513727.0, + "step": 7400 + }, + { + "epoch": 0.8127608170437074, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.222214698791504, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6869553327560425, + "num_tokens": 191541044.0, + "step": 7401 + }, + { + "epoch": 0.8128706347463212, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 1.7596325874328613, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7330176830291748, + "num_tokens": 191575243.0, + "step": 7402 + }, + { + "epoch": 0.8129804524489348, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.1680049896240234, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7147390842437744, + "num_tokens": 191599817.0, + "step": 7403 + }, + { + "epoch": 0.8130902701515484, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 1.9373151063919067, + "learning_rate": 1e-06, + "loss": 1.0788, + "mean_token_accuracy": 0.6741560697555542, + "num_tokens": 191633627.0, + "step": 7404 + }, + { + "epoch": 0.8132000878541621, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.1619884967803955, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6894283294677734, + "num_tokens": 191661966.0, + "step": 7405 + }, + { + "epoch": 0.8133099055567757, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.124443769454956, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.7026119828224182, + "num_tokens": 191689951.0, + "step": 7406 + }, + { + "epoch": 0.8134197232593894, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.182882785797119, + "learning_rate": 1e-06, + "loss": 1.0588, + "mean_token_accuracy": 0.6831458210945129, + "num_tokens": 191716482.0, + "step": 7407 + }, + { + "epoch": 0.813529540962003, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.284747362136841, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7037297487258911, + "num_tokens": 191741044.0, + "step": 7408 + }, + { + "epoch": 0.8136393586646168, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.0654757022857666, + "learning_rate": 1e-06, + "loss": 1.113, + "mean_token_accuracy": 0.6735248565673828, + "num_tokens": 191772667.0, + "step": 7409 + }, + { + "epoch": 0.8137491763672304, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.0031423568725586, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7121178507804871, + "num_tokens": 191801028.0, + "step": 7410 + }, + { + "epoch": 0.8138589940698441, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.1893422603607178, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7062374353408813, + "num_tokens": 191825693.0, + "step": 7411 + }, + { + "epoch": 0.8139688117724577, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.1672682762145996, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7177361845970154, + "num_tokens": 191850044.0, + "step": 7412 + }, + { + "epoch": 0.8140786294750714, + "ewc_loss": 6.389617919921875e-05, + "grad_norm": 2.3370537757873535, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7086396217346191, + "num_tokens": 191873183.0, + "step": 7413 + }, + { + "epoch": 0.814188447177685, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.060044765472412, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7083200216293335, + "num_tokens": 191899944.0, + "step": 7414 + }, + { + "epoch": 0.8142982648802987, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.116758108139038, + "learning_rate": 1e-06, + "loss": 1.0621, + "mean_token_accuracy": 0.6782729625701904, + "num_tokens": 191927847.0, + "step": 7415 + }, + { + "epoch": 0.8144080825829124, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.184617280960083, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7045536041259766, + "num_tokens": 191953831.0, + "step": 7416 + }, + { + "epoch": 0.8145179002855261, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.19932222366333, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.6992777585983276, + "num_tokens": 191979703.0, + "step": 7417 + }, + { + "epoch": 0.8146277179881397, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.065094470977783, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7234722375869751, + "num_tokens": 192007014.0, + "step": 7418 + }, + { + "epoch": 0.8147375356907534, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.1143600940704346, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6862607002258301, + "num_tokens": 192034396.0, + "step": 7419 + }, + { + "epoch": 0.814847353393367, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9838544130325317, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6904815435409546, + "num_tokens": 192065478.0, + "step": 7420 + }, + { + "epoch": 0.8149571710959806, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.1802783012390137, + "learning_rate": 1e-06, + "loss": 1.0515, + "mean_token_accuracy": 0.692611813545227, + "num_tokens": 192091541.0, + "step": 7421 + }, + { + "epoch": 0.8150669887985943, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.371521234512329, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7194491028785706, + "num_tokens": 192113111.0, + "step": 7422 + }, + { + "epoch": 0.8151768065012079, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.192138433456421, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6951311826705933, + "num_tokens": 192139874.0, + "step": 7423 + }, + { + "epoch": 0.8152866242038217, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9564274549484253, + "learning_rate": 1e-06, + "loss": 1.0601, + "mean_token_accuracy": 0.6862661838531494, + "num_tokens": 192173021.0, + "step": 7424 + }, + { + "epoch": 0.8153964419064353, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.086878538131714, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7263890504837036, + "num_tokens": 192200042.0, + "step": 7425 + }, + { + "epoch": 0.815506259609049, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9339510202407837, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.704301118850708, + "num_tokens": 192233652.0, + "step": 7426 + }, + { + "epoch": 0.8156160773116626, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.211033821105957, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.6940795183181763, + "num_tokens": 192259177.0, + "step": 7427 + }, + { + "epoch": 0.8157258950142763, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.032496452331543, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6901313662528992, + "num_tokens": 192289215.0, + "step": 7428 + }, + { + "epoch": 0.8158357127168899, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.1801819801330566, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7280482053756714, + "num_tokens": 192311361.0, + "step": 7429 + }, + { + "epoch": 0.8159455304195036, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.173588514328003, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7185488343238831, + "num_tokens": 192336211.0, + "step": 7430 + }, + { + "epoch": 0.8160553481221173, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.2398149967193604, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7146570682525635, + "num_tokens": 192359588.0, + "step": 7431 + }, + { + "epoch": 0.816165165824731, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.1653223037719727, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6930846571922302, + "num_tokens": 192387109.0, + "step": 7432 + }, + { + "epoch": 0.8162749835273446, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.177257776260376, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7041640281677246, + "num_tokens": 192414021.0, + "step": 7433 + }, + { + "epoch": 0.8163848012299583, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.966577172279358, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6945894956588745, + "num_tokens": 192444406.0, + "step": 7434 + }, + { + "epoch": 0.8164946189325719, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.06270432472229, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6894521713256836, + "num_tokens": 192472462.0, + "step": 7435 + }, + { + "epoch": 0.8166044366351856, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.08471941947937, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.6992415189743042, + "num_tokens": 192501003.0, + "step": 7436 + }, + { + "epoch": 0.8167142543377992, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.3310983180999756, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7222204208374023, + "num_tokens": 192521744.0, + "step": 7437 + }, + { + "epoch": 0.816824072040413, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9203084707260132, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7206872701644897, + "num_tokens": 192551757.0, + "step": 7438 + }, + { + "epoch": 0.8169338897430266, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9167029857635498, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6915491223335266, + "num_tokens": 192585787.0, + "step": 7439 + }, + { + "epoch": 0.8170437074456403, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.5231313705444336, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.719253659248352, + "num_tokens": 192605935.0, + "step": 7440 + }, + { + "epoch": 0.8171535251482539, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.255866527557373, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7129544019699097, + "num_tokens": 192629921.0, + "step": 7441 + }, + { + "epoch": 0.8172633428508675, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.9701247215270996, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7183977365493774, + "num_tokens": 192660071.0, + "step": 7442 + }, + { + "epoch": 0.8173731605534812, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.392673969268799, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6789969801902771, + "num_tokens": 192686540.0, + "step": 7443 + }, + { + "epoch": 0.8174829782560948, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.0130631923675537, + "learning_rate": 1e-06, + "loss": 1.0512, + "mean_token_accuracy": 0.6892854571342468, + "num_tokens": 192718722.0, + "step": 7444 + }, + { + "epoch": 0.8175927959587086, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.3929924964904785, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6984761953353882, + "num_tokens": 192743632.0, + "step": 7445 + }, + { + "epoch": 0.8177026136613222, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.1782050132751465, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7319513559341431, + "num_tokens": 192768444.0, + "step": 7446 + }, + { + "epoch": 0.8178124313639359, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.2417056560516357, + "learning_rate": 1e-06, + "loss": 1.0679, + "mean_token_accuracy": 0.6869537830352783, + "num_tokens": 192795814.0, + "step": 7447 + }, + { + "epoch": 0.8179222490665495, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.073751926422119, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7092294096946716, + "num_tokens": 192821027.0, + "step": 7448 + }, + { + "epoch": 0.8180320667691632, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 2.19754958152771, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6916325092315674, + "num_tokens": 192846075.0, + "step": 7449 + }, + { + "epoch": 0.8181418844717768, + "ewc_loss": 6.437301635742188e-05, + "grad_norm": 1.8814860582351685, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7183278203010559, + "num_tokens": 192875195.0, + "step": 7450 + }, + { + "epoch": 0.8182517021743905, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1985249519348145, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7029885053634644, + "num_tokens": 192899409.0, + "step": 7451 + }, + { + "epoch": 0.8183615198770041, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1049551963806152, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7087879180908203, + "num_tokens": 192924224.0, + "step": 7452 + }, + { + "epoch": 0.8184713375796179, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.5052170753479004, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7059457302093506, + "num_tokens": 192944988.0, + "step": 7453 + }, + { + "epoch": 0.8185811552822315, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.3116207122802734, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7306824922561646, + "num_tokens": 192967089.0, + "step": 7454 + }, + { + "epoch": 0.8186909729848452, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1645617485046387, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7043724060058594, + "num_tokens": 192993389.0, + "step": 7455 + }, + { + "epoch": 0.8188007906874588, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.318138360977173, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7101058959960938, + "num_tokens": 193016217.0, + "step": 7456 + }, + { + "epoch": 0.8189106083900725, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1263790130615234, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6970163583755493, + "num_tokens": 193043656.0, + "step": 7457 + }, + { + "epoch": 0.8190204260926861, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.332166910171509, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.710591197013855, + "num_tokens": 193066359.0, + "step": 7458 + }, + { + "epoch": 0.8191302437952998, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.070578098297119, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6913425326347351, + "num_tokens": 193096540.0, + "step": 7459 + }, + { + "epoch": 0.8192400614979135, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.324958324432373, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7023279070854187, + "num_tokens": 193119171.0, + "step": 7460 + }, + { + "epoch": 0.8193498792005272, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9907430410385132, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7005503177642822, + "num_tokens": 193149032.0, + "step": 7461 + }, + { + "epoch": 0.8194596969031408, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.127124547958374, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7097725868225098, + "num_tokens": 193174665.0, + "step": 7462 + }, + { + "epoch": 0.8195695146057544, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.317436695098877, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7093709111213684, + "num_tokens": 193198821.0, + "step": 7463 + }, + { + "epoch": 0.8196793323083681, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.324039936065674, + "learning_rate": 1e-06, + "loss": 0.7836, + "mean_token_accuracy": 0.7529077529907227, + "num_tokens": 193219550.0, + "step": 7464 + }, + { + "epoch": 0.8197891500109817, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1261844635009766, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.718551754951477, + "num_tokens": 193245055.0, + "step": 7465 + }, + { + "epoch": 0.8198989677135954, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9345262050628662, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.716828465461731, + "num_tokens": 193274167.0, + "step": 7466 + }, + { + "epoch": 0.8200087854162091, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.247810125350952, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7233383655548096, + "num_tokens": 193297555.0, + "step": 7467 + }, + { + "epoch": 0.8201186031188228, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.14050555229187, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7100622653961182, + "num_tokens": 193324070.0, + "step": 7468 + }, + { + "epoch": 0.8202284208214364, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4044082164764404, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6982938051223755, + "num_tokens": 193347653.0, + "step": 7469 + }, + { + "epoch": 0.8203382385240501, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1771905422210693, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7103525400161743, + "num_tokens": 193372170.0, + "step": 7470 + }, + { + "epoch": 0.8204480562266637, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.13496732711792, + "learning_rate": 1e-06, + "loss": 1.0967, + "mean_token_accuracy": 0.6689761877059937, + "num_tokens": 193404706.0, + "step": 7471 + }, + { + "epoch": 0.8205578739292774, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1257898807525635, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7168282270431519, + "num_tokens": 193431115.0, + "step": 7472 + }, + { + "epoch": 0.820667691631891, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1916046142578125, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6945804357528687, + "num_tokens": 193458507.0, + "step": 7473 + }, + { + "epoch": 0.8207775093345048, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0702638626098633, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.7016834020614624, + "num_tokens": 193488381.0, + "step": 7474 + }, + { + "epoch": 0.8208873270371184, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0409748554229736, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.697060763835907, + "num_tokens": 193519142.0, + "step": 7475 + }, + { + "epoch": 0.8209971447397321, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2688839435577393, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7164733409881592, + "num_tokens": 193543492.0, + "step": 7476 + }, + { + "epoch": 0.8211069624423457, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.228935718536377, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6982455849647522, + "num_tokens": 193567255.0, + "step": 7477 + }, + { + "epoch": 0.8212167801449594, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4365899562835693, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.710508406162262, + "num_tokens": 193588674.0, + "step": 7478 + }, + { + "epoch": 0.821326597847573, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0748775005340576, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.695917010307312, + "num_tokens": 193617385.0, + "step": 7479 + }, + { + "epoch": 0.8214364155501866, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1434903144836426, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7329323291778564, + "num_tokens": 193641889.0, + "step": 7480 + }, + { + "epoch": 0.8215462332528004, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4439709186553955, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6913317441940308, + "num_tokens": 193663586.0, + "step": 7481 + }, + { + "epoch": 0.821656050955414, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.403970718383789, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7090502977371216, + "num_tokens": 193685218.0, + "step": 7482 + }, + { + "epoch": 0.8217658686580277, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.228872299194336, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7188832759857178, + "num_tokens": 193708146.0, + "step": 7483 + }, + { + "epoch": 0.8218756863606413, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.475409507751465, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7134217619895935, + "num_tokens": 193727761.0, + "step": 7484 + }, + { + "epoch": 0.821985504063255, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2522261142730713, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.700776219367981, + "num_tokens": 193751345.0, + "step": 7485 + }, + { + "epoch": 0.8220953217658686, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0559074878692627, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7300305366516113, + "num_tokens": 193781026.0, + "step": 7486 + }, + { + "epoch": 0.8222051394684823, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1801652908325195, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7140426635742188, + "num_tokens": 193807025.0, + "step": 7487 + }, + { + "epoch": 0.8223149571710959, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.113365888595581, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7332131862640381, + "num_tokens": 193831146.0, + "step": 7488 + }, + { + "epoch": 0.8224247748737097, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.203587532043457, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7308341264724731, + "num_tokens": 193856325.0, + "step": 7489 + }, + { + "epoch": 0.8225345925763233, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1483051776885986, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.6992741227149963, + "num_tokens": 193881662.0, + "step": 7490 + }, + { + "epoch": 0.822644410278937, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.152757167816162, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.702343225479126, + "num_tokens": 193909299.0, + "step": 7491 + }, + { + "epoch": 0.8227542279815506, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.8666448593139648, + "learning_rate": 1e-06, + "loss": 1.1087, + "mean_token_accuracy": 0.6686937808990479, + "num_tokens": 193945491.0, + "step": 7492 + }, + { + "epoch": 0.8228640456841643, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9795119762420654, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.6981954574584961, + "num_tokens": 193972892.0, + "step": 7493 + }, + { + "epoch": 0.8229738633867779, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4144458770751953, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6986445188522339, + "num_tokens": 193994630.0, + "step": 7494 + }, + { + "epoch": 0.8230836810893916, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.342334032058716, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7255195379257202, + "num_tokens": 194016765.0, + "step": 7495 + }, + { + "epoch": 0.8231934987920053, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1272201538085938, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7145929932594299, + "num_tokens": 194041630.0, + "step": 7496 + }, + { + "epoch": 0.823303316494619, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.6865689754486084, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7044594883918762, + "num_tokens": 194059924.0, + "step": 7497 + }, + { + "epoch": 0.8234131341972326, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1022162437438965, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7054111957550049, + "num_tokens": 194087610.0, + "step": 7498 + }, + { + "epoch": 0.8235229518998463, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1410984992980957, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7206758260726929, + "num_tokens": 194112974.0, + "step": 7499 + }, + { + "epoch": 0.8236327696024599, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.8950223922729492, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7250491976737976, + "num_tokens": 194145084.0, + "step": 7500 + }, + { + "epoch": 0.8237425873050735, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0238521099090576, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.709013819694519, + "num_tokens": 194170949.0, + "step": 7501 + }, + { + "epoch": 0.8238524050076872, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.331613063812256, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7096976041793823, + "num_tokens": 194194584.0, + "step": 7502 + }, + { + "epoch": 0.823962222710301, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1915245056152344, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7017259001731873, + "num_tokens": 194219238.0, + "step": 7503 + }, + { + "epoch": 0.8240720404129146, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.171553134918213, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7183276414871216, + "num_tokens": 194245490.0, + "step": 7504 + }, + { + "epoch": 0.8241818581155282, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0131306648254395, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6981642842292786, + "num_tokens": 194277489.0, + "step": 7505 + }, + { + "epoch": 0.8242916758181419, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0139646530151367, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7267395257949829, + "num_tokens": 194304550.0, + "step": 7506 + }, + { + "epoch": 0.8244014935207555, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.295919895172119, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.724399209022522, + "num_tokens": 194328241.0, + "step": 7507 + }, + { + "epoch": 0.8245113112233692, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2867531776428223, + "learning_rate": 1e-06, + "loss": 1.0892, + "mean_token_accuracy": 0.6894319653511047, + "num_tokens": 194353809.0, + "step": 7508 + }, + { + "epoch": 0.8246211289259828, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4441277980804443, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.741539716720581, + "num_tokens": 194374441.0, + "step": 7509 + }, + { + "epoch": 0.8247309466285966, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.109773635864258, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7172707319259644, + "num_tokens": 194401876.0, + "step": 7510 + }, + { + "epoch": 0.8248407643312102, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.3601605892181396, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7277898788452148, + "num_tokens": 194423064.0, + "step": 7511 + }, + { + "epoch": 0.8249505820338239, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1296489238739014, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7206111550331116, + "num_tokens": 194449636.0, + "step": 7512 + }, + { + "epoch": 0.8250603997364375, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.142733573913574, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7174947261810303, + "num_tokens": 194476248.0, + "step": 7513 + }, + { + "epoch": 0.8251702174390512, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.127415657043457, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7003575563430786, + "num_tokens": 194504609.0, + "step": 7514 + }, + { + "epoch": 0.8252800351416648, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2641451358795166, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7098333835601807, + "num_tokens": 194526147.0, + "step": 7515 + }, + { + "epoch": 0.8253898528442785, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1419148445129395, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.7022252082824707, + "num_tokens": 194552878.0, + "step": 7516 + }, + { + "epoch": 0.8254996705468921, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.051954746246338, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.7003453373908997, + "num_tokens": 194581847.0, + "step": 7517 + }, + { + "epoch": 0.8256094882495059, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9821199178695679, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7064355611801147, + "num_tokens": 194610977.0, + "step": 7518 + }, + { + "epoch": 0.8257193059521195, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.238281011581421, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7556787729263306, + "num_tokens": 194633139.0, + "step": 7519 + }, + { + "epoch": 0.8258291236547332, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2103347778320312, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6964300870895386, + "num_tokens": 194658417.0, + "step": 7520 + }, + { + "epoch": 0.8259389413573468, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.30659556388855, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7072934508323669, + "num_tokens": 194682723.0, + "step": 7521 + }, + { + "epoch": 0.8260487590599604, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.215261936187744, + "learning_rate": 1e-06, + "loss": 1.086, + "mean_token_accuracy": 0.6744741201400757, + "num_tokens": 194709719.0, + "step": 7522 + }, + { + "epoch": 0.8261585767625741, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.262430429458618, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7234182953834534, + "num_tokens": 194735545.0, + "step": 7523 + }, + { + "epoch": 0.8262683944651877, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.080760955810547, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7100837230682373, + "num_tokens": 194764463.0, + "step": 7524 + }, + { + "epoch": 0.8263782121678015, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.23812198638916, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.7064018845558167, + "num_tokens": 194788477.0, + "step": 7525 + }, + { + "epoch": 0.8264880298704151, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9961879253387451, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7145289778709412, + "num_tokens": 194818075.0, + "step": 7526 + }, + { + "epoch": 0.8265978475730288, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1258151531219482, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7115710377693176, + "num_tokens": 194844525.0, + "step": 7527 + }, + { + "epoch": 0.8267076652756424, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9246785640716553, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.692401647567749, + "num_tokens": 194875526.0, + "step": 7528 + }, + { + "epoch": 0.8268174829782561, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0595808029174805, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7064825892448425, + "num_tokens": 194902455.0, + "step": 7529 + }, + { + "epoch": 0.8269273006808697, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2820372581481934, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7284085750579834, + "num_tokens": 194924955.0, + "step": 7530 + }, + { + "epoch": 0.8270371183834834, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9525920152664185, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.716274619102478, + "num_tokens": 194956923.0, + "step": 7531 + }, + { + "epoch": 0.8271469360860971, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1061978340148926, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7124730944633484, + "num_tokens": 194982918.0, + "step": 7532 + }, + { + "epoch": 0.8272567537887108, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.219449520111084, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.6970887184143066, + "num_tokens": 195006390.0, + "step": 7533 + }, + { + "epoch": 0.8273665714913244, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2012839317321777, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7222591638565063, + "num_tokens": 195030450.0, + "step": 7534 + }, + { + "epoch": 0.8274763891939381, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1278183460235596, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6860741376876831, + "num_tokens": 195057157.0, + "step": 7535 + }, + { + "epoch": 0.8275862068965517, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1472926139831543, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.6883882880210876, + "num_tokens": 195086748.0, + "step": 7536 + }, + { + "epoch": 0.8276960245991654, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.986326813697815, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7049504518508911, + "num_tokens": 195116699.0, + "step": 7537 + }, + { + "epoch": 0.827805842301779, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.048521041870117, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7094827890396118, + "num_tokens": 195144108.0, + "step": 7538 + }, + { + "epoch": 0.8279156600043928, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2162959575653076, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7127728462219238, + "num_tokens": 195167984.0, + "step": 7539 + }, + { + "epoch": 0.8280254777070064, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.626544952392578, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7420094609260559, + "num_tokens": 195186092.0, + "step": 7540 + }, + { + "epoch": 0.82813529540962, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0811941623687744, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7044651508331299, + "num_tokens": 195213490.0, + "step": 7541 + }, + { + "epoch": 0.8282451131122337, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.487730026245117, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.706305980682373, + "num_tokens": 195235215.0, + "step": 7542 + }, + { + "epoch": 0.8283549308148473, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.9733474254608154, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6888529658317566, + "num_tokens": 195268719.0, + "step": 7543 + }, + { + "epoch": 0.828464748517461, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0203731060028076, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6856461763381958, + "num_tokens": 195302496.0, + "step": 7544 + }, + { + "epoch": 0.8285745662200746, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2077903747558594, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.7053251266479492, + "num_tokens": 195328694.0, + "step": 7545 + }, + { + "epoch": 0.8286843839226883, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.20253849029541, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7170575261116028, + "num_tokens": 195352880.0, + "step": 7546 + }, + { + "epoch": 0.828794201625302, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 1.929909110069275, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7354080677032471, + "num_tokens": 195381758.0, + "step": 7547 + }, + { + "epoch": 0.8289040193279157, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2437384128570557, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6940613985061646, + "num_tokens": 195406281.0, + "step": 7548 + }, + { + "epoch": 0.8290138370305293, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2331840991973877, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6936044692993164, + "num_tokens": 195431651.0, + "step": 7549 + }, + { + "epoch": 0.829123654733143, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.04122257232666, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7127161026000977, + "num_tokens": 195460148.0, + "step": 7550 + }, + { + "epoch": 0.8292334724357566, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.1830694675445557, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7460824847221375, + "num_tokens": 195484865.0, + "step": 7551 + }, + { + "epoch": 0.8293432901383703, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2047786712646484, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7119608521461487, + "num_tokens": 195510871.0, + "step": 7552 + }, + { + "epoch": 0.8294531078409839, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0678837299346924, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6969279050827026, + "num_tokens": 195540329.0, + "step": 7553 + }, + { + "epoch": 0.8295629255435977, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.2187488079071045, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7304427623748779, + "num_tokens": 195563569.0, + "step": 7554 + }, + { + "epoch": 0.8296727432462113, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.5273935794830322, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7118537425994873, + "num_tokens": 195584615.0, + "step": 7555 + }, + { + "epoch": 0.829782560948825, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.0685617923736572, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.685707688331604, + "num_tokens": 195612512.0, + "step": 7556 + }, + { + "epoch": 0.8298923786514386, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.489745855331421, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7169661521911621, + "num_tokens": 195632874.0, + "step": 7557 + }, + { + "epoch": 0.8300021963540523, + "ewc_loss": 6.4849853515625e-05, + "grad_norm": 2.4274349212646484, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6953673958778381, + "num_tokens": 195655467.0, + "step": 7558 + }, + { + "epoch": 0.8301120140566659, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.9381110668182373, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.7077151536941528, + "num_tokens": 195687886.0, + "step": 7559 + }, + { + "epoch": 0.8302218317592795, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.304142713546753, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7059742212295532, + "num_tokens": 195711794.0, + "step": 7560 + }, + { + "epoch": 0.8303316494618933, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.045755624771118, + "learning_rate": 1e-06, + "loss": 1.071, + "mean_token_accuracy": 0.6798588037490845, + "num_tokens": 195742931.0, + "step": 7561 + }, + { + "epoch": 0.830441467164507, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.558521270751953, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7172299027442932, + "num_tokens": 195762619.0, + "step": 7562 + }, + { + "epoch": 0.8305512848671206, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.2643918991088867, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7296240329742432, + "num_tokens": 195784610.0, + "step": 7563 + }, + { + "epoch": 0.8306611025697342, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1428158283233643, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.7053027749061584, + "num_tokens": 195810897.0, + "step": 7564 + }, + { + "epoch": 0.8307709202723479, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.8447939157485962, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.687006950378418, + "num_tokens": 195846421.0, + "step": 7565 + }, + { + "epoch": 0.8308807379749615, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.338639497756958, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7018694877624512, + "num_tokens": 195868398.0, + "step": 7566 + }, + { + "epoch": 0.8309905556775752, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.3952507972717285, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.6979721188545227, + "num_tokens": 195889234.0, + "step": 7567 + }, + { + "epoch": 0.8311003733801889, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.5394389629364014, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7054625749588013, + "num_tokens": 195908887.0, + "step": 7568 + }, + { + "epoch": 0.8312101910828026, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.445183753967285, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6921932697296143, + "num_tokens": 195932480.0, + "step": 7569 + }, + { + "epoch": 0.8313200087854162, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.211280107498169, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7038307785987854, + "num_tokens": 195959682.0, + "step": 7570 + }, + { + "epoch": 0.8314298264880299, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.9717386960983276, + "learning_rate": 1e-06, + "loss": 1.0335, + "mean_token_accuracy": 0.688997745513916, + "num_tokens": 195992913.0, + "step": 7571 + }, + { + "epoch": 0.8315396441906435, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.393139123916626, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.710350513458252, + "num_tokens": 196013906.0, + "step": 7572 + }, + { + "epoch": 0.8316494618932572, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.3565361499786377, + "learning_rate": 1e-06, + "loss": 1.0251, + "mean_token_accuracy": 0.6863783001899719, + "num_tokens": 196039076.0, + "step": 7573 + }, + { + "epoch": 0.8317592795958708, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.9632588624954224, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7026392221450806, + "num_tokens": 196070521.0, + "step": 7574 + }, + { + "epoch": 0.8318690972984845, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1286840438842773, + "learning_rate": 1e-06, + "loss": 1.0664, + "mean_token_accuracy": 0.6786413192749023, + "num_tokens": 196097730.0, + "step": 7575 + }, + { + "epoch": 0.8319789150010982, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.314716100692749, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7395800352096558, + "num_tokens": 196117974.0, + "step": 7576 + }, + { + "epoch": 0.8320887327037119, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.131579875946045, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.7051995992660522, + "num_tokens": 196146404.0, + "step": 7577 + }, + { + "epoch": 0.8321985504063255, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.2367846965789795, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7146111130714417, + "num_tokens": 196171219.0, + "step": 7578 + }, + { + "epoch": 0.8323083681089392, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.3877675533294678, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6853516101837158, + "num_tokens": 196194458.0, + "step": 7579 + }, + { + "epoch": 0.8324181858115528, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.143322467803955, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7254176139831543, + "num_tokens": 196219471.0, + "step": 7580 + }, + { + "epoch": 0.8325280035141664, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1949164867401123, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6950327157974243, + "num_tokens": 196243947.0, + "step": 7581 + }, + { + "epoch": 0.8326378212167801, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1254723072052, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7022966146469116, + "num_tokens": 196270203.0, + "step": 7582 + }, + { + "epoch": 0.8327476389193939, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1794064044952393, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7112441062927246, + "num_tokens": 196297066.0, + "step": 7583 + }, + { + "epoch": 0.8328574566220075, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.3602898120880127, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7141376733779907, + "num_tokens": 196318379.0, + "step": 7584 + }, + { + "epoch": 0.8329672743246211, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.4619522094726562, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7185161709785461, + "num_tokens": 196338137.0, + "step": 7585 + }, + { + "epoch": 0.8330770920272348, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.0490245819091797, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.6971067190170288, + "num_tokens": 196367424.0, + "step": 7586 + }, + { + "epoch": 0.8331869097298484, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.2194199562072754, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7173452377319336, + "num_tokens": 196391314.0, + "step": 7587 + }, + { + "epoch": 0.8332967274324621, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1669363975524902, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7010648250579834, + "num_tokens": 196416374.0, + "step": 7588 + }, + { + "epoch": 0.8334065451350757, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 2.1389729976654053, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7305690050125122, + "num_tokens": 196441639.0, + "step": 7589 + }, + { + "epoch": 0.8335163628376895, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.9080461263656616, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7162758111953735, + "num_tokens": 196471450.0, + "step": 7590 + }, + { + "epoch": 0.8336261805403031, + "ewc_loss": 6.532669067382812e-05, + "grad_norm": 1.9745553731918335, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6893591284751892, + "num_tokens": 196500994.0, + "step": 7591 + }, + { + "epoch": 0.8337359982429168, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.107050657272339, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7116990089416504, + "num_tokens": 196526947.0, + "step": 7592 + }, + { + "epoch": 0.8338458159455304, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3168692588806152, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7088418006896973, + "num_tokens": 196550249.0, + "step": 7593 + }, + { + "epoch": 0.8339556336481441, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.263253688812256, + "learning_rate": 1e-06, + "loss": 1.0724, + "mean_token_accuracy": 0.6869360208511353, + "num_tokens": 196577281.0, + "step": 7594 + }, + { + "epoch": 0.8340654513507577, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.202260732650757, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7077337503433228, + "num_tokens": 196602095.0, + "step": 7595 + }, + { + "epoch": 0.8341752690533714, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9917960166931152, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6954498887062073, + "num_tokens": 196629629.0, + "step": 7596 + }, + { + "epoch": 0.8342850867559851, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.457010269165039, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7068502902984619, + "num_tokens": 196652009.0, + "step": 7597 + }, + { + "epoch": 0.8343949044585988, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.163386821746826, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.724486768245697, + "num_tokens": 196677268.0, + "step": 7598 + }, + { + "epoch": 0.8345047221612124, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2246313095092773, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6972498893737793, + "num_tokens": 196703110.0, + "step": 7599 + }, + { + "epoch": 0.834614539863826, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.855188012123108, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7023699879646301, + "num_tokens": 196735845.0, + "step": 7600 + }, + { + "epoch": 0.8347243575664397, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.081587791442871, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.703626275062561, + "num_tokens": 196764382.0, + "step": 7601 + }, + { + "epoch": 0.8348341752690533, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3499462604522705, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7224687933921814, + "num_tokens": 196785955.0, + "step": 7602 + }, + { + "epoch": 0.834943992971667, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0495223999023438, + "learning_rate": 1e-06, + "loss": 1.0889, + "mean_token_accuracy": 0.6762968301773071, + "num_tokens": 196815121.0, + "step": 7603 + }, + { + "epoch": 0.8350538106742806, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9088242053985596, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7132588028907776, + "num_tokens": 196844968.0, + "step": 7604 + }, + { + "epoch": 0.8351636283768944, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.147022008895874, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7038913369178772, + "num_tokens": 196872347.0, + "step": 7605 + }, + { + "epoch": 0.835273446079508, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.311650514602661, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6968600153923035, + "num_tokens": 196896899.0, + "step": 7606 + }, + { + "epoch": 0.8353832637821217, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.086707830429077, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7044543027877808, + "num_tokens": 196925898.0, + "step": 7607 + }, + { + "epoch": 0.8354930814847353, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2466306686401367, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.6996533870697021, + "num_tokens": 196949512.0, + "step": 7608 + }, + { + "epoch": 0.835602899187349, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.526374340057373, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7233409285545349, + "num_tokens": 196969131.0, + "step": 7609 + }, + { + "epoch": 0.8357127168899626, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1099090576171875, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6938991546630859, + "num_tokens": 196996637.0, + "step": 7610 + }, + { + "epoch": 0.8358225345925763, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.4305646419525146, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7096922397613525, + "num_tokens": 197017944.0, + "step": 7611 + }, + { + "epoch": 0.83593235229519, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3990399837493896, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7186915278434753, + "num_tokens": 197039417.0, + "step": 7612 + }, + { + "epoch": 0.8360421699978037, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1661453247070312, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7107616066932678, + "num_tokens": 197063264.0, + "step": 7613 + }, + { + "epoch": 0.8361519877004173, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3131508827209473, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7306082248687744, + "num_tokens": 197085752.0, + "step": 7614 + }, + { + "epoch": 0.836261805403031, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.105076313018799, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6923828125, + "num_tokens": 197114138.0, + "step": 7615 + }, + { + "epoch": 0.8363716231056446, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.392098903656006, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.682232141494751, + "num_tokens": 197136693.0, + "step": 7616 + }, + { + "epoch": 0.8364814408082583, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1214282512664795, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7236617803573608, + "num_tokens": 197161869.0, + "step": 7617 + }, + { + "epoch": 0.8365912585108719, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0051746368408203, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7150716781616211, + "num_tokens": 197192424.0, + "step": 7618 + }, + { + "epoch": 0.8367010762134857, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9487618207931519, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7119146585464478, + "num_tokens": 197221800.0, + "step": 7619 + }, + { + "epoch": 0.8368108939160993, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.091581106185913, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7081950902938843, + "num_tokens": 197249437.0, + "step": 7620 + }, + { + "epoch": 0.836920711618713, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.344714641571045, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7180619239807129, + "num_tokens": 197271744.0, + "step": 7621 + }, + { + "epoch": 0.8370305293213266, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2595927715301514, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7112919688224792, + "num_tokens": 197295207.0, + "step": 7622 + }, + { + "epoch": 0.8371403470239402, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.290956497192383, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7093180418014526, + "num_tokens": 197319632.0, + "step": 7623 + }, + { + "epoch": 0.8372501647265539, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1687231063842773, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6850875616073608, + "num_tokens": 197350363.0, + "step": 7624 + }, + { + "epoch": 0.8373599824291675, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2029054164886475, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7241843938827515, + "num_tokens": 197375405.0, + "step": 7625 + }, + { + "epoch": 0.8374698001317813, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1473591327667236, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6946761012077332, + "num_tokens": 197401800.0, + "step": 7626 + }, + { + "epoch": 0.8375796178343949, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2572805881500244, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.692814826965332, + "num_tokens": 197425361.0, + "step": 7627 + }, + { + "epoch": 0.8376894355370086, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.092617988586426, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.7107018828392029, + "num_tokens": 197451855.0, + "step": 7628 + }, + { + "epoch": 0.8377992532396222, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3433263301849365, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.691322922706604, + "num_tokens": 197475433.0, + "step": 7629 + }, + { + "epoch": 0.8379090709422359, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.5862278938293457, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7180458307266235, + "num_tokens": 197495708.0, + "step": 7630 + }, + { + "epoch": 0.8380188886448495, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9136182069778442, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6969006061553955, + "num_tokens": 197529731.0, + "step": 7631 + }, + { + "epoch": 0.8381287063474632, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.272287368774414, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6897282600402832, + "num_tokens": 197555343.0, + "step": 7632 + }, + { + "epoch": 0.8382385240500769, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.021686315536499, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7129641175270081, + "num_tokens": 197585097.0, + "step": 7633 + }, + { + "epoch": 0.8383483417526906, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.076082944869995, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7418484687805176, + "num_tokens": 197610696.0, + "step": 7634 + }, + { + "epoch": 0.8384581594553042, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.042320728302002, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7018827795982361, + "num_tokens": 197638076.0, + "step": 7635 + }, + { + "epoch": 0.8385679771579179, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.4920191764831543, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7084883451461792, + "num_tokens": 197659044.0, + "step": 7636 + }, + { + "epoch": 0.8386777948605315, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.130504608154297, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7071034908294678, + "num_tokens": 197684359.0, + "step": 7637 + }, + { + "epoch": 0.8387876125631452, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.407094717025757, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7110898494720459, + "num_tokens": 197706907.0, + "step": 7638 + }, + { + "epoch": 0.8388974302657588, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1180577278137207, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7084075808525085, + "num_tokens": 197734448.0, + "step": 7639 + }, + { + "epoch": 0.8390072479683724, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1118104457855225, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7061893343925476, + "num_tokens": 197760172.0, + "step": 7640 + }, + { + "epoch": 0.8391170656709862, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2248268127441406, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7057144641876221, + "num_tokens": 197786884.0, + "step": 7641 + }, + { + "epoch": 0.8392268833735999, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.276029109954834, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7120654582977295, + "num_tokens": 197809095.0, + "step": 7642 + }, + { + "epoch": 0.8393367010762135, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1282131671905518, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7128716111183167, + "num_tokens": 197835389.0, + "step": 7643 + }, + { + "epoch": 0.8394465187788271, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.52193546295166, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7328878045082092, + "num_tokens": 197857238.0, + "step": 7644 + }, + { + "epoch": 0.8395563364814408, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.8598096370697021, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.6951713562011719, + "num_tokens": 197891332.0, + "step": 7645 + }, + { + "epoch": 0.8396661541840544, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1901049613952637, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6913293600082397, + "num_tokens": 197917516.0, + "step": 7646 + }, + { + "epoch": 0.8397759718866681, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.359468460083008, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7234175801277161, + "num_tokens": 197937808.0, + "step": 7647 + }, + { + "epoch": 0.8398857895892818, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9912910461425781, + "learning_rate": 1e-06, + "loss": 1.0852, + "mean_token_accuracy": 0.6725452542304993, + "num_tokens": 197970058.0, + "step": 7648 + }, + { + "epoch": 0.8399956072918955, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.043414354324341, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6932018995285034, + "num_tokens": 198000588.0, + "step": 7649 + }, + { + "epoch": 0.8401054249945091, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9514340162277222, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6950041055679321, + "num_tokens": 198030599.0, + "step": 7650 + }, + { + "epoch": 0.8402152426971228, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1736509799957275, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7103360891342163, + "num_tokens": 198055412.0, + "step": 7651 + }, + { + "epoch": 0.8403250603997364, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.179448366165161, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.723021388053894, + "num_tokens": 198079265.0, + "step": 7652 + }, + { + "epoch": 0.8404348781023501, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.5404489040374756, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7141333818435669, + "num_tokens": 198099381.0, + "step": 7653 + }, + { + "epoch": 0.8405446958049637, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.265017032623291, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.7016482353210449, + "num_tokens": 198122592.0, + "step": 7654 + }, + { + "epoch": 0.8406545135075775, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3678572177886963, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7096095085144043, + "num_tokens": 198146233.0, + "step": 7655 + }, + { + "epoch": 0.8407643312101911, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.316953659057617, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7110517024993896, + "num_tokens": 198170128.0, + "step": 7656 + }, + { + "epoch": 0.8408741489128048, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.192807197570801, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7273033261299133, + "num_tokens": 198195310.0, + "step": 7657 + }, + { + "epoch": 0.8409839666154184, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.125840902328491, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6999132037162781, + "num_tokens": 198221303.0, + "step": 7658 + }, + { + "epoch": 0.8410937843180321, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1618716716766357, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6904320120811462, + "num_tokens": 198247562.0, + "step": 7659 + }, + { + "epoch": 0.8412036020206457, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0506560802459717, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6880442500114441, + "num_tokens": 198276906.0, + "step": 7660 + }, + { + "epoch": 0.8413134197232593, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.607611656188965, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7136696577072144, + "num_tokens": 198295104.0, + "step": 7661 + }, + { + "epoch": 0.8414232374258731, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0599348545074463, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7028032541275024, + "num_tokens": 198323331.0, + "step": 7662 + }, + { + "epoch": 0.8415330551284868, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.177288770675659, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7046325206756592, + "num_tokens": 198350668.0, + "step": 7663 + }, + { + "epoch": 0.8416428728311004, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9031085968017578, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7151201963424683, + "num_tokens": 198381292.0, + "step": 7664 + }, + { + "epoch": 0.841752690533714, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2380411624908447, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7117998600006104, + "num_tokens": 198406274.0, + "step": 7665 + }, + { + "epoch": 0.8418625082363277, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2557966709136963, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.6995355486869812, + "num_tokens": 198430130.0, + "step": 7666 + }, + { + "epoch": 0.8419723259389413, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9401546716690063, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7263051271438599, + "num_tokens": 198459777.0, + "step": 7667 + }, + { + "epoch": 0.842082143641555, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2095015048980713, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7001227140426636, + "num_tokens": 198485192.0, + "step": 7668 + }, + { + "epoch": 0.8421919613441686, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.506727457046509, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7300590872764587, + "num_tokens": 198503231.0, + "step": 7669 + }, + { + "epoch": 0.8423017790467824, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2515242099761963, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7202377319335938, + "num_tokens": 198528863.0, + "step": 7670 + }, + { + "epoch": 0.842411596749396, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.167879581451416, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7177740335464478, + "num_tokens": 198553857.0, + "step": 7671 + }, + { + "epoch": 0.8425214144520097, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0566422939300537, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6997901201248169, + "num_tokens": 198583558.0, + "step": 7672 + }, + { + "epoch": 0.8426312321546233, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.1745200157165527, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7278293967247009, + "num_tokens": 198609291.0, + "step": 7673 + }, + { + "epoch": 0.842741049857237, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.098383903503418, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7119437456130981, + "num_tokens": 198636900.0, + "step": 7674 + }, + { + "epoch": 0.8428508675598506, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 1.9215220212936401, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6932851672172546, + "num_tokens": 198668433.0, + "step": 7675 + }, + { + "epoch": 0.8429606852624643, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.19319486618042, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7151228189468384, + "num_tokens": 198694562.0, + "step": 7676 + }, + { + "epoch": 0.843070502965078, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.3975167274475098, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7068238258361816, + "num_tokens": 198717369.0, + "step": 7677 + }, + { + "epoch": 0.8431803206676917, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.0648300647735596, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7124069333076477, + "num_tokens": 198744856.0, + "step": 7678 + }, + { + "epoch": 0.8432901383703053, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.163980007171631, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.6906154751777649, + "num_tokens": 198771538.0, + "step": 7679 + }, + { + "epoch": 0.843399956072919, + "ewc_loss": 6.580352783203125e-05, + "grad_norm": 2.2358217239379883, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6934481263160706, + "num_tokens": 198796885.0, + "step": 7680 + }, + { + "epoch": 0.8435097737755326, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.296815872192383, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7159112691879272, + "num_tokens": 198818940.0, + "step": 7681 + }, + { + "epoch": 0.8436195914781462, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.1875176429748535, + "learning_rate": 1e-06, + "loss": 1.0514, + "mean_token_accuracy": 0.6771480441093445, + "num_tokens": 198845612.0, + "step": 7682 + }, + { + "epoch": 0.8437294091807599, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.04142689704895, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7104401588439941, + "num_tokens": 198874429.0, + "step": 7683 + }, + { + "epoch": 0.8438392268833736, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.33651065826416, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7122383713722229, + "num_tokens": 198899257.0, + "step": 7684 + }, + { + "epoch": 0.8439490445859873, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.248781442642212, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7139322757720947, + "num_tokens": 198923303.0, + "step": 7685 + }, + { + "epoch": 0.8440588622886009, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.185880661010742, + "learning_rate": 1e-06, + "loss": 1.1146, + "mean_token_accuracy": 0.6626691818237305, + "num_tokens": 198953795.0, + "step": 7686 + }, + { + "epoch": 0.8441686799912146, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.142080545425415, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7032659649848938, + "num_tokens": 198981725.0, + "step": 7687 + }, + { + "epoch": 0.8442784976938282, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.224571943283081, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7127354145050049, + "num_tokens": 199005153.0, + "step": 7688 + }, + { + "epoch": 0.8443883153964419, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.4080827236175537, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7269949913024902, + "num_tokens": 199027042.0, + "step": 7689 + }, + { + "epoch": 0.8444981330990555, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.3937695026397705, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7130590081214905, + "num_tokens": 199050146.0, + "step": 7690 + }, + { + "epoch": 0.8446079508016693, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.359376907348633, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6991183161735535, + "num_tokens": 199073404.0, + "step": 7691 + }, + { + "epoch": 0.8447177685042829, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.1410295963287354, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6996957063674927, + "num_tokens": 199101008.0, + "step": 7692 + }, + { + "epoch": 0.8448275862068966, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.2459607124328613, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6948485970497131, + "num_tokens": 199126165.0, + "step": 7693 + }, + { + "epoch": 0.8449374039095102, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.0570878982543945, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7064090967178345, + "num_tokens": 199153533.0, + "step": 7694 + }, + { + "epoch": 0.8450472216121239, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.065685749053955, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7389941215515137, + "num_tokens": 199179424.0, + "step": 7695 + }, + { + "epoch": 0.8451570393147375, + "ewc_loss": 6.628036499023438e-05, + "grad_norm": 2.4075000286102295, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7003738880157471, + "num_tokens": 199202764.0, + "step": 7696 + }, + { + "epoch": 0.8452668570173512, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2665693759918213, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7178075909614563, + "num_tokens": 199227940.0, + "step": 7697 + }, + { + "epoch": 0.8453766747199648, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.186507225036621, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7119029760360718, + "num_tokens": 199252255.0, + "step": 7698 + }, + { + "epoch": 0.8454864924225786, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0521469116210938, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7143090963363647, + "num_tokens": 199278817.0, + "step": 7699 + }, + { + "epoch": 0.8455963101251922, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.176175355911255, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7099449634552002, + "num_tokens": 199304937.0, + "step": 7700 + }, + { + "epoch": 0.8457061278278059, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.8252712488174438, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7306057214736938, + "num_tokens": 199337317.0, + "step": 7701 + }, + { + "epoch": 0.8458159455304195, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1863901615142822, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7081761360168457, + "num_tokens": 199363824.0, + "step": 7702 + }, + { + "epoch": 0.8459257632330331, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2188050746917725, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7073506116867065, + "num_tokens": 199388320.0, + "step": 7703 + }, + { + "epoch": 0.8460355809356468, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.155182123184204, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.6979653239250183, + "num_tokens": 199415523.0, + "step": 7704 + }, + { + "epoch": 0.8461453986382604, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.277374744415283, + "learning_rate": 1e-06, + "loss": 1.0848, + "mean_token_accuracy": 0.6905659437179565, + "num_tokens": 199441645.0, + "step": 7705 + }, + { + "epoch": 0.8462552163408742, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.5146539211273193, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7192025184631348, + "num_tokens": 199462863.0, + "step": 7706 + }, + { + "epoch": 0.8463650340434878, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3950469493865967, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7305992841720581, + "num_tokens": 199484387.0, + "step": 7707 + }, + { + "epoch": 0.8464748517461015, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1173415184020996, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.704546332359314, + "num_tokens": 199513679.0, + "step": 7708 + }, + { + "epoch": 0.8465846694487151, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1857402324676514, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6882753372192383, + "num_tokens": 199542808.0, + "step": 7709 + }, + { + "epoch": 0.8466944871513288, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3430404663085938, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6929981708526611, + "num_tokens": 199566968.0, + "step": 7710 + }, + { + "epoch": 0.8468043048539424, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1870265007019043, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7183736562728882, + "num_tokens": 199593692.0, + "step": 7711 + }, + { + "epoch": 0.8469141225565561, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0063607692718506, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7051538825035095, + "num_tokens": 199622402.0, + "step": 7712 + }, + { + "epoch": 0.8470239402591698, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.7177939414978027, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7195601463317871, + "num_tokens": 199641522.0, + "step": 7713 + }, + { + "epoch": 0.8471337579617835, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.123166561126709, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7097662687301636, + "num_tokens": 199667927.0, + "step": 7714 + }, + { + "epoch": 0.8472435756643971, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3117177486419678, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.6902507543563843, + "num_tokens": 199693552.0, + "step": 7715 + }, + { + "epoch": 0.8473533933670108, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.268784523010254, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.685369610786438, + "num_tokens": 199716886.0, + "step": 7716 + }, + { + "epoch": 0.8474632110696244, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.4150314331054688, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7113760709762573, + "num_tokens": 199737928.0, + "step": 7717 + }, + { + "epoch": 0.8475730287722381, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0321927070617676, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7106965780258179, + "num_tokens": 199766914.0, + "step": 7718 + }, + { + "epoch": 0.8476828464748517, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.938101053237915, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7089153528213501, + "num_tokens": 199797223.0, + "step": 7719 + }, + { + "epoch": 0.8477926641774655, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2690765857696533, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.737415075302124, + "num_tokens": 199819060.0, + "step": 7720 + }, + { + "epoch": 0.8479024818800791, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2460076808929443, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7131322026252747, + "num_tokens": 199845241.0, + "step": 7721 + }, + { + "epoch": 0.8480122995826928, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2050724029541016, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7244999408721924, + "num_tokens": 199868686.0, + "step": 7722 + }, + { + "epoch": 0.8481221172853064, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.536355495452881, + "learning_rate": 1e-06, + "loss": 1.0705, + "mean_token_accuracy": 0.6816515922546387, + "num_tokens": 199892312.0, + "step": 7723 + }, + { + "epoch": 0.84823193498792, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.280094623565674, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7099865674972534, + "num_tokens": 199917331.0, + "step": 7724 + }, + { + "epoch": 0.8483417526905337, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2061407566070557, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7093896865844727, + "num_tokens": 199942069.0, + "step": 7725 + }, + { + "epoch": 0.8484515703931473, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1778573989868164, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.6939765810966492, + "num_tokens": 199967921.0, + "step": 7726 + }, + { + "epoch": 0.848561388095761, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1389214992523193, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7200847268104553, + "num_tokens": 199993258.0, + "step": 7727 + }, + { + "epoch": 0.8486712057983747, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9544214010238647, + "learning_rate": 1e-06, + "loss": 1.0642, + "mean_token_accuracy": 0.6756030917167664, + "num_tokens": 200025395.0, + "step": 7728 + }, + { + "epoch": 0.8487810235009884, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.5881800651550293, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6937330365180969, + "num_tokens": 200047088.0, + "step": 7729 + }, + { + "epoch": 0.848890841203602, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1162259578704834, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7176483273506165, + "num_tokens": 200073044.0, + "step": 7730 + }, + { + "epoch": 0.8490006589062157, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.994038701057434, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7036827802658081, + "num_tokens": 200103693.0, + "step": 7731 + }, + { + "epoch": 0.8491104766088293, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.097834825515747, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7187845706939697, + "num_tokens": 200129605.0, + "step": 7732 + }, + { + "epoch": 0.849220294311443, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0675036907196045, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.7018359899520874, + "num_tokens": 200159419.0, + "step": 7733 + }, + { + "epoch": 0.8493301120140566, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.272315740585327, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6858346462249756, + "num_tokens": 200183858.0, + "step": 7734 + }, + { + "epoch": 0.8494399297166704, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9995681047439575, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6861780881881714, + "num_tokens": 200215080.0, + "step": 7735 + }, + { + "epoch": 0.849549747419284, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.410508632659912, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.6913512945175171, + "num_tokens": 200236944.0, + "step": 7736 + }, + { + "epoch": 0.8496595651218977, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1212410926818848, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7041791677474976, + "num_tokens": 200265185.0, + "step": 7737 + }, + { + "epoch": 0.8497693828245113, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0077481269836426, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7056605219841003, + "num_tokens": 200293716.0, + "step": 7738 + }, + { + "epoch": 0.849879200527125, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1886868476867676, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7026153802871704, + "num_tokens": 200318403.0, + "step": 7739 + }, + { + "epoch": 0.8499890182297386, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0782313346862793, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7337764501571655, + "num_tokens": 200343584.0, + "step": 7740 + }, + { + "epoch": 0.8500988359323522, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9462578296661377, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7075515985488892, + "num_tokens": 200375251.0, + "step": 7741 + }, + { + "epoch": 0.850208653634966, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0406875610351562, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6929197311401367, + "num_tokens": 200406054.0, + "step": 7742 + }, + { + "epoch": 0.8503184713375797, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.4241104125976562, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.7088107466697693, + "num_tokens": 200427875.0, + "step": 7743 + }, + { + "epoch": 0.8504282890401933, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0268447399139404, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7271510362625122, + "num_tokens": 200454987.0, + "step": 7744 + }, + { + "epoch": 0.8505381067428069, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.201223611831665, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7304759621620178, + "num_tokens": 200478848.0, + "step": 7745 + }, + { + "epoch": 0.8506479244454206, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2064099311828613, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7171701788902283, + "num_tokens": 200502938.0, + "step": 7746 + }, + { + "epoch": 0.8507577421480342, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.195402145385742, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7306017875671387, + "num_tokens": 200527962.0, + "step": 7747 + }, + { + "epoch": 0.8508675598506479, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1924407482147217, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.7055470943450928, + "num_tokens": 200554384.0, + "step": 7748 + }, + { + "epoch": 0.8509773775532616, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.172179937362671, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.6968674063682556, + "num_tokens": 200580256.0, + "step": 7749 + }, + { + "epoch": 0.8510871952558753, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.4189116954803467, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.703411340713501, + "num_tokens": 200602913.0, + "step": 7750 + }, + { + "epoch": 0.8511970129584889, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.001830577850342, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.697262167930603, + "num_tokens": 200637997.0, + "step": 7751 + }, + { + "epoch": 0.8513068306611026, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.247131109237671, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7193544507026672, + "num_tokens": 200663680.0, + "step": 7752 + }, + { + "epoch": 0.8514166483637162, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.128610372543335, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.723525881767273, + "num_tokens": 200687698.0, + "step": 7753 + }, + { + "epoch": 0.8515264660663299, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0509119033813477, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6941328048706055, + "num_tokens": 200718240.0, + "step": 7754 + }, + { + "epoch": 0.8516362837689435, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2422428131103516, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6867868304252625, + "num_tokens": 200743003.0, + "step": 7755 + }, + { + "epoch": 0.8517461014715572, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.121781587600708, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6848622560501099, + "num_tokens": 200772187.0, + "step": 7756 + }, + { + "epoch": 0.8518559191741709, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.228477954864502, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7243853807449341, + "num_tokens": 200800494.0, + "step": 7757 + }, + { + "epoch": 0.8519657368767846, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.420698642730713, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7217370867729187, + "num_tokens": 200820127.0, + "step": 7758 + }, + { + "epoch": 0.8520755545793982, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.569694757461548, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7361807823181152, + "num_tokens": 200838496.0, + "step": 7759 + }, + { + "epoch": 0.8521853722820119, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2624001502990723, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.7061175107955933, + "num_tokens": 200863720.0, + "step": 7760 + }, + { + "epoch": 0.8522951899846255, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2294392585754395, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7087247371673584, + "num_tokens": 200890496.0, + "step": 7761 + }, + { + "epoch": 0.8524050076872391, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.047997236251831, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7125869989395142, + "num_tokens": 200920864.0, + "step": 7762 + }, + { + "epoch": 0.8525148253898528, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.291851043701172, + "learning_rate": 1e-06, + "loss": 1.0501, + "mean_token_accuracy": 0.6823366284370422, + "num_tokens": 200945141.0, + "step": 7763 + }, + { + "epoch": 0.8526246430924665, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.146827459335327, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7084940671920776, + "num_tokens": 200973178.0, + "step": 7764 + }, + { + "epoch": 0.8527344607950802, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1851353645324707, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7198830842971802, + "num_tokens": 200996047.0, + "step": 7765 + }, + { + "epoch": 0.8528442784976938, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2218918800354004, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7020227313041687, + "num_tokens": 201021146.0, + "step": 7766 + }, + { + "epoch": 0.8529540962003075, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2177884578704834, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7221126556396484, + "num_tokens": 201046614.0, + "step": 7767 + }, + { + "epoch": 0.8530639139029211, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2472479343414307, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7354094982147217, + "num_tokens": 201068484.0, + "step": 7768 + }, + { + "epoch": 0.8531737316055348, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.35528564453125, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.71474289894104, + "num_tokens": 201089170.0, + "step": 7769 + }, + { + "epoch": 0.8532835493081484, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.343496561050415, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6943082809448242, + "num_tokens": 201112542.0, + "step": 7770 + }, + { + "epoch": 0.8533933670107622, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.204512119293213, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7356255054473877, + "num_tokens": 201136180.0, + "step": 7771 + }, + { + "epoch": 0.8535031847133758, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.385841131210327, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7041304111480713, + "num_tokens": 201159136.0, + "step": 7772 + }, + { + "epoch": 0.8536130024159895, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.043290853500366, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6954866647720337, + "num_tokens": 201188466.0, + "step": 7773 + }, + { + "epoch": 0.8537228201186031, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.11942982673645, + "learning_rate": 1e-06, + "loss": 1.0232, + "mean_token_accuracy": 0.6901187896728516, + "num_tokens": 201214870.0, + "step": 7774 + }, + { + "epoch": 0.8538326378212168, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3150599002838135, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7159431576728821, + "num_tokens": 201238516.0, + "step": 7775 + }, + { + "epoch": 0.8539424555238304, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.159306764602661, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7087677121162415, + "num_tokens": 201264492.0, + "step": 7776 + }, + { + "epoch": 0.8540522732264441, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.799515962600708, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6885087490081787, + "num_tokens": 201300458.0, + "step": 7777 + }, + { + "epoch": 0.8541620909290578, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.732090711593628, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7131569385528564, + "num_tokens": 201337032.0, + "step": 7778 + }, + { + "epoch": 0.8542719086316715, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3431453704833984, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7109193801879883, + "num_tokens": 201360124.0, + "step": 7779 + }, + { + "epoch": 0.8543817263342851, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.573815107345581, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7076473236083984, + "num_tokens": 201379881.0, + "step": 7780 + }, + { + "epoch": 0.8544915440368988, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1277365684509277, + "learning_rate": 1e-06, + "loss": 1.0837, + "mean_token_accuracy": 0.6703161001205444, + "num_tokens": 201407020.0, + "step": 7781 + }, + { + "epoch": 0.8546013617395124, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.4291422367095947, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7176618576049805, + "num_tokens": 201430308.0, + "step": 7782 + }, + { + "epoch": 0.854711179442126, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2280120849609375, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6925472021102905, + "num_tokens": 201456423.0, + "step": 7783 + }, + { + "epoch": 0.8548209971447397, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1727652549743652, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7181318998336792, + "num_tokens": 201479965.0, + "step": 7784 + }, + { + "epoch": 0.8549308148473534, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.150596857070923, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.686296284198761, + "num_tokens": 201506398.0, + "step": 7785 + }, + { + "epoch": 0.8550406325499671, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0874009132385254, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7066947817802429, + "num_tokens": 201533902.0, + "step": 7786 + }, + { + "epoch": 0.8551504502525807, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3224775791168213, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.721470832824707, + "num_tokens": 201555508.0, + "step": 7787 + }, + { + "epoch": 0.8552602679551944, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2529852390289307, + "learning_rate": 1e-06, + "loss": 1.0792, + "mean_token_accuracy": 0.6770973205566406, + "num_tokens": 201583580.0, + "step": 7788 + }, + { + "epoch": 0.855370085657808, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.090506076812744, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6989893317222595, + "num_tokens": 201614497.0, + "step": 7789 + }, + { + "epoch": 0.8554799033604217, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2243895530700684, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7152106761932373, + "num_tokens": 201637724.0, + "step": 7790 + }, + { + "epoch": 0.8555897210630353, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0849456787109375, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7210820913314819, + "num_tokens": 201663407.0, + "step": 7791 + }, + { + "epoch": 0.855699538765649, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.509697437286377, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7030110359191895, + "num_tokens": 201683632.0, + "step": 7792 + }, + { + "epoch": 0.8558093564682627, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3974874019622803, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.720012903213501, + "num_tokens": 201704739.0, + "step": 7793 + }, + { + "epoch": 0.8559191741708764, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.311352252960205, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7092475295066833, + "num_tokens": 201727432.0, + "step": 7794 + }, + { + "epoch": 0.85602899187349, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.3678672313690186, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7065643072128296, + "num_tokens": 201751819.0, + "step": 7795 + }, + { + "epoch": 0.8561388095761037, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.8404492139816284, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7133774161338806, + "num_tokens": 201783478.0, + "step": 7796 + }, + { + "epoch": 0.8562486272787173, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1323719024658203, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7106194496154785, + "num_tokens": 201808626.0, + "step": 7797 + }, + { + "epoch": 0.856358444981331, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.341588020324707, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7147467136383057, + "num_tokens": 201830947.0, + "step": 7798 + }, + { + "epoch": 0.8564682626839446, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.331815242767334, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.694394052028656, + "num_tokens": 201855299.0, + "step": 7799 + }, + { + "epoch": 0.8565780803865584, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.224557876586914, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.714033305644989, + "num_tokens": 201881901.0, + "step": 7800 + }, + { + "epoch": 0.856687898089172, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.976410150527954, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7019033432006836, + "num_tokens": 201914176.0, + "step": 7801 + }, + { + "epoch": 0.8567977157917857, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.024855852127075, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7017380595207214, + "num_tokens": 201941746.0, + "step": 7802 + }, + { + "epoch": 0.8569075334943993, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0768558979034424, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7234978079795837, + "num_tokens": 201968023.0, + "step": 7803 + }, + { + "epoch": 0.8570173511970129, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1262011528015137, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7119941115379333, + "num_tokens": 201992326.0, + "step": 7804 + }, + { + "epoch": 0.8571271688996266, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.7710182666778564, + "learning_rate": 1e-06, + "loss": 1.0831, + "mean_token_accuracy": 0.6708412170410156, + "num_tokens": 202030782.0, + "step": 7805 + }, + { + "epoch": 0.8572369866022402, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.054945707321167, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7061411142349243, + "num_tokens": 202058757.0, + "step": 7806 + }, + { + "epoch": 0.857346804304854, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2871203422546387, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7038753032684326, + "num_tokens": 202083103.0, + "step": 7807 + }, + { + "epoch": 0.8574566220074676, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.10249662399292, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.695827305316925, + "num_tokens": 202112823.0, + "step": 7808 + }, + { + "epoch": 0.8575664397100813, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 1.9042373895645142, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7261928915977478, + "num_tokens": 202143862.0, + "step": 7809 + }, + { + "epoch": 0.8576762574126949, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9091544151306152, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.6855025291442871, + "num_tokens": 202176427.0, + "step": 7810 + }, + { + "epoch": 0.8577860751153086, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.035074234008789, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7235330939292908, + "num_tokens": 202205462.0, + "step": 7811 + }, + { + "epoch": 0.8578958928179222, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.328005790710449, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7123071551322937, + "num_tokens": 202228124.0, + "step": 7812 + }, + { + "epoch": 0.8580057105205359, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.211078405380249, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7115950584411621, + "num_tokens": 202253879.0, + "step": 7813 + }, + { + "epoch": 0.8581155282231496, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9725035429000854, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.7008323669433594, + "num_tokens": 202283863.0, + "step": 7814 + }, + { + "epoch": 0.8582253459257633, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2120306491851807, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7112119197845459, + "num_tokens": 202308816.0, + "step": 7815 + }, + { + "epoch": 0.8583351636283769, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.6049280166625977, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7235637307167053, + "num_tokens": 202328376.0, + "step": 7816 + }, + { + "epoch": 0.8584449813309906, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.115326404571533, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7121149301528931, + "num_tokens": 202355544.0, + "step": 7817 + }, + { + "epoch": 0.8585547990336042, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0573410987854004, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7289087772369385, + "num_tokens": 202381441.0, + "step": 7818 + }, + { + "epoch": 0.8586646167362179, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1851418018341064, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7305768728256226, + "num_tokens": 202408014.0, + "step": 7819 + }, + { + "epoch": 0.8587744344388315, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.076169729232788, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7457267045974731, + "num_tokens": 202435919.0, + "step": 7820 + }, + { + "epoch": 0.8588842521414451, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.234388589859009, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7208877801895142, + "num_tokens": 202458627.0, + "step": 7821 + }, + { + "epoch": 0.8589940698440589, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.186669111251831, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7096871733665466, + "num_tokens": 202484299.0, + "step": 7822 + }, + { + "epoch": 0.8591038875466726, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.125998020172119, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.711962878704071, + "num_tokens": 202510414.0, + "step": 7823 + }, + { + "epoch": 0.8592137052492862, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2771787643432617, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7152291536331177, + "num_tokens": 202533513.0, + "step": 7824 + }, + { + "epoch": 0.8593235229518998, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.246957778930664, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7006444931030273, + "num_tokens": 202559780.0, + "step": 7825 + }, + { + "epoch": 0.8594333406545135, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.9561525583267212, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7029500007629395, + "num_tokens": 202589175.0, + "step": 7826 + }, + { + "epoch": 0.8595431583571271, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.4348256587982178, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7173624038696289, + "num_tokens": 202611568.0, + "step": 7827 + }, + { + "epoch": 0.8596529760597408, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0058979988098145, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6791519522666931, + "num_tokens": 202640644.0, + "step": 7828 + }, + { + "epoch": 0.8597627937623545, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.8589704036712646, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7024804353713989, + "num_tokens": 202674639.0, + "step": 7829 + }, + { + "epoch": 0.8598726114649682, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1942763328552246, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.6935930252075195, + "num_tokens": 202700287.0, + "step": 7830 + }, + { + "epoch": 0.8599824291675818, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 1.825603723526001, + "learning_rate": 1e-06, + "loss": 1.0482, + "mean_token_accuracy": 0.6823043823242188, + "num_tokens": 202736629.0, + "step": 7831 + }, + { + "epoch": 0.8600922468701955, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.301086187362671, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7179175615310669, + "num_tokens": 202757926.0, + "step": 7832 + }, + { + "epoch": 0.8602020645728091, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1874890327453613, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6994705200195312, + "num_tokens": 202784249.0, + "step": 7833 + }, + { + "epoch": 0.8603118822754228, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.2163140773773193, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6975302696228027, + "num_tokens": 202809701.0, + "step": 7834 + }, + { + "epoch": 0.8604216999780364, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0510518550872803, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7014304399490356, + "num_tokens": 202837782.0, + "step": 7835 + }, + { + "epoch": 0.8605315176806502, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.1421477794647217, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7263371348381042, + "num_tokens": 202864615.0, + "step": 7836 + }, + { + "epoch": 0.8606413353832638, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.094557046890259, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7078863382339478, + "num_tokens": 202891672.0, + "step": 7837 + }, + { + "epoch": 0.8607511530858775, + "ewc_loss": 6.67572021484375e-05, + "grad_norm": 2.0546789169311523, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.6988593339920044, + "num_tokens": 202922336.0, + "step": 7838 + }, + { + "epoch": 0.8608609707884911, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.090644598007202, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.6968891620635986, + "num_tokens": 202949143.0, + "step": 7839 + }, + { + "epoch": 0.8609707884911048, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.418144702911377, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7419352531433105, + "num_tokens": 202967791.0, + "step": 7840 + }, + { + "epoch": 0.8610806061937184, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2541005611419678, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6938896179199219, + "num_tokens": 202990516.0, + "step": 7841 + }, + { + "epoch": 0.861190423896332, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2026429176330566, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.701858639717102, + "num_tokens": 203014639.0, + "step": 7842 + }, + { + "epoch": 0.8613002415989458, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0656397342681885, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7130591869354248, + "num_tokens": 203042275.0, + "step": 7843 + }, + { + "epoch": 0.8614100593015594, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.280797004699707, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6967008113861084, + "num_tokens": 203066848.0, + "step": 7844 + }, + { + "epoch": 0.8615198770041731, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 1.9988315105438232, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6863529086112976, + "num_tokens": 203097751.0, + "step": 7845 + }, + { + "epoch": 0.8616296947067867, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.158050537109375, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.692534863948822, + "num_tokens": 203124755.0, + "step": 7846 + }, + { + "epoch": 0.8617395124094004, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.414276361465454, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.7052392959594727, + "num_tokens": 203146497.0, + "step": 7847 + }, + { + "epoch": 0.861849330112014, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.4009931087493896, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7141318321228027, + "num_tokens": 203167468.0, + "step": 7848 + }, + { + "epoch": 0.8619591478146277, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 1.9429662227630615, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7194852828979492, + "num_tokens": 203198814.0, + "step": 7849 + }, + { + "epoch": 0.8620689655172413, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1401429176330566, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7064852714538574, + "num_tokens": 203224152.0, + "step": 7850 + }, + { + "epoch": 0.8621787832198551, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1517014503479004, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7323921322822571, + "num_tokens": 203248679.0, + "step": 7851 + }, + { + "epoch": 0.8622886009224687, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 1.9616844654083252, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6851903200149536, + "num_tokens": 203279320.0, + "step": 7852 + }, + { + "epoch": 0.8623984186250824, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.207979440689087, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7164428234100342, + "num_tokens": 203303420.0, + "step": 7853 + }, + { + "epoch": 0.862508236327696, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.3014519214630127, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7149759531021118, + "num_tokens": 203327137.0, + "step": 7854 + }, + { + "epoch": 0.8626180540303097, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1464273929595947, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7165968418121338, + "num_tokens": 203353839.0, + "step": 7855 + }, + { + "epoch": 0.8627278717329233, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0244054794311523, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.6906100511550903, + "num_tokens": 203383795.0, + "step": 7856 + }, + { + "epoch": 0.862837689435537, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.3223016262054443, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7152645587921143, + "num_tokens": 203407162.0, + "step": 7857 + }, + { + "epoch": 0.8629475071381507, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.176459789276123, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7014976739883423, + "num_tokens": 203434099.0, + "step": 7858 + }, + { + "epoch": 0.8630573248407644, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.3623275756835938, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.717420220375061, + "num_tokens": 203456548.0, + "step": 7859 + }, + { + "epoch": 0.863167142543378, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.109079122543335, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.718885600566864, + "num_tokens": 203484169.0, + "step": 7860 + }, + { + "epoch": 0.8632769602459917, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2425668239593506, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6956422328948975, + "num_tokens": 203509125.0, + "step": 7861 + }, + { + "epoch": 0.8633867779486053, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.106110095977783, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7160648107528687, + "num_tokens": 203534755.0, + "step": 7862 + }, + { + "epoch": 0.863496595651219, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.258594512939453, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.6971813440322876, + "num_tokens": 203558568.0, + "step": 7863 + }, + { + "epoch": 0.8636064133538326, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2501227855682373, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7083094120025635, + "num_tokens": 203582837.0, + "step": 7864 + }, + { + "epoch": 0.8637162310564463, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0711660385131836, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7175043821334839, + "num_tokens": 203611467.0, + "step": 7865 + }, + { + "epoch": 0.86382604875906, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2548470497131348, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.71197110414505, + "num_tokens": 203636023.0, + "step": 7866 + }, + { + "epoch": 0.8639358664616736, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2143259048461914, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7023417949676514, + "num_tokens": 203660568.0, + "step": 7867 + }, + { + "epoch": 0.8640456841642873, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2319562435150146, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.693225622177124, + "num_tokens": 203688505.0, + "step": 7868 + }, + { + "epoch": 0.8641555018669009, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0396320819854736, + "learning_rate": 1e-06, + "loss": 1.0673, + "mean_token_accuracy": 0.6827660799026489, + "num_tokens": 203718239.0, + "step": 7869 + }, + { + "epoch": 0.8642653195695146, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.162019968032837, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7158055901527405, + "num_tokens": 203743844.0, + "step": 7870 + }, + { + "epoch": 0.8643751372721282, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2864012718200684, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7093397974967957, + "num_tokens": 203767153.0, + "step": 7871 + }, + { + "epoch": 0.864484954974742, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0113048553466797, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7069416046142578, + "num_tokens": 203794917.0, + "step": 7872 + }, + { + "epoch": 0.8645947726773556, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.360640048980713, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7062548398971558, + "num_tokens": 203817801.0, + "step": 7873 + }, + { + "epoch": 0.8647045903799693, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.222590684890747, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7009161710739136, + "num_tokens": 203841656.0, + "step": 7874 + }, + { + "epoch": 0.8648144080825829, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.392380475997925, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7221764326095581, + "num_tokens": 203864128.0, + "step": 7875 + }, + { + "epoch": 0.8649242257851966, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.2946088314056396, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7255445122718811, + "num_tokens": 203886915.0, + "step": 7876 + }, + { + "epoch": 0.8650340434878102, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.1083993911743164, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7166447639465332, + "num_tokens": 203912550.0, + "step": 7877 + }, + { + "epoch": 0.8651438611904239, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.3408524990081787, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7116589546203613, + "num_tokens": 203935089.0, + "step": 7878 + }, + { + "epoch": 0.8652536788930375, + "ewc_loss": 6.723403930664062e-05, + "grad_norm": 2.0197596549987793, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7377647161483765, + "num_tokens": 203963748.0, + "step": 7879 + }, + { + "epoch": 0.8653634965956513, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.4464542865753174, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7246313691139221, + "num_tokens": 203984878.0, + "step": 7880 + }, + { + "epoch": 0.8654733142982649, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.0276222229003906, + "learning_rate": 1e-06, + "loss": 1.0801, + "mean_token_accuracy": 0.6824580430984497, + "num_tokens": 204014695.0, + "step": 7881 + }, + { + "epoch": 0.8655831320008786, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.378021240234375, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7217491865158081, + "num_tokens": 204035281.0, + "step": 7882 + }, + { + "epoch": 0.8656929497034922, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.216982841491699, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7174203991889954, + "num_tokens": 204059679.0, + "step": 7883 + }, + { + "epoch": 0.8658027674061058, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.3815155029296875, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.71673583984375, + "num_tokens": 204081454.0, + "step": 7884 + }, + { + "epoch": 0.8659125851087195, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.3207037448883057, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7254914045333862, + "num_tokens": 204103325.0, + "step": 7885 + }, + { + "epoch": 0.8660224028113331, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.641308307647705, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7068262696266174, + "num_tokens": 204124511.0, + "step": 7886 + }, + { + "epoch": 0.8661322205139469, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.106381893157959, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7077295184135437, + "num_tokens": 204151632.0, + "step": 7887 + }, + { + "epoch": 0.8662420382165605, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.1314971446990967, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6871521472930908, + "num_tokens": 204177986.0, + "step": 7888 + }, + { + "epoch": 0.8663518559191742, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.164027452468872, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7004520893096924, + "num_tokens": 204205111.0, + "step": 7889 + }, + { + "epoch": 0.8664616736217878, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.2771217823028564, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7048704028129578, + "num_tokens": 204227601.0, + "step": 7890 + }, + { + "epoch": 0.8665714913244015, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.2515509128570557, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7234680652618408, + "num_tokens": 204250245.0, + "step": 7891 + }, + { + "epoch": 0.8666813090270151, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.333548069000244, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7093539237976074, + "num_tokens": 204274328.0, + "step": 7892 + }, + { + "epoch": 0.8667911267296288, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.3674097061157227, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7220755815505981, + "num_tokens": 204297123.0, + "step": 7893 + }, + { + "epoch": 0.8669009444322425, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 1.9430644512176514, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7096316814422607, + "num_tokens": 204327408.0, + "step": 7894 + }, + { + "epoch": 0.8670107621348562, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.4375948905944824, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7190250158309937, + "num_tokens": 204349246.0, + "step": 7895 + }, + { + "epoch": 0.8671205798374698, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.5658881664276123, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.728998064994812, + "num_tokens": 204367333.0, + "step": 7896 + }, + { + "epoch": 0.8672303975400835, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 1.861409068107605, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7276749610900879, + "num_tokens": 204400348.0, + "step": 7897 + }, + { + "epoch": 0.8673402152426971, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.244389057159424, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6893858909606934, + "num_tokens": 204425388.0, + "step": 7898 + }, + { + "epoch": 0.8674500329453108, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.4224727153778076, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.6935107707977295, + "num_tokens": 204449858.0, + "step": 7899 + }, + { + "epoch": 0.8675598506479244, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.164246082305908, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6862990856170654, + "num_tokens": 204476824.0, + "step": 7900 + }, + { + "epoch": 0.8676696683505382, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 1.9736634492874146, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6861002445220947, + "num_tokens": 204505730.0, + "step": 7901 + }, + { + "epoch": 0.8677794860531518, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.6185991764068604, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7060492634773254, + "num_tokens": 204526003.0, + "step": 7902 + }, + { + "epoch": 0.8678893037557655, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.4376962184906006, + "learning_rate": 1e-06, + "loss": 1.0419, + "mean_token_accuracy": 0.6981608271598816, + "num_tokens": 204548218.0, + "step": 7903 + }, + { + "epoch": 0.8679991214583791, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.082017183303833, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6956117749214172, + "num_tokens": 204576206.0, + "step": 7904 + }, + { + "epoch": 0.8681089391609927, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 1.9573029279708862, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6993708610534668, + "num_tokens": 204608378.0, + "step": 7905 + }, + { + "epoch": 0.8682187568636064, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.525822401046753, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6891599297523499, + "num_tokens": 204628158.0, + "step": 7906 + }, + { + "epoch": 0.86832857456622, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.393735885620117, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7164772748947144, + "num_tokens": 204650319.0, + "step": 7907 + }, + { + "epoch": 0.8684383922688337, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.6454977989196777, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7204383611679077, + "num_tokens": 204668863.0, + "step": 7908 + }, + { + "epoch": 0.8685482099714474, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2681429386138916, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7204437255859375, + "num_tokens": 204692670.0, + "step": 7909 + }, + { + "epoch": 0.8686580276740611, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.378720998764038, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6889233589172363, + "num_tokens": 204715673.0, + "step": 7910 + }, + { + "epoch": 0.8687678453766747, + "ewc_loss": 6.771087646484375e-05, + "grad_norm": 2.3658406734466553, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7338061332702637, + "num_tokens": 204738523.0, + "step": 7911 + }, + { + "epoch": 0.8688776630792884, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0560123920440674, + "learning_rate": 1e-06, + "loss": 1.0692, + "mean_token_accuracy": 0.6741931438446045, + "num_tokens": 204767872.0, + "step": 7912 + }, + { + "epoch": 0.868987480781902, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1537864208221436, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6967554688453674, + "num_tokens": 204795170.0, + "step": 7913 + }, + { + "epoch": 0.8690972984845157, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.9830402135849, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6851270794868469, + "num_tokens": 204825372.0, + "step": 7914 + }, + { + "epoch": 0.8692071161871293, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0660154819488525, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6992190480232239, + "num_tokens": 204853010.0, + "step": 7915 + }, + { + "epoch": 0.8693169338897431, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0354700088500977, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7053549885749817, + "num_tokens": 204883722.0, + "step": 7916 + }, + { + "epoch": 0.8694267515923567, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2097244262695312, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7099158763885498, + "num_tokens": 204908704.0, + "step": 7917 + }, + { + "epoch": 0.8695365692949704, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.280243396759033, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7144625782966614, + "num_tokens": 204931118.0, + "step": 7918 + }, + { + "epoch": 0.869646386997584, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0644619464874268, + "learning_rate": 1e-06, + "loss": 1.0447, + "mean_token_accuracy": 0.6786820292472839, + "num_tokens": 204959245.0, + "step": 7919 + }, + { + "epoch": 0.8697562047001977, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.4892921447753906, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7046639919281006, + "num_tokens": 204980960.0, + "step": 7920 + }, + { + "epoch": 0.8698660224028113, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.9660825729370117, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7137286067008972, + "num_tokens": 205012385.0, + "step": 7921 + }, + { + "epoch": 0.869975840105425, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.997209072113037, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7087478637695312, + "num_tokens": 205042316.0, + "step": 7922 + }, + { + "epoch": 0.8700856578080387, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.144021987915039, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7047160267829895, + "num_tokens": 205069593.0, + "step": 7923 + }, + { + "epoch": 0.8701954755106523, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1223936080932617, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7123212814331055, + "num_tokens": 205095869.0, + "step": 7924 + }, + { + "epoch": 0.870305293213266, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.309396266937256, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6914612054824829, + "num_tokens": 205120650.0, + "step": 7925 + }, + { + "epoch": 0.8704151109158796, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.3473353385925293, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.6993486285209656, + "num_tokens": 205145191.0, + "step": 7926 + }, + { + "epoch": 0.8705249286184933, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2146108150482178, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7165507078170776, + "num_tokens": 205171048.0, + "step": 7927 + }, + { + "epoch": 0.8706347463211069, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0491297245025635, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7177072167396545, + "num_tokens": 205197124.0, + "step": 7928 + }, + { + "epoch": 0.8707445640237206, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.8987613916397095, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7081371545791626, + "num_tokens": 205228334.0, + "step": 7929 + }, + { + "epoch": 0.8708543817263343, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.193706512451172, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7237244844436646, + "num_tokens": 205252781.0, + "step": 7930 + }, + { + "epoch": 0.870964199428948, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1482272148132324, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.7021051049232483, + "num_tokens": 205279568.0, + "step": 7931 + }, + { + "epoch": 0.8710740171315616, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.4834518432617188, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7195504307746887, + "num_tokens": 205298739.0, + "step": 7932 + }, + { + "epoch": 0.8711838348341753, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.213341236114502, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7272971868515015, + "num_tokens": 205323281.0, + "step": 7933 + }, + { + "epoch": 0.8712936525367889, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.3891384601593018, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7063500285148621, + "num_tokens": 205346345.0, + "step": 7934 + }, + { + "epoch": 0.8714034702394026, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.375380039215088, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7155895233154297, + "num_tokens": 205367389.0, + "step": 7935 + }, + { + "epoch": 0.8715132879420162, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2395265102386475, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7017812132835388, + "num_tokens": 205390695.0, + "step": 7936 + }, + { + "epoch": 0.87162310564463, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2284257411956787, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6933248043060303, + "num_tokens": 205415769.0, + "step": 7937 + }, + { + "epoch": 0.8717329233472436, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.260331153869629, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7197780609130859, + "num_tokens": 205438970.0, + "step": 7938 + }, + { + "epoch": 0.8718427410498573, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1737213134765625, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.706829309463501, + "num_tokens": 205463899.0, + "step": 7939 + }, + { + "epoch": 0.8719525587524709, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0395212173461914, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7002071142196655, + "num_tokens": 205493986.0, + "step": 7940 + }, + { + "epoch": 0.8720623764550846, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.4759719371795654, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.70754474401474, + "num_tokens": 205516520.0, + "step": 7941 + }, + { + "epoch": 0.8721721941576982, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1858575344085693, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.6926838159561157, + "num_tokens": 205544956.0, + "step": 7942 + }, + { + "epoch": 0.8722820118603118, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.4863483905792236, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7046756148338318, + "num_tokens": 205565308.0, + "step": 7943 + }, + { + "epoch": 0.8723918295629255, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.073246717453003, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7005438804626465, + "num_tokens": 205592370.0, + "step": 7944 + }, + { + "epoch": 0.8725016472655392, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.9704786539077759, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6977821588516235, + "num_tokens": 205622412.0, + "step": 7945 + }, + { + "epoch": 0.8726114649681529, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.3339641094207764, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7061751484870911, + "num_tokens": 205646197.0, + "step": 7946 + }, + { + "epoch": 0.8727212826707665, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.257841110229492, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.712076723575592, + "num_tokens": 205669733.0, + "step": 7947 + }, + { + "epoch": 0.8728311003733802, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.9903208017349243, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7069715261459351, + "num_tokens": 205697984.0, + "step": 7948 + }, + { + "epoch": 0.8729409180759938, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.3577258586883545, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6892379522323608, + "num_tokens": 205722254.0, + "step": 7949 + }, + { + "epoch": 0.8730507357786075, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2008352279663086, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.707315981388092, + "num_tokens": 205748538.0, + "step": 7950 + }, + { + "epoch": 0.8731605534812211, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2475669384002686, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7350528240203857, + "num_tokens": 205772601.0, + "step": 7951 + }, + { + "epoch": 0.8732703711838349, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2179481983184814, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.6954854726791382, + "num_tokens": 205796423.0, + "step": 7952 + }, + { + "epoch": 0.8733801888864485, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.9967894554138184, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7257125377655029, + "num_tokens": 205826257.0, + "step": 7953 + }, + { + "epoch": 0.8734900065890622, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1822879314422607, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6974674463272095, + "num_tokens": 205850383.0, + "step": 7954 + }, + { + "epoch": 0.8735998242916758, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.5696418285369873, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7061591148376465, + "num_tokens": 205869212.0, + "step": 7955 + }, + { + "epoch": 0.8737096419942895, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.3882782459259033, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7120345830917358, + "num_tokens": 205891962.0, + "step": 7956 + }, + { + "epoch": 0.8738194596969031, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.32584810256958, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7165437340736389, + "num_tokens": 205916887.0, + "step": 7957 + }, + { + "epoch": 0.8739292773995168, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0543248653411865, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.711030900478363, + "num_tokens": 205943337.0, + "step": 7958 + }, + { + "epoch": 0.8740390951021305, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1151950359344482, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7018235921859741, + "num_tokens": 205969059.0, + "step": 7959 + }, + { + "epoch": 0.8741489128047442, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.978023648262024, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.722139835357666, + "num_tokens": 205998041.0, + "step": 7960 + }, + { + "epoch": 0.8742587305073578, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.231771230697632, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.717498779296875, + "num_tokens": 206022857.0, + "step": 7961 + }, + { + "epoch": 0.8743685482099715, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0589354038238525, + "learning_rate": 1e-06, + "loss": 1.0716, + "mean_token_accuracy": 0.6850057244300842, + "num_tokens": 206051159.0, + "step": 7962 + }, + { + "epoch": 0.8744783659125851, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.019737720489502, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7198511362075806, + "num_tokens": 206079039.0, + "step": 7963 + }, + { + "epoch": 0.8745881836151987, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1987926959991455, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7049809694290161, + "num_tokens": 206103595.0, + "step": 7964 + }, + { + "epoch": 0.8746980013178124, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2371387481689453, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7003222107887268, + "num_tokens": 206128831.0, + "step": 7965 + }, + { + "epoch": 0.8748078190204261, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2298896312713623, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7029857635498047, + "num_tokens": 206153310.0, + "step": 7966 + }, + { + "epoch": 0.8749176367230398, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.0378589630126953, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7278492450714111, + "num_tokens": 206179791.0, + "step": 7967 + }, + { + "epoch": 0.8750274544256534, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.573322057723999, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7182961106300354, + "num_tokens": 206198392.0, + "step": 7968 + }, + { + "epoch": 0.8751372721282671, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1035006046295166, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7086212635040283, + "num_tokens": 206226572.0, + "step": 7969 + }, + { + "epoch": 0.8752470898308807, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.354067325592041, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7346045970916748, + "num_tokens": 206247127.0, + "step": 7970 + }, + { + "epoch": 0.8753569075334944, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2061493396759033, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7155718803405762, + "num_tokens": 206271724.0, + "step": 7971 + }, + { + "epoch": 0.875466725236108, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 1.911329746246338, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7005150318145752, + "num_tokens": 206302578.0, + "step": 7972 + }, + { + "epoch": 0.8755765429387217, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.2979955673217773, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7145838141441345, + "num_tokens": 206328020.0, + "step": 7973 + }, + { + "epoch": 0.8756863606413354, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.097782611846924, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7198748588562012, + "num_tokens": 206353497.0, + "step": 7974 + }, + { + "epoch": 0.8757961783439491, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.1760659217834473, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7103345394134521, + "num_tokens": 206377663.0, + "step": 7975 + }, + { + "epoch": 0.8759059960465627, + "ewc_loss": 6.818771362304688e-05, + "grad_norm": 2.104203701019287, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7047106027603149, + "num_tokens": 206404111.0, + "step": 7976 + }, + { + "epoch": 0.8760158137491764, + "ewc_loss": 6.914138793945312e-05, + "grad_norm": 2.377634286880493, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7073222398757935, + "num_tokens": 206425714.0, + "step": 7977 + }, + { + "epoch": 0.87612563145179, + "ewc_loss": 6.914138793945312e-05, + "grad_norm": 2.0706474781036377, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7001453638076782, + "num_tokens": 206454801.0, + "step": 7978 + }, + { + "epoch": 0.8762354491544037, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2387304306030273, + "learning_rate": 1e-06, + "loss": 1.0633, + "mean_token_accuracy": 0.6865749359130859, + "num_tokens": 206480724.0, + "step": 7979 + }, + { + "epoch": 0.8763452668570173, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.073789358139038, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7167189717292786, + "num_tokens": 206508803.0, + "step": 7980 + }, + { + "epoch": 0.8764550845596311, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2285008430480957, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7053498029708862, + "num_tokens": 206532791.0, + "step": 7981 + }, + { + "epoch": 0.8765649022622447, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.7076566219329834, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.709688127040863, + "num_tokens": 206550452.0, + "step": 7982 + }, + { + "epoch": 0.8766747199648584, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.14898681640625, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7045232653617859, + "num_tokens": 206575311.0, + "step": 7983 + }, + { + "epoch": 0.876784537667472, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9270968437194824, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6940307021141052, + "num_tokens": 206606387.0, + "step": 7984 + }, + { + "epoch": 0.8768943553700856, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1080284118652344, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7245151400566101, + "num_tokens": 206631078.0, + "step": 7985 + }, + { + "epoch": 0.8770041730726993, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.255854845046997, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.7035843133926392, + "num_tokens": 206656326.0, + "step": 7986 + }, + { + "epoch": 0.8771139907753129, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9934173822402954, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6982381939888, + "num_tokens": 206687052.0, + "step": 7987 + }, + { + "epoch": 0.8772238084779267, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.494499921798706, + "learning_rate": 1e-06, + "loss": 1.0643, + "mean_token_accuracy": 0.6755398511886597, + "num_tokens": 206708479.0, + "step": 7988 + }, + { + "epoch": 0.8773336261805403, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1236460208892822, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.69272381067276, + "num_tokens": 206734578.0, + "step": 7989 + }, + { + "epoch": 0.877443443883154, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0356435775756836, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7214108109474182, + "num_tokens": 206761012.0, + "step": 7990 + }, + { + "epoch": 0.8775532615857676, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.23464035987854, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7283691167831421, + "num_tokens": 206782739.0, + "step": 7991 + }, + { + "epoch": 0.8776630792883813, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9387950897216797, + "learning_rate": 1e-06, + "loss": 1.0535, + "mean_token_accuracy": 0.6788891553878784, + "num_tokens": 206816363.0, + "step": 7992 + }, + { + "epoch": 0.8777728969909949, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2510018348693848, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7227000594139099, + "num_tokens": 206840364.0, + "step": 7993 + }, + { + "epoch": 0.8778827146936086, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.31501841545105, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6900486946105957, + "num_tokens": 206868088.0, + "step": 7994 + }, + { + "epoch": 0.8779925323962223, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4982352256774902, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7067304849624634, + "num_tokens": 206889482.0, + "step": 7995 + }, + { + "epoch": 0.878102350098836, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.3094053268432617, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7068259119987488, + "num_tokens": 206912349.0, + "step": 7996 + }, + { + "epoch": 0.8782121678014496, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0743825435638428, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7207531332969666, + "num_tokens": 206937852.0, + "step": 7997 + }, + { + "epoch": 0.8783219855040633, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2342400550842285, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7314988970756531, + "num_tokens": 206961008.0, + "step": 7998 + }, + { + "epoch": 0.8784318032066769, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.3320908546447754, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.699964165687561, + "num_tokens": 206983322.0, + "step": 7999 + }, + { + "epoch": 0.8785416209092906, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2799484729766846, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7060595750808716, + "num_tokens": 207007351.0, + "step": 8000 + }, + { + "epoch": 0.8786514386119042, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9349292516708374, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7102501392364502, + "num_tokens": 207038395.0, + "step": 8001 + }, + { + "epoch": 0.8787612563145178, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2851338386535645, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7130162119865417, + "num_tokens": 207062632.0, + "step": 8002 + }, + { + "epoch": 0.8788710740171316, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9488152265548706, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6888041496276855, + "num_tokens": 207092243.0, + "step": 8003 + }, + { + "epoch": 0.8789808917197452, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9566503763198853, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.6990770101547241, + "num_tokens": 207122819.0, + "step": 8004 + }, + { + "epoch": 0.8790907094223589, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2683234214782715, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7151871919631958, + "num_tokens": 207147033.0, + "step": 8005 + }, + { + "epoch": 0.8792005271249725, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.186849355697632, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6893078088760376, + "num_tokens": 207172566.0, + "step": 8006 + }, + { + "epoch": 0.8793103448275862, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9816484451293945, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7043171525001526, + "num_tokens": 207201224.0, + "step": 8007 + }, + { + "epoch": 0.8794201625301998, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0668785572052, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7229114770889282, + "num_tokens": 207228077.0, + "step": 8008 + }, + { + "epoch": 0.8795299802328135, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.061891794204712, + "learning_rate": 1e-06, + "loss": 1.0375, + "mean_token_accuracy": 0.6821863055229187, + "num_tokens": 207256588.0, + "step": 8009 + }, + { + "epoch": 0.8796397979354272, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2319633960723877, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7182564735412598, + "num_tokens": 207278042.0, + "step": 8010 + }, + { + "epoch": 0.8797496156380409, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1486189365386963, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6880127787590027, + "num_tokens": 207305518.0, + "step": 8011 + }, + { + "epoch": 0.8798594333406545, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9561657905578613, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7093825936317444, + "num_tokens": 207335128.0, + "step": 8012 + }, + { + "epoch": 0.8799692510432682, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4478445053100586, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7181894183158875, + "num_tokens": 207357774.0, + "step": 8013 + }, + { + "epoch": 0.8800790687458818, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.213620662689209, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7136953473091125, + "num_tokens": 207383683.0, + "step": 8014 + }, + { + "epoch": 0.8801888864484955, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.983737826347351, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7142300605773926, + "num_tokens": 207413913.0, + "step": 8015 + }, + { + "epoch": 0.8802987041511091, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.843726634979248, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7082086205482483, + "num_tokens": 207449426.0, + "step": 8016 + }, + { + "epoch": 0.8804085218537229, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1465938091278076, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7071471214294434, + "num_tokens": 207476016.0, + "step": 8017 + }, + { + "epoch": 0.8805183395563365, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.354421615600586, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7155174016952515, + "num_tokens": 207497955.0, + "step": 8018 + }, + { + "epoch": 0.8806281572589502, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.481519937515259, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7412352561950684, + "num_tokens": 207516932.0, + "step": 8019 + }, + { + "epoch": 0.8807379749615638, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1049447059631348, + "learning_rate": 1e-06, + "loss": 1.0417, + "mean_token_accuracy": 0.6884809732437134, + "num_tokens": 207545331.0, + "step": 8020 + }, + { + "epoch": 0.8808477926641775, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.3458120822906494, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7129508256912231, + "num_tokens": 207568324.0, + "step": 8021 + }, + { + "epoch": 0.8809576103667911, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2631635665893555, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7231384515762329, + "num_tokens": 207590390.0, + "step": 8022 + }, + { + "epoch": 0.8810674280694047, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.104423999786377, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7270370721817017, + "num_tokens": 207614861.0, + "step": 8023 + }, + { + "epoch": 0.8811772457720185, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.036867380142212, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7025676369667053, + "num_tokens": 207643294.0, + "step": 8024 + }, + { + "epoch": 0.8812870634746321, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4937429428100586, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7038681507110596, + "num_tokens": 207663473.0, + "step": 8025 + }, + { + "epoch": 0.8813968811772458, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0694026947021484, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7132947444915771, + "num_tokens": 207693923.0, + "step": 8026 + }, + { + "epoch": 0.8815066988798594, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.096115827560425, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.730305016040802, + "num_tokens": 207722358.0, + "step": 8027 + }, + { + "epoch": 0.8816165165824731, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.196537733078003, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7098608016967773, + "num_tokens": 207748579.0, + "step": 8028 + }, + { + "epoch": 0.8817263342850867, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.3642570972442627, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7239817976951599, + "num_tokens": 207771070.0, + "step": 8029 + }, + { + "epoch": 0.8818361519877004, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.5817341804504395, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7202882170677185, + "num_tokens": 207789767.0, + "step": 8030 + }, + { + "epoch": 0.881945969690314, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.338253974914551, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7214983701705933, + "num_tokens": 207810612.0, + "step": 8031 + }, + { + "epoch": 0.8820557873929278, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.209277868270874, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6964643001556396, + "num_tokens": 207836096.0, + "step": 8032 + }, + { + "epoch": 0.8821656050955414, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.844144582748413, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7043153643608093, + "num_tokens": 207868083.0, + "step": 8033 + }, + { + "epoch": 0.8822754227981551, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2543416023254395, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7217909097671509, + "num_tokens": 207892881.0, + "step": 8034 + }, + { + "epoch": 0.8823852405007687, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9529435634613037, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6955263614654541, + "num_tokens": 207924786.0, + "step": 8035 + }, + { + "epoch": 0.8824950582033824, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.5420448780059814, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7244106531143188, + "num_tokens": 207946705.0, + "step": 8036 + }, + { + "epoch": 0.882604875905996, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2863481044769287, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7323911786079407, + "num_tokens": 207968344.0, + "step": 8037 + }, + { + "epoch": 0.8827146936086097, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.3222007751464844, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7114447355270386, + "num_tokens": 207991588.0, + "step": 8038 + }, + { + "epoch": 0.8828245113112234, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0843021869659424, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6961108446121216, + "num_tokens": 208020156.0, + "step": 8039 + }, + { + "epoch": 0.8829343290138371, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.5763564109802246, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7112997770309448, + "num_tokens": 208039776.0, + "step": 8040 + }, + { + "epoch": 0.8830441467164507, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.8237369060516357, + "learning_rate": 1e-06, + "loss": 1.0709, + "mean_token_accuracy": 0.6830687522888184, + "num_tokens": 208078399.0, + "step": 8041 + }, + { + "epoch": 0.8831539644190644, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4201443195343018, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7141942977905273, + "num_tokens": 208098993.0, + "step": 8042 + }, + { + "epoch": 0.883263782121678, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.951720952987671, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.709372878074646, + "num_tokens": 208132064.0, + "step": 8043 + }, + { + "epoch": 0.8833735998242916, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1015214920043945, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7313507795333862, + "num_tokens": 208158784.0, + "step": 8044 + }, + { + "epoch": 0.8834834175269053, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1806161403656006, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.7026359438896179, + "num_tokens": 208184449.0, + "step": 8045 + }, + { + "epoch": 0.883593235229519, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4575302600860596, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7228690385818481, + "num_tokens": 208204624.0, + "step": 8046 + }, + { + "epoch": 0.8837030529321327, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4389841556549072, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7011496424674988, + "num_tokens": 208227598.0, + "step": 8047 + }, + { + "epoch": 0.8838128706347463, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.339099884033203, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7274036407470703, + "num_tokens": 208251223.0, + "step": 8048 + }, + { + "epoch": 0.88392268833736, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.5388758182525635, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7021307349205017, + "num_tokens": 208271542.0, + "step": 8049 + }, + { + "epoch": 0.8840325060399736, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.112098455429077, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7024881839752197, + "num_tokens": 208300464.0, + "step": 8050 + }, + { + "epoch": 0.8841423237425873, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.466693878173828, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7155138254165649, + "num_tokens": 208320189.0, + "step": 8051 + }, + { + "epoch": 0.8842521414452009, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.888363242149353, + "learning_rate": 1e-06, + "loss": 1.0689, + "mean_token_accuracy": 0.6909281015396118, + "num_tokens": 208355152.0, + "step": 8052 + }, + { + "epoch": 0.8843619591478147, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9690974950790405, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6888037323951721, + "num_tokens": 208383984.0, + "step": 8053 + }, + { + "epoch": 0.8844717768504283, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.096456527709961, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7175100445747375, + "num_tokens": 208410915.0, + "step": 8054 + }, + { + "epoch": 0.884581594553042, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1553235054016113, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7455947995185852, + "num_tokens": 208436148.0, + "step": 8055 + }, + { + "epoch": 0.8846914122556556, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1239230632781982, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7115481495857239, + "num_tokens": 208465224.0, + "step": 8056 + }, + { + "epoch": 0.8848012299582693, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9086657762527466, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7348217368125916, + "num_tokens": 208494111.0, + "step": 8057 + }, + { + "epoch": 0.8849110476608829, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.354635238647461, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7275071144104004, + "num_tokens": 208514390.0, + "step": 8058 + }, + { + "epoch": 0.8850208653634966, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.992274522781372, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.6996781229972839, + "num_tokens": 208543421.0, + "step": 8059 + }, + { + "epoch": 0.8851306830661102, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1390795707702637, + "learning_rate": 1e-06, + "loss": 1.0561, + "mean_token_accuracy": 0.6868785619735718, + "num_tokens": 208570744.0, + "step": 8060 + }, + { + "epoch": 0.885240500768724, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1573901176452637, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7189109325408936, + "num_tokens": 208596170.0, + "step": 8061 + }, + { + "epoch": 0.8853503184713376, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.084134817123413, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6971378326416016, + "num_tokens": 208625206.0, + "step": 8062 + }, + { + "epoch": 0.8854601361739513, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4082493782043457, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7260340452194214, + "num_tokens": 208645910.0, + "step": 8063 + }, + { + "epoch": 0.8855699538765649, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.080801486968994, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7074638605117798, + "num_tokens": 208672805.0, + "step": 8064 + }, + { + "epoch": 0.8856797715791785, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.9848533868789673, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7060012221336365, + "num_tokens": 208703429.0, + "step": 8065 + }, + { + "epoch": 0.8857895892817922, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1517486572265625, + "learning_rate": 1e-06, + "loss": 1.0647, + "mean_token_accuracy": 0.6835051774978638, + "num_tokens": 208732360.0, + "step": 8066 + }, + { + "epoch": 0.8858994069844058, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1656041145324707, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.7022227048873901, + "num_tokens": 208759178.0, + "step": 8067 + }, + { + "epoch": 0.8860092246870196, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.507392644882202, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7156553864479065, + "num_tokens": 208778917.0, + "step": 8068 + }, + { + "epoch": 0.8861190423896332, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0186233520507812, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7109541893005371, + "num_tokens": 208808384.0, + "step": 8069 + }, + { + "epoch": 0.8862288600922469, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4597272872924805, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7138448357582092, + "num_tokens": 208829418.0, + "step": 8070 + }, + { + "epoch": 0.8863386777948605, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0738584995269775, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.7041300535202026, + "num_tokens": 208857521.0, + "step": 8071 + }, + { + "epoch": 0.8864484954974742, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.228466510772705, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6974616050720215, + "num_tokens": 208881906.0, + "step": 8072 + }, + { + "epoch": 0.8865583132000878, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 1.8433139324188232, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.734416127204895, + "num_tokens": 208914194.0, + "step": 8073 + }, + { + "epoch": 0.8866681309027015, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2625982761383057, + "learning_rate": 1e-06, + "loss": 1.0511, + "mean_token_accuracy": 0.6849042177200317, + "num_tokens": 208939704.0, + "step": 8074 + }, + { + "epoch": 0.8867779486053152, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1711764335632324, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7314398884773254, + "num_tokens": 208963714.0, + "step": 8075 + }, + { + "epoch": 0.8868877663079289, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.113036870956421, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6914211511611938, + "num_tokens": 208992718.0, + "step": 8076 + }, + { + "epoch": 0.8869975840105425, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0377254486083984, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7178994417190552, + "num_tokens": 209021300.0, + "step": 8077 + }, + { + "epoch": 0.8871074017131562, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1757776737213135, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.698505699634552, + "num_tokens": 209048287.0, + "step": 8078 + }, + { + "epoch": 0.8872172194157698, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.4785375595092773, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7062787413597107, + "num_tokens": 209069545.0, + "step": 8079 + }, + { + "epoch": 0.8873270371183835, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.098477840423584, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7035529613494873, + "num_tokens": 209095028.0, + "step": 8080 + }, + { + "epoch": 0.8874368548209971, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.2207300662994385, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7245222330093384, + "num_tokens": 209118193.0, + "step": 8081 + }, + { + "epoch": 0.8875466725236109, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.010270833969116, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7076253890991211, + "num_tokens": 209146980.0, + "step": 8082 + }, + { + "epoch": 0.8876564902262245, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.0890331268310547, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7012059688568115, + "num_tokens": 209174812.0, + "step": 8083 + }, + { + "epoch": 0.8877663079288381, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1930887699127197, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7211159467697144, + "num_tokens": 209200809.0, + "step": 8084 + }, + { + "epoch": 0.8878761256314518, + "ewc_loss": 6.961822509765625e-05, + "grad_norm": 2.1436264514923096, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7326720952987671, + "num_tokens": 209225031.0, + "step": 8085 + }, + { + "epoch": 0.8879859433340654, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1633317470550537, + "learning_rate": 1e-06, + "loss": 1.0598, + "mean_token_accuracy": 0.6797809600830078, + "num_tokens": 209253581.0, + "step": 8086 + }, + { + "epoch": 0.8880957610366791, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1036508083343506, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6853026747703552, + "num_tokens": 209283539.0, + "step": 8087 + }, + { + "epoch": 0.8882055787392927, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9725775718688965, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6898786425590515, + "num_tokens": 209313896.0, + "step": 8088 + }, + { + "epoch": 0.8883153964419065, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.063406467437744, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7224153280258179, + "num_tokens": 209338742.0, + "step": 8089 + }, + { + "epoch": 0.8884252141445201, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.459902048110962, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7141432762145996, + "num_tokens": 209358816.0, + "step": 8090 + }, + { + "epoch": 0.8885350318471338, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.213243007659912, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7121158838272095, + "num_tokens": 209383834.0, + "step": 8091 + }, + { + "epoch": 0.8886448495497474, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1185896396636963, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.6966375112533569, + "num_tokens": 209410554.0, + "step": 8092 + }, + { + "epoch": 0.8887546672523611, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.485253095626831, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7168201208114624, + "num_tokens": 209430014.0, + "step": 8093 + }, + { + "epoch": 0.8888644849549747, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3587749004364014, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7086895108222961, + "num_tokens": 209450901.0, + "step": 8094 + }, + { + "epoch": 0.8889743026575884, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0409045219421387, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7000583410263062, + "num_tokens": 209480423.0, + "step": 8095 + }, + { + "epoch": 0.889084120360202, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0034968852996826, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.718350887298584, + "num_tokens": 209509147.0, + "step": 8096 + }, + { + "epoch": 0.8891939380628158, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9383654594421387, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7145475745201111, + "num_tokens": 209539620.0, + "step": 8097 + }, + { + "epoch": 0.8893037557654294, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3528032302856445, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7322039604187012, + "num_tokens": 209560177.0, + "step": 8098 + }, + { + "epoch": 0.8894135734680431, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.127230644226074, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7138575911521912, + "num_tokens": 209585869.0, + "step": 8099 + }, + { + "epoch": 0.8895233911706567, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.4447882175445557, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6977911591529846, + "num_tokens": 209607506.0, + "step": 8100 + }, + { + "epoch": 0.8896332088732704, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2025582790374756, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7045791745185852, + "num_tokens": 209634975.0, + "step": 8101 + }, + { + "epoch": 0.889743026575884, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1763815879821777, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7093437910079956, + "num_tokens": 209661090.0, + "step": 8102 + }, + { + "epoch": 0.8898528442784976, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1337239742279053, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7034155130386353, + "num_tokens": 209687591.0, + "step": 8103 + }, + { + "epoch": 0.8899626619811114, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3658788204193115, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7179050445556641, + "num_tokens": 209709367.0, + "step": 8104 + }, + { + "epoch": 0.890072479683725, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9555941820144653, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7081539630889893, + "num_tokens": 209740280.0, + "step": 8105 + }, + { + "epoch": 0.8901822973863387, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0718350410461426, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7257844805717468, + "num_tokens": 209766365.0, + "step": 8106 + }, + { + "epoch": 0.8902921150889523, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.21018648147583, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7091712355613708, + "num_tokens": 209792007.0, + "step": 8107 + }, + { + "epoch": 0.890401932791566, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.931241512298584, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7092354893684387, + "num_tokens": 209820382.0, + "step": 8108 + }, + { + "epoch": 0.8905117504941796, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.356717586517334, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7011938095092773, + "num_tokens": 209844582.0, + "step": 8109 + }, + { + "epoch": 0.8906215681967933, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0024492740631104, + "learning_rate": 1e-06, + "loss": 1.0936, + "mean_token_accuracy": 0.6724014282226562, + "num_tokens": 209876586.0, + "step": 8110 + }, + { + "epoch": 0.890731385899407, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.241363048553467, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7205746173858643, + "num_tokens": 209901208.0, + "step": 8111 + }, + { + "epoch": 0.8908412036020207, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2893104553222656, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7424026727676392, + "num_tokens": 209922889.0, + "step": 8112 + }, + { + "epoch": 0.8909510213046343, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.15657377243042, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.707161009311676, + "num_tokens": 209948901.0, + "step": 8113 + }, + { + "epoch": 0.891060839007248, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2735602855682373, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7082866430282593, + "num_tokens": 209972783.0, + "step": 8114 + }, + { + "epoch": 0.8911706567098616, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9589670896530151, + "learning_rate": 1e-06, + "loss": 1.0557, + "mean_token_accuracy": 0.6803780794143677, + "num_tokens": 210003463.0, + "step": 8115 + }, + { + "epoch": 0.8912804744124753, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.284982681274414, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7110165357589722, + "num_tokens": 210026671.0, + "step": 8116 + }, + { + "epoch": 0.8913902921150889, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1693477630615234, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7266378998756409, + "num_tokens": 210049798.0, + "step": 8117 + }, + { + "epoch": 0.8915001098177027, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1589455604553223, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6988372802734375, + "num_tokens": 210076020.0, + "step": 8118 + }, + { + "epoch": 0.8916099275203163, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9942492246627808, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6920979022979736, + "num_tokens": 210106761.0, + "step": 8119 + }, + { + "epoch": 0.89171974522293, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.17781400680542, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7240491509437561, + "num_tokens": 210132344.0, + "step": 8120 + }, + { + "epoch": 0.8918295629255436, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1489651203155518, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7054443359375, + "num_tokens": 210159175.0, + "step": 8121 + }, + { + "epoch": 0.8919393806281573, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0854852199554443, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6926865577697754, + "num_tokens": 210189422.0, + "step": 8122 + }, + { + "epoch": 0.8920491983307709, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0035407543182373, + "learning_rate": 1e-06, + "loss": 1.0426, + "mean_token_accuracy": 0.689495861530304, + "num_tokens": 210220355.0, + "step": 8123 + }, + { + "epoch": 0.8921590160333845, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2656290531158447, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7148793935775757, + "num_tokens": 210243366.0, + "step": 8124 + }, + { + "epoch": 0.8922688337359982, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.006542921066284, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7016830444335938, + "num_tokens": 210273857.0, + "step": 8125 + }, + { + "epoch": 0.892378651438612, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3342158794403076, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6947646141052246, + "num_tokens": 210298244.0, + "step": 8126 + }, + { + "epoch": 0.8924884691412256, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0830941200256348, + "learning_rate": 1e-06, + "loss": 1.0484, + "mean_token_accuracy": 0.6855037212371826, + "num_tokens": 210329923.0, + "step": 8127 + }, + { + "epoch": 0.8925982868438392, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.056973695755005, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.7031691074371338, + "num_tokens": 210358538.0, + "step": 8128 + }, + { + "epoch": 0.8927081045464529, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.01334285736084, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7093518376350403, + "num_tokens": 210386966.0, + "step": 8129 + }, + { + "epoch": 0.8928179222490665, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.117619752883911, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.694344699382782, + "num_tokens": 210414246.0, + "step": 8130 + }, + { + "epoch": 0.8929277399516802, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.33195161819458, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6946917176246643, + "num_tokens": 210437784.0, + "step": 8131 + }, + { + "epoch": 0.8930375576542938, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1869802474975586, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7026798129081726, + "num_tokens": 210463775.0, + "step": 8132 + }, + { + "epoch": 0.8931473753569076, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.0717782974243164, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.6952264308929443, + "num_tokens": 210492247.0, + "step": 8133 + }, + { + "epoch": 0.8932571930595212, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.267430543899536, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.6992545127868652, + "num_tokens": 210516236.0, + "step": 8134 + }, + { + "epoch": 0.8933670107621349, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3100051879882812, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7241230010986328, + "num_tokens": 210539167.0, + "step": 8135 + }, + { + "epoch": 0.8934768284647485, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2619855403900146, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7178175449371338, + "num_tokens": 210562684.0, + "step": 8136 + }, + { + "epoch": 0.8935866461673622, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.891100287437439, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.722184419631958, + "num_tokens": 210596516.0, + "step": 8137 + }, + { + "epoch": 0.8936964638699758, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1075456142425537, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7239018678665161, + "num_tokens": 210623295.0, + "step": 8138 + }, + { + "epoch": 0.8938062815725895, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.4116387367248535, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.726231575012207, + "num_tokens": 210643221.0, + "step": 8139 + }, + { + "epoch": 0.8939160992752032, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.151087760925293, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6976528167724609, + "num_tokens": 210670898.0, + "step": 8140 + }, + { + "epoch": 0.8940259169778169, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.372370719909668, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7076531052589417, + "num_tokens": 210694162.0, + "step": 8141 + }, + { + "epoch": 0.8941357346804305, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1705474853515625, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.6967383623123169, + "num_tokens": 210719930.0, + "step": 8142 + }, + { + "epoch": 0.8942455523830442, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2351269721984863, + "learning_rate": 1e-06, + "loss": 1.0382, + "mean_token_accuracy": 0.6876371502876282, + "num_tokens": 210747367.0, + "step": 8143 + }, + { + "epoch": 0.8943553700856578, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.150371789932251, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.7089650630950928, + "num_tokens": 210773069.0, + "step": 8144 + }, + { + "epoch": 0.8944651877882714, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.082817316055298, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7136589884757996, + "num_tokens": 210801344.0, + "step": 8145 + }, + { + "epoch": 0.8945750054908851, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.8524868488311768, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7173155546188354, + "num_tokens": 210835548.0, + "step": 8146 + }, + { + "epoch": 0.8946848231934988, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.290086507797241, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7274349927902222, + "num_tokens": 210857281.0, + "step": 8147 + }, + { + "epoch": 0.8947946408961125, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.120032787322998, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.7086721658706665, + "num_tokens": 210884292.0, + "step": 8148 + }, + { + "epoch": 0.8949044585987261, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2329306602478027, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7083530426025391, + "num_tokens": 210908448.0, + "step": 8149 + }, + { + "epoch": 0.8950142763013398, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3429317474365234, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6926299333572388, + "num_tokens": 210932876.0, + "step": 8150 + }, + { + "epoch": 0.8951240940039534, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9164983034133911, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7155435085296631, + "num_tokens": 210963703.0, + "step": 8151 + }, + { + "epoch": 0.8952339117065671, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1809818744659424, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7118020057678223, + "num_tokens": 210990458.0, + "step": 8152 + }, + { + "epoch": 0.8953437294091807, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9361811876296997, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.695825457572937, + "num_tokens": 211021700.0, + "step": 8153 + }, + { + "epoch": 0.8954535471117944, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3604416847229004, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7069557905197144, + "num_tokens": 211044523.0, + "step": 8154 + }, + { + "epoch": 0.8955633648144081, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1799402236938477, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7120382785797119, + "num_tokens": 211067809.0, + "step": 8155 + }, + { + "epoch": 0.8956731825170218, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2101027965545654, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.704245924949646, + "num_tokens": 211093559.0, + "step": 8156 + }, + { + "epoch": 0.8957830002196354, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.191472053527832, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7199347615242004, + "num_tokens": 211120767.0, + "step": 8157 + }, + { + "epoch": 0.8958928179222491, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2137699127197266, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7170740365982056, + "num_tokens": 211144996.0, + "step": 8158 + }, + { + "epoch": 0.8960026356248627, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.988312840461731, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7182263135910034, + "num_tokens": 211175424.0, + "step": 8159 + }, + { + "epoch": 0.8961124533274764, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 1.9861127138137817, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.706592321395874, + "num_tokens": 211204392.0, + "step": 8160 + }, + { + "epoch": 0.89622227103009, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.2696964740753174, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7048965096473694, + "num_tokens": 211229550.0, + "step": 8161 + }, + { + "epoch": 0.8963320887327038, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.157038688659668, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7210034132003784, + "num_tokens": 211255545.0, + "step": 8162 + }, + { + "epoch": 0.8964419064353174, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.029545307159424, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.700019359588623, + "num_tokens": 211283602.0, + "step": 8163 + }, + { + "epoch": 0.896551724137931, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.529628276824951, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7076128125190735, + "num_tokens": 211302935.0, + "step": 8164 + }, + { + "epoch": 0.8966615418405447, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.6497912406921387, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7080526351928711, + "num_tokens": 211321835.0, + "step": 8165 + }, + { + "epoch": 0.8967713595431583, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.027190685272217, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7008436918258667, + "num_tokens": 211351891.0, + "step": 8166 + }, + { + "epoch": 0.896881177245772, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.355328321456909, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6999374628067017, + "num_tokens": 211372892.0, + "step": 8167 + }, + { + "epoch": 0.8969909949483856, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.233262062072754, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7184491157531738, + "num_tokens": 211394671.0, + "step": 8168 + }, + { + "epoch": 0.8971008126509994, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.010061502456665, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7155846357345581, + "num_tokens": 211422879.0, + "step": 8169 + }, + { + "epoch": 0.897210630353613, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.034045934677124, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7047184109687805, + "num_tokens": 211450565.0, + "step": 8170 + }, + { + "epoch": 0.8973204480562267, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.38627290725708, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7034639120101929, + "num_tokens": 211471829.0, + "step": 8171 + }, + { + "epoch": 0.8974302657588403, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3190596103668213, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7254579067230225, + "num_tokens": 211495522.0, + "step": 8172 + }, + { + "epoch": 0.897540083461454, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.137282609939575, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.714326024055481, + "num_tokens": 211522860.0, + "step": 8173 + }, + { + "epoch": 0.8976499011640676, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.174818992614746, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6972055435180664, + "num_tokens": 211548583.0, + "step": 8174 + }, + { + "epoch": 0.8977597188666813, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1948299407958984, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7093074321746826, + "num_tokens": 211572444.0, + "step": 8175 + }, + { + "epoch": 0.897869536569295, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3943064212799072, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7260607481002808, + "num_tokens": 211594694.0, + "step": 8176 + }, + { + "epoch": 0.8979793542719087, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.534379005432129, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7059345245361328, + "num_tokens": 211614529.0, + "step": 8177 + }, + { + "epoch": 0.8980891719745223, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.1781280040740967, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7155141830444336, + "num_tokens": 211640705.0, + "step": 8178 + }, + { + "epoch": 0.898198989677136, + "ewc_loss": 7.009506225585938e-05, + "grad_norm": 2.3068315982818604, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7136799097061157, + "num_tokens": 211664393.0, + "step": 8179 + }, + { + "epoch": 0.8983088073797496, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.134317398071289, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7168102264404297, + "num_tokens": 211689982.0, + "step": 8180 + }, + { + "epoch": 0.8984186250823633, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.5928547382354736, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7136515378952026, + "num_tokens": 211708693.0, + "step": 8181 + }, + { + "epoch": 0.8985284427849769, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3082516193389893, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6966938972473145, + "num_tokens": 211732231.0, + "step": 8182 + }, + { + "epoch": 0.8986382604875905, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9553333520889282, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7188184261322021, + "num_tokens": 211760191.0, + "step": 8183 + }, + { + "epoch": 0.8987480781902043, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1463422775268555, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6976985335350037, + "num_tokens": 211786463.0, + "step": 8184 + }, + { + "epoch": 0.898857895892818, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1963179111480713, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7121766209602356, + "num_tokens": 211812926.0, + "step": 8185 + }, + { + "epoch": 0.8989677135954316, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.294656276702881, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.731853187084198, + "num_tokens": 211834368.0, + "step": 8186 + }, + { + "epoch": 0.8990775312980452, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.5729808807373047, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.715811014175415, + "num_tokens": 211853863.0, + "step": 8187 + }, + { + "epoch": 0.8991873490006589, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.29423451423645, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6990987658500671, + "num_tokens": 211877620.0, + "step": 8188 + }, + { + "epoch": 0.8992971667032725, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.058635711669922, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6945505738258362, + "num_tokens": 211908189.0, + "step": 8189 + }, + { + "epoch": 0.8994069844058862, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.199587821960449, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6987878084182739, + "num_tokens": 211935489.0, + "step": 8190 + }, + { + "epoch": 0.8995168021084999, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3189656734466553, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7128205299377441, + "num_tokens": 211958611.0, + "step": 8191 + }, + { + "epoch": 0.8996266198111136, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1169865131378174, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7088561058044434, + "num_tokens": 211983462.0, + "step": 8192 + }, + { + "epoch": 0.8997364375137272, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.016003131866455, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7257314920425415, + "num_tokens": 212011557.0, + "step": 8193 + }, + { + "epoch": 0.8998462552163409, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.894572377204895, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7269818782806396, + "num_tokens": 212042302.0, + "step": 8194 + }, + { + "epoch": 0.8999560729189545, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.326134204864502, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7056962251663208, + "num_tokens": 212065357.0, + "step": 8195 + }, + { + "epoch": 0.9000658906215682, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1665234565734863, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7083470225334167, + "num_tokens": 212092573.0, + "step": 8196 + }, + { + "epoch": 0.9001757083241818, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0786423683166504, + "learning_rate": 1e-06, + "loss": 1.0589, + "mean_token_accuracy": 0.6810108423233032, + "num_tokens": 212120452.0, + "step": 8197 + }, + { + "epoch": 0.9002855260267956, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.119932174682617, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7199616432189941, + "num_tokens": 212146205.0, + "step": 8198 + }, + { + "epoch": 0.9003953437294092, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.275128126144409, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7200227975845337, + "num_tokens": 212170022.0, + "step": 8199 + }, + { + "epoch": 0.9005051614320229, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.102879285812378, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7050666213035583, + "num_tokens": 212197763.0, + "step": 8200 + }, + { + "epoch": 0.9006149791346365, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.234651565551758, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6999340057373047, + "num_tokens": 212223065.0, + "step": 8201 + }, + { + "epoch": 0.9007247968372502, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3241164684295654, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7066019177436829, + "num_tokens": 212245488.0, + "step": 8202 + }, + { + "epoch": 0.9008346145398638, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0973665714263916, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6959911584854126, + "num_tokens": 212274724.0, + "step": 8203 + }, + { + "epoch": 0.9009444322424774, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.214339256286621, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7354686260223389, + "num_tokens": 212296850.0, + "step": 8204 + }, + { + "epoch": 0.9010542499450912, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.104783058166504, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6853713989257812, + "num_tokens": 212324861.0, + "step": 8205 + }, + { + "epoch": 0.9011640676477048, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2665812969207764, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6965790390968323, + "num_tokens": 212350599.0, + "step": 8206 + }, + { + "epoch": 0.9012738853503185, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3099608421325684, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7145218849182129, + "num_tokens": 212374498.0, + "step": 8207 + }, + { + "epoch": 0.9013837030529321, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.26239013671875, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.6927330493927002, + "num_tokens": 212400734.0, + "step": 8208 + }, + { + "epoch": 0.9014935207555458, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.082401752471924, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7279404401779175, + "num_tokens": 212428540.0, + "step": 8209 + }, + { + "epoch": 0.9016033384581594, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3683278560638428, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7228270173072815, + "num_tokens": 212450911.0, + "step": 8210 + }, + { + "epoch": 0.9017131561607731, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3086888790130615, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7342929840087891, + "num_tokens": 212473053.0, + "step": 8211 + }, + { + "epoch": 0.9018229738633867, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2299606800079346, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6995946168899536, + "num_tokens": 212498830.0, + "step": 8212 + }, + { + "epoch": 0.9019327915660005, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9194035530090332, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6983122825622559, + "num_tokens": 212530401.0, + "step": 8213 + }, + { + "epoch": 0.9020426092686141, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.109531879425049, + "learning_rate": 1e-06, + "loss": 1.0385, + "mean_token_accuracy": 0.6898859739303589, + "num_tokens": 212558520.0, + "step": 8214 + }, + { + "epoch": 0.9021524269712278, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.115107536315918, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6937536001205444, + "num_tokens": 212586735.0, + "step": 8215 + }, + { + "epoch": 0.9022622446738414, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0778861045837402, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6983502507209778, + "num_tokens": 212613134.0, + "step": 8216 + }, + { + "epoch": 0.9023720623764551, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.145411968231201, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.7005820274353027, + "num_tokens": 212638391.0, + "step": 8217 + }, + { + "epoch": 0.9024818800790687, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2622294425964355, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7189902067184448, + "num_tokens": 212663725.0, + "step": 8218 + }, + { + "epoch": 0.9025916977816824, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.087916135787964, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.711821973323822, + "num_tokens": 212693559.0, + "step": 8219 + }, + { + "epoch": 0.9027015154842961, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.247083902359009, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.710578978061676, + "num_tokens": 212716262.0, + "step": 8220 + }, + { + "epoch": 0.9028113331869098, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.029970645904541, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7162097692489624, + "num_tokens": 212747179.0, + "step": 8221 + }, + { + "epoch": 0.9029211508895234, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0470852851867676, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7105557918548584, + "num_tokens": 212774828.0, + "step": 8222 + }, + { + "epoch": 0.903030968592137, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9552466869354248, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6931860446929932, + "num_tokens": 212808125.0, + "step": 8223 + }, + { + "epoch": 0.9031407862947507, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2115108966827393, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6873084902763367, + "num_tokens": 212834200.0, + "step": 8224 + }, + { + "epoch": 0.9032506039973643, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1534616947174072, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.6847065687179565, + "num_tokens": 212861783.0, + "step": 8225 + }, + { + "epoch": 0.903360421699978, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9437614679336548, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.6935041546821594, + "num_tokens": 212893688.0, + "step": 8226 + }, + { + "epoch": 0.9034702394025917, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.009629487991333, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7096117734909058, + "num_tokens": 212923304.0, + "step": 8227 + }, + { + "epoch": 0.9035800571052054, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0253889560699463, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6881144642829895, + "num_tokens": 212955307.0, + "step": 8228 + }, + { + "epoch": 0.903689874807819, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.276169776916504, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7098678350448608, + "num_tokens": 212976112.0, + "step": 8229 + }, + { + "epoch": 0.9037996925104327, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.569521188735962, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7170039415359497, + "num_tokens": 212995197.0, + "step": 8230 + }, + { + "epoch": 0.9039095102130463, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1083478927612305, + "learning_rate": 1e-06, + "loss": 1.0962, + "mean_token_accuracy": 0.677377462387085, + "num_tokens": 213024273.0, + "step": 8231 + }, + { + "epoch": 0.90401932791566, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9407274723052979, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7164537906646729, + "num_tokens": 213053233.0, + "step": 8232 + }, + { + "epoch": 0.9041291456182736, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.4175312519073486, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7279614210128784, + "num_tokens": 213074531.0, + "step": 8233 + }, + { + "epoch": 0.9042389633208874, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.031153917312622, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7043907046318054, + "num_tokens": 213103128.0, + "step": 8234 + }, + { + "epoch": 0.904348781023501, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.354642868041992, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7153046131134033, + "num_tokens": 213124284.0, + "step": 8235 + }, + { + "epoch": 0.9044585987261147, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9638023376464844, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7398698925971985, + "num_tokens": 213153427.0, + "step": 8236 + }, + { + "epoch": 0.9045684164287283, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0876810550689697, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.6967822909355164, + "num_tokens": 213180649.0, + "step": 8237 + }, + { + "epoch": 0.904678234131342, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9329146146774292, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6943100094795227, + "num_tokens": 213212215.0, + "step": 8238 + }, + { + "epoch": 0.9047880518339556, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.000227689743042, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6902556419372559, + "num_tokens": 213242563.0, + "step": 8239 + }, + { + "epoch": 0.9048978695365693, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1359567642211914, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6886295080184937, + "num_tokens": 213271046.0, + "step": 8240 + }, + { + "epoch": 0.9050076872391829, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.9943639039993286, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7206436395645142, + "num_tokens": 213301255.0, + "step": 8241 + }, + { + "epoch": 0.9051175049417967, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.3136398792266846, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7283031940460205, + "num_tokens": 213324833.0, + "step": 8242 + }, + { + "epoch": 0.9052273226444103, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.066286325454712, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6952872276306152, + "num_tokens": 213351024.0, + "step": 8243 + }, + { + "epoch": 0.905337140347024, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0821473598480225, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6906077861785889, + "num_tokens": 213379797.0, + "step": 8244 + }, + { + "epoch": 0.9054469580496376, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0666615962982178, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6884902119636536, + "num_tokens": 213409556.0, + "step": 8245 + }, + { + "epoch": 0.9055567757522512, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.174135208129883, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7091907262802124, + "num_tokens": 213433519.0, + "step": 8246 + }, + { + "epoch": 0.9056665934548649, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.025857448577881, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6941779851913452, + "num_tokens": 213462154.0, + "step": 8247 + }, + { + "epoch": 0.9057764111574785, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2193925380706787, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7127034068107605, + "num_tokens": 213485521.0, + "step": 8248 + }, + { + "epoch": 0.9058862288600923, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.056729555130005, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7020959854125977, + "num_tokens": 213514699.0, + "step": 8249 + }, + { + "epoch": 0.9059960465627059, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.2274444103240967, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.6972694993019104, + "num_tokens": 213539877.0, + "step": 8250 + }, + { + "epoch": 0.9061058642653196, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.132683515548706, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6985271573066711, + "num_tokens": 213567267.0, + "step": 8251 + }, + { + "epoch": 0.9062156819679332, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1566553115844727, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7189704179763794, + "num_tokens": 213593174.0, + "step": 8252 + }, + { + "epoch": 0.9063254996705469, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.1597557067871094, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.7048825025558472, + "num_tokens": 213619039.0, + "step": 8253 + }, + { + "epoch": 0.9064353173731605, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.20249605178833, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.717833399772644, + "num_tokens": 213644412.0, + "step": 8254 + }, + { + "epoch": 0.9065451350757742, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.288673162460327, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7139123678207397, + "num_tokens": 213668556.0, + "step": 8255 + }, + { + "epoch": 0.9066549527783879, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 1.983598232269287, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7464736104011536, + "num_tokens": 213694420.0, + "step": 8256 + }, + { + "epoch": 0.9067647704810016, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.232211112976074, + "learning_rate": 1e-06, + "loss": 1.0203, + "mean_token_accuracy": 0.689910888671875, + "num_tokens": 213721246.0, + "step": 8257 + }, + { + "epoch": 0.9068745881836152, + "ewc_loss": 7.05718994140625e-05, + "grad_norm": 2.0915768146514893, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7173354625701904, + "num_tokens": 213747046.0, + "step": 8258 + }, + { + "epoch": 0.9069844058862289, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.298593282699585, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.7008413076400757, + "num_tokens": 213771105.0, + "step": 8259 + }, + { + "epoch": 0.9070942235888425, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.167785406112671, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7249952554702759, + "num_tokens": 213795878.0, + "step": 8260 + }, + { + "epoch": 0.9072040412914562, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0423600673675537, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6948255300521851, + "num_tokens": 213824083.0, + "step": 8261 + }, + { + "epoch": 0.9073138589940698, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0629749298095703, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.717889130115509, + "num_tokens": 213851743.0, + "step": 8262 + }, + { + "epoch": 0.9074236766966836, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.457239866256714, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7088702917098999, + "num_tokens": 213872726.0, + "step": 8263 + }, + { + "epoch": 0.9075334943992972, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9083393812179565, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.7005765438079834, + "num_tokens": 213904422.0, + "step": 8264 + }, + { + "epoch": 0.9076433121019108, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2378931045532227, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7101007699966431, + "num_tokens": 213929673.0, + "step": 8265 + }, + { + "epoch": 0.9077531298045245, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.246962547302246, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7161514163017273, + "num_tokens": 213952516.0, + "step": 8266 + }, + { + "epoch": 0.9078629475071381, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9012080430984497, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6939245462417603, + "num_tokens": 213982810.0, + "step": 8267 + }, + { + "epoch": 0.9079727652097518, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1348824501037598, + "learning_rate": 1e-06, + "loss": 1.0535, + "mean_token_accuracy": 0.6865918636322021, + "num_tokens": 214010916.0, + "step": 8268 + }, + { + "epoch": 0.9080825829123654, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2868454456329346, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7044815421104431, + "num_tokens": 214035488.0, + "step": 8269 + }, + { + "epoch": 0.9081924006149792, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9851404428482056, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7231054306030273, + "num_tokens": 214064658.0, + "step": 8270 + }, + { + "epoch": 0.9083022183175928, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1816506385803223, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7048194408416748, + "num_tokens": 214090493.0, + "step": 8271 + }, + { + "epoch": 0.9084120360202065, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.4736437797546387, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7235819697380066, + "num_tokens": 214109319.0, + "step": 8272 + }, + { + "epoch": 0.9085218537228201, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.122319221496582, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7070026397705078, + "num_tokens": 214135566.0, + "step": 8273 + }, + { + "epoch": 0.9086316714254338, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.259352207183838, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7175281047821045, + "num_tokens": 214158638.0, + "step": 8274 + }, + { + "epoch": 0.9087414891280474, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0390729904174805, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6968082189559937, + "num_tokens": 214186129.0, + "step": 8275 + }, + { + "epoch": 0.9088513068306611, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0179080963134766, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.6993201971054077, + "num_tokens": 214216102.0, + "step": 8276 + }, + { + "epoch": 0.9089611245332747, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.031191825866699, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.704322338104248, + "num_tokens": 214244267.0, + "step": 8277 + }, + { + "epoch": 0.9090709422358885, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.962951421737671, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6961996555328369, + "num_tokens": 214274438.0, + "step": 8278 + }, + { + "epoch": 0.9091807599385021, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.911943793296814, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6946972608566284, + "num_tokens": 214306660.0, + "step": 8279 + }, + { + "epoch": 0.9092905776411158, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.219499111175537, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7441418170928955, + "num_tokens": 214331101.0, + "step": 8280 + }, + { + "epoch": 0.9094003953437294, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1274759769439697, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7221652269363403, + "num_tokens": 214358743.0, + "step": 8281 + }, + { + "epoch": 0.909510213046343, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1366822719573975, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7005617618560791, + "num_tokens": 214386075.0, + "step": 8282 + }, + { + "epoch": 0.9096200307489567, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.6580586433410645, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.710114598274231, + "num_tokens": 214405293.0, + "step": 8283 + }, + { + "epoch": 0.9097298484515703, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0395760536193848, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7066479921340942, + "num_tokens": 214432685.0, + "step": 8284 + }, + { + "epoch": 0.9098396661541841, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1987361907958984, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7030402421951294, + "num_tokens": 214458723.0, + "step": 8285 + }, + { + "epoch": 0.9099494838567977, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.094299077987671, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7162044048309326, + "num_tokens": 214486020.0, + "step": 8286 + }, + { + "epoch": 0.9100593015594114, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.353814125061035, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7050029039382935, + "num_tokens": 214509254.0, + "step": 8287 + }, + { + "epoch": 0.910169119262025, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0307724475860596, + "learning_rate": 1e-06, + "loss": 1.044, + "mean_token_accuracy": 0.6889779567718506, + "num_tokens": 214540063.0, + "step": 8288 + }, + { + "epoch": 0.9102789369646387, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2490475177764893, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6980735063552856, + "num_tokens": 214565092.0, + "step": 8289 + }, + { + "epoch": 0.9103887546672523, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.8753336668014526, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7051154375076294, + "num_tokens": 214598113.0, + "step": 8290 + }, + { + "epoch": 0.910498572369866, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2502102851867676, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7102770209312439, + "num_tokens": 214621759.0, + "step": 8291 + }, + { + "epoch": 0.9106083900724797, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1848256587982178, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7219316959381104, + "num_tokens": 214646977.0, + "step": 8292 + }, + { + "epoch": 0.9107182077750934, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3311142921447754, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7361698150634766, + "num_tokens": 214667009.0, + "step": 8293 + }, + { + "epoch": 0.910828025477707, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.288482189178467, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6953586339950562, + "num_tokens": 214690203.0, + "step": 8294 + }, + { + "epoch": 0.9109378431803207, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.453493118286133, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7285721302032471, + "num_tokens": 214711257.0, + "step": 8295 + }, + { + "epoch": 0.9110476608829343, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9731392860412598, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7056360244750977, + "num_tokens": 214740590.0, + "step": 8296 + }, + { + "epoch": 0.911157478585548, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9606696367263794, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.715316653251648, + "num_tokens": 214772673.0, + "step": 8297 + }, + { + "epoch": 0.9112672962881616, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.271022319793701, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6973452568054199, + "num_tokens": 214796315.0, + "step": 8298 + }, + { + "epoch": 0.9113771139907754, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.254023790359497, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6863059997558594, + "num_tokens": 214822372.0, + "step": 8299 + }, + { + "epoch": 0.911486931693389, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9217808246612549, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6933345794677734, + "num_tokens": 214855213.0, + "step": 8300 + }, + { + "epoch": 0.9115967493960027, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.8490749597549438, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7044001817703247, + "num_tokens": 214890209.0, + "step": 8301 + }, + { + "epoch": 0.9117065670986163, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9398235082626343, + "learning_rate": 1e-06, + "loss": 1.06, + "mean_token_accuracy": 0.6865828037261963, + "num_tokens": 214924274.0, + "step": 8302 + }, + { + "epoch": 0.91181638480123, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.35089373588562, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.7004531025886536, + "num_tokens": 214946732.0, + "step": 8303 + }, + { + "epoch": 0.9119262025038436, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.546931743621826, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7331787347793579, + "num_tokens": 214965979.0, + "step": 8304 + }, + { + "epoch": 0.9120360202064572, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3586912155151367, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7041959166526794, + "num_tokens": 214989147.0, + "step": 8305 + }, + { + "epoch": 0.9121458379090709, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.336703300476074, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7149976491928101, + "num_tokens": 215012035.0, + "step": 8306 + }, + { + "epoch": 0.9122556556116846, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.251251220703125, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7241166830062866, + "num_tokens": 215036677.0, + "step": 8307 + }, + { + "epoch": 0.9123654733142983, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3396480083465576, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7268210649490356, + "num_tokens": 215057133.0, + "step": 8308 + }, + { + "epoch": 0.9124752910169119, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1025123596191406, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.6970138549804688, + "num_tokens": 215082831.0, + "step": 8309 + }, + { + "epoch": 0.9125851087195256, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.195525884628296, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7062028646469116, + "num_tokens": 215106983.0, + "step": 8310 + }, + { + "epoch": 0.9126949264221392, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.170567274093628, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7215763330459595, + "num_tokens": 215131950.0, + "step": 8311 + }, + { + "epoch": 0.9128047441247529, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.383310317993164, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7080759406089783, + "num_tokens": 215153394.0, + "step": 8312 + }, + { + "epoch": 0.9129145618273665, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0242178440093994, + "learning_rate": 1e-06, + "loss": 1.0504, + "mean_token_accuracy": 0.6785432696342468, + "num_tokens": 215184587.0, + "step": 8313 + }, + { + "epoch": 0.9130243795299803, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1373672485351562, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6996909379959106, + "num_tokens": 215212894.0, + "step": 8314 + }, + { + "epoch": 0.9131341972325939, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1662826538085938, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.6971166729927063, + "num_tokens": 215240789.0, + "step": 8315 + }, + { + "epoch": 0.9132440149352076, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.280586004257202, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6875687837600708, + "num_tokens": 215265364.0, + "step": 8316 + }, + { + "epoch": 0.9133538326378212, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.5041072368621826, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7084652781486511, + "num_tokens": 215285525.0, + "step": 8317 + }, + { + "epoch": 0.9134636503404349, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.248621702194214, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7208309173583984, + "num_tokens": 215308355.0, + "step": 8318 + }, + { + "epoch": 0.9135734680430485, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3840928077697754, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6972498893737793, + "num_tokens": 215336239.0, + "step": 8319 + }, + { + "epoch": 0.9136832857456622, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.388453960418701, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7077836990356445, + "num_tokens": 215360536.0, + "step": 8320 + }, + { + "epoch": 0.9137931034482759, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2057504653930664, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7110397815704346, + "num_tokens": 215386149.0, + "step": 8321 + }, + { + "epoch": 0.9139029211508896, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.368623971939087, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7219593524932861, + "num_tokens": 215407934.0, + "step": 8322 + }, + { + "epoch": 0.9140127388535032, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2101657390594482, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7292228937149048, + "num_tokens": 215435377.0, + "step": 8323 + }, + { + "epoch": 0.9141225565561168, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3216874599456787, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6986693143844604, + "num_tokens": 215461261.0, + "step": 8324 + }, + { + "epoch": 0.9142323742587305, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.3239991664886475, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7244566082954407, + "num_tokens": 215483834.0, + "step": 8325 + }, + { + "epoch": 0.9143421919613441, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1815712451934814, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6977943181991577, + "num_tokens": 215508844.0, + "step": 8326 + }, + { + "epoch": 0.9144520096639578, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9965893030166626, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7310029864311218, + "num_tokens": 215536103.0, + "step": 8327 + }, + { + "epoch": 0.9145618273665715, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.5000967979431152, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7294377088546753, + "num_tokens": 215555919.0, + "step": 8328 + }, + { + "epoch": 0.9146716450691852, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2910966873168945, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7074853181838989, + "num_tokens": 215580581.0, + "step": 8329 + }, + { + "epoch": 0.9147814627717988, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.101625442504883, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7243374586105347, + "num_tokens": 215605680.0, + "step": 8330 + }, + { + "epoch": 0.9148912804744125, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0091018676757812, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6932539939880371, + "num_tokens": 215634982.0, + "step": 8331 + }, + { + "epoch": 0.9150010981770261, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 1.9408791065216064, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6933900117874146, + "num_tokens": 215667249.0, + "step": 8332 + }, + { + "epoch": 0.9151109158796398, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2687795162200928, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.723273754119873, + "num_tokens": 215690494.0, + "step": 8333 + }, + { + "epoch": 0.9152207335822534, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1257944107055664, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6983270049095154, + "num_tokens": 215716541.0, + "step": 8334 + }, + { + "epoch": 0.9153305512848671, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.208226203918457, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7076995372772217, + "num_tokens": 215741018.0, + "step": 8335 + }, + { + "epoch": 0.9154403689874808, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.2857813835144043, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7153940200805664, + "num_tokens": 215765566.0, + "step": 8336 + }, + { + "epoch": 0.9155501866900945, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.2918975353240967, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7199540138244629, + "num_tokens": 215789434.0, + "step": 8337 + }, + { + "epoch": 0.9156600043927081, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.0587360858917236, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7169933319091797, + "num_tokens": 215816457.0, + "step": 8338 + }, + { + "epoch": 0.9157698220953218, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.332700252532959, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7255074977874756, + "num_tokens": 215839046.0, + "step": 8339 + }, + { + "epoch": 0.9158796397979354, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.200071334838867, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.726862907409668, + "num_tokens": 215861389.0, + "step": 8340 + }, + { + "epoch": 0.915989457500549, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.439265489578247, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7183834910392761, + "num_tokens": 215881662.0, + "step": 8341 + }, + { + "epoch": 0.9160992752031627, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.4519553184509277, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7253950238227844, + "num_tokens": 215901617.0, + "step": 8342 + }, + { + "epoch": 0.9162090929057765, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.188001871109009, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7152339220046997, + "num_tokens": 215927452.0, + "step": 8343 + }, + { + "epoch": 0.9163189106083901, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.2213051319122314, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7154916524887085, + "num_tokens": 215953475.0, + "step": 8344 + }, + { + "epoch": 0.9164287283110037, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.33414363861084, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7033425569534302, + "num_tokens": 215978069.0, + "step": 8345 + }, + { + "epoch": 0.9165385460136174, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.075702667236328, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7171316742897034, + "num_tokens": 216005963.0, + "step": 8346 + }, + { + "epoch": 0.916648363716231, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1324360370635986, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7162829041481018, + "num_tokens": 216032217.0, + "step": 8347 + }, + { + "epoch": 0.9167581814188447, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0669403076171875, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.682138204574585, + "num_tokens": 216061479.0, + "step": 8348 + }, + { + "epoch": 0.9168679991214583, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.046251058578491, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7056148052215576, + "num_tokens": 216090079.0, + "step": 8349 + }, + { + "epoch": 0.9169778168240721, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.1738593578338623, + "learning_rate": 1e-06, + "loss": 1.034, + "mean_token_accuracy": 0.6877315044403076, + "num_tokens": 216117217.0, + "step": 8350 + }, + { + "epoch": 0.9170876345266857, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.0706286430358887, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7252781987190247, + "num_tokens": 216144407.0, + "step": 8351 + }, + { + "epoch": 0.9171974522292994, + "ewc_loss": 7.104873657226562e-05, + "grad_norm": 2.281163215637207, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.72612065076828, + "num_tokens": 216168373.0, + "step": 8352 + }, + { + "epoch": 0.917307269931913, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.3373398780822754, + "learning_rate": 1e-06, + "loss": 1.0769, + "mean_token_accuracy": 0.6709868907928467, + "num_tokens": 216193251.0, + "step": 8353 + }, + { + "epoch": 0.9174170876345267, + "ewc_loss": 7.152557373046875e-05, + "grad_norm": 2.4646239280700684, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7263373136520386, + "num_tokens": 216213242.0, + "step": 8354 + }, + { + "epoch": 0.9175269053371403, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.025786876678467, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7172212600708008, + "num_tokens": 216240426.0, + "step": 8355 + }, + { + "epoch": 0.917636723039754, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.989911437034607, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6978745460510254, + "num_tokens": 216270015.0, + "step": 8356 + }, + { + "epoch": 0.9177465407423677, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.980433702468872, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7048083543777466, + "num_tokens": 216300090.0, + "step": 8357 + }, + { + "epoch": 0.9178563584449814, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9543042182922363, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7113610506057739, + "num_tokens": 216330103.0, + "step": 8358 + }, + { + "epoch": 0.917966176147595, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0643484592437744, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7197399139404297, + "num_tokens": 216356619.0, + "step": 8359 + }, + { + "epoch": 0.9180759938502087, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3518946170806885, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7086161375045776, + "num_tokens": 216379171.0, + "step": 8360 + }, + { + "epoch": 0.9181858115528223, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9905868768692017, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.6946181654930115, + "num_tokens": 216409387.0, + "step": 8361 + }, + { + "epoch": 0.918295629255436, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.233170986175537, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7166155576705933, + "num_tokens": 216433635.0, + "step": 8362 + }, + { + "epoch": 0.9184054469580496, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9248408079147339, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6980802416801453, + "num_tokens": 216466485.0, + "step": 8363 + }, + { + "epoch": 0.9185152646606632, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.24515438079834, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7165592908859253, + "num_tokens": 216488907.0, + "step": 8364 + }, + { + "epoch": 0.918625082363277, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9056856632232666, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6992838978767395, + "num_tokens": 216521259.0, + "step": 8365 + }, + { + "epoch": 0.9187349000658906, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.6127710342407227, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7290324568748474, + "num_tokens": 216538864.0, + "step": 8366 + }, + { + "epoch": 0.9188447177685043, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0361545085906982, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7082920074462891, + "num_tokens": 216568362.0, + "step": 8367 + }, + { + "epoch": 0.9189545354711179, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9476044178009033, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7054189443588257, + "num_tokens": 216598722.0, + "step": 8368 + }, + { + "epoch": 0.9190643531737316, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.233053684234619, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6978030204772949, + "num_tokens": 216623241.0, + "step": 8369 + }, + { + "epoch": 0.9191741708763452, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0508904457092285, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7239203453063965, + "num_tokens": 216650065.0, + "step": 8370 + }, + { + "epoch": 0.9192839885789589, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.209907054901123, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6899474859237671, + "num_tokens": 216675984.0, + "step": 8371 + }, + { + "epoch": 0.9193938062815726, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9740452766418457, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.7046854496002197, + "num_tokens": 216706335.0, + "step": 8372 + }, + { + "epoch": 0.9195036239841863, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.914993405342102, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6857573390007019, + "num_tokens": 216738678.0, + "step": 8373 + }, + { + "epoch": 0.9196134416867999, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9739829301834106, + "learning_rate": 1e-06, + "loss": 0.7598, + "mean_token_accuracy": 0.765208899974823, + "num_tokens": 216764265.0, + "step": 8374 + }, + { + "epoch": 0.9197232593894136, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2971696853637695, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7297313213348389, + "num_tokens": 216787019.0, + "step": 8375 + }, + { + "epoch": 0.9198330770920272, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1971683502197266, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7066761255264282, + "num_tokens": 216810780.0, + "step": 8376 + }, + { + "epoch": 0.9199428947946409, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3889620304107666, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6925002336502075, + "num_tokens": 216833727.0, + "step": 8377 + }, + { + "epoch": 0.9200527124972545, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.195493698120117, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6877306699752808, + "num_tokens": 216860498.0, + "step": 8378 + }, + { + "epoch": 0.9201625301998683, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.623286008834839, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7237303256988525, + "num_tokens": 216880677.0, + "step": 8379 + }, + { + "epoch": 0.9202723479024819, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2543675899505615, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7135733366012573, + "num_tokens": 216903748.0, + "step": 8380 + }, + { + "epoch": 0.9203821656050956, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.123811721801758, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6910303831100464, + "num_tokens": 216935680.0, + "step": 8381 + }, + { + "epoch": 0.9204919833077092, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1775591373443604, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7081864476203918, + "num_tokens": 216960966.0, + "step": 8382 + }, + { + "epoch": 0.9206018010103229, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.225667715072632, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7156319618225098, + "num_tokens": 216984655.0, + "step": 8383 + }, + { + "epoch": 0.9207116187129365, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.962110161781311, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7279993891716003, + "num_tokens": 217013859.0, + "step": 8384 + }, + { + "epoch": 0.9208214364155501, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3202505111694336, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6923982501029968, + "num_tokens": 217039191.0, + "step": 8385 + }, + { + "epoch": 0.9209312541181639, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.192890167236328, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7054848670959473, + "num_tokens": 217065249.0, + "step": 8386 + }, + { + "epoch": 0.9210410718207775, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.174894094467163, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6984127163887024, + "num_tokens": 217092553.0, + "step": 8387 + }, + { + "epoch": 0.9211508895233912, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.202272415161133, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7123181819915771, + "num_tokens": 217117661.0, + "step": 8388 + }, + { + "epoch": 0.9212607072260048, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1253561973571777, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6822812557220459, + "num_tokens": 217146021.0, + "step": 8389 + }, + { + "epoch": 0.9213705249286185, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2234017848968506, + "learning_rate": 1e-06, + "loss": 1.0328, + "mean_token_accuracy": 0.6981163620948792, + "num_tokens": 217171003.0, + "step": 8390 + }, + { + "epoch": 0.9214803426312321, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.184514045715332, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7010204195976257, + "num_tokens": 217194716.0, + "step": 8391 + }, + { + "epoch": 0.9215901603338458, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1812515258789062, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6944674253463745, + "num_tokens": 217220352.0, + "step": 8392 + }, + { + "epoch": 0.9216999780364594, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.258110761642456, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7290254831314087, + "num_tokens": 217241491.0, + "step": 8393 + }, + { + "epoch": 0.9218097957390732, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3102481365203857, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7248984575271606, + "num_tokens": 217261562.0, + "step": 8394 + }, + { + "epoch": 0.9219196134416868, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8555941581726074, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7097381353378296, + "num_tokens": 217294792.0, + "step": 8395 + }, + { + "epoch": 0.9220294311443005, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0429751873016357, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7118347883224487, + "num_tokens": 217322645.0, + "step": 8396 + }, + { + "epoch": 0.9221392488469141, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8542380332946777, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.7024194002151489, + "num_tokens": 217357087.0, + "step": 8397 + }, + { + "epoch": 0.9222490665495278, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2554194927215576, + "learning_rate": 1e-06, + "loss": 1.1003, + "mean_token_accuracy": 0.673121988773346, + "num_tokens": 217383181.0, + "step": 8398 + }, + { + "epoch": 0.9223588842521414, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.6731934547424316, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7143256664276123, + "num_tokens": 217401323.0, + "step": 8399 + }, + { + "epoch": 0.922468701954755, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9938907623291016, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7023543119430542, + "num_tokens": 217430911.0, + "step": 8400 + }, + { + "epoch": 0.9225785196573688, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.989926815032959, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6928113698959351, + "num_tokens": 217460849.0, + "step": 8401 + }, + { + "epoch": 0.9226883373599825, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.103405475616455, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7173957824707031, + "num_tokens": 217487254.0, + "step": 8402 + }, + { + "epoch": 0.9227981550625961, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1162235736846924, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.6766855716705322, + "num_tokens": 217515029.0, + "step": 8403 + }, + { + "epoch": 0.9229079727652097, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.21585750579834, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7007594108581543, + "num_tokens": 217539472.0, + "step": 8404 + }, + { + "epoch": 0.9230177904678234, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.139070510864258, + "learning_rate": 1e-06, + "loss": 1.0833, + "mean_token_accuracy": 0.6970628499984741, + "num_tokens": 217565277.0, + "step": 8405 + }, + { + "epoch": 0.923127608170437, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.132783889770508, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7027257680892944, + "num_tokens": 217591244.0, + "step": 8406 + }, + { + "epoch": 0.9232374258730507, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0306801795959473, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7070969343185425, + "num_tokens": 217619176.0, + "step": 8407 + }, + { + "epoch": 0.9233472435756644, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9838608503341675, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7220135927200317, + "num_tokens": 217646909.0, + "step": 8408 + }, + { + "epoch": 0.9234570612782781, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.4515342712402344, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7112847566604614, + "num_tokens": 217667989.0, + "step": 8409 + }, + { + "epoch": 0.9235668789808917, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0581791400909424, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7239340543746948, + "num_tokens": 217694095.0, + "step": 8410 + }, + { + "epoch": 0.9236766966835054, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.045698404312134, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.6848466396331787, + "num_tokens": 217723548.0, + "step": 8411 + }, + { + "epoch": 0.923786514386119, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.310741662979126, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7086390256881714, + "num_tokens": 217747359.0, + "step": 8412 + }, + { + "epoch": 0.9238963320887327, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2424142360687256, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6998311281204224, + "num_tokens": 217772721.0, + "step": 8413 + }, + { + "epoch": 0.9240061497913463, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2398629188537598, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6919155120849609, + "num_tokens": 217797517.0, + "step": 8414 + }, + { + "epoch": 0.9241159674939601, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.015519857406616, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6933271288871765, + "num_tokens": 217826190.0, + "step": 8415 + }, + { + "epoch": 0.9242257851965737, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.256251811981201, + "learning_rate": 1e-06, + "loss": 1.0292, + "mean_token_accuracy": 0.6864655017852783, + "num_tokens": 217852227.0, + "step": 8416 + }, + { + "epoch": 0.9243356028991874, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0907676219940186, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7032226324081421, + "num_tokens": 217880720.0, + "step": 8417 + }, + { + "epoch": 0.924445420601801, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.407407760620117, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7090909481048584, + "num_tokens": 217901861.0, + "step": 8418 + }, + { + "epoch": 0.9245552383044147, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.034092903137207, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6957274675369263, + "num_tokens": 217931798.0, + "step": 8419 + }, + { + "epoch": 0.9246650560070283, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.154939889907837, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6914072036743164, + "num_tokens": 217959279.0, + "step": 8420 + }, + { + "epoch": 0.924774873709642, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.175394058227539, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7063108086585999, + "num_tokens": 217984120.0, + "step": 8421 + }, + { + "epoch": 0.9248846914122557, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0999596118927, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7298390865325928, + "num_tokens": 218010085.0, + "step": 8422 + }, + { + "epoch": 0.9249945091148694, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9303274154663086, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7045578956604004, + "num_tokens": 218042423.0, + "step": 8423 + }, + { + "epoch": 0.925104326817483, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.279902935028076, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7000045776367188, + "num_tokens": 218067588.0, + "step": 8424 + }, + { + "epoch": 0.9252141445200966, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9069961309432983, + "learning_rate": 1e-06, + "loss": 1.1335, + "mean_token_accuracy": 0.6637966632843018, + "num_tokens": 218101759.0, + "step": 8425 + }, + { + "epoch": 0.9253239622227103, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.123150587081909, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.702788233757019, + "num_tokens": 218129303.0, + "step": 8426 + }, + { + "epoch": 0.9254337799253239, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.272963523864746, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7020807862281799, + "num_tokens": 218154835.0, + "step": 8427 + }, + { + "epoch": 0.9255435976279376, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.061960220336914, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6839653849601746, + "num_tokens": 218183636.0, + "step": 8428 + }, + { + "epoch": 0.9256534153305512, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3816111087799072, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7135874032974243, + "num_tokens": 218207019.0, + "step": 8429 + }, + { + "epoch": 0.925763233033165, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.257766008377075, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7189612984657288, + "num_tokens": 218229583.0, + "step": 8430 + }, + { + "epoch": 0.9258730507357786, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.156801223754883, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7037237882614136, + "num_tokens": 218256194.0, + "step": 8431 + }, + { + "epoch": 0.9259828684383923, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3445045948028564, + "learning_rate": 1e-06, + "loss": 1.0252, + "mean_token_accuracy": 0.685064435005188, + "num_tokens": 218280046.0, + "step": 8432 + }, + { + "epoch": 0.9260926861410059, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.4944329261779785, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.717565655708313, + "num_tokens": 218299251.0, + "step": 8433 + }, + { + "epoch": 0.9262025038436196, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.259169340133667, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7139817476272583, + "num_tokens": 218323170.0, + "step": 8434 + }, + { + "epoch": 0.9263123215462332, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.314162492752075, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7128739953041077, + "num_tokens": 218346012.0, + "step": 8435 + }, + { + "epoch": 0.9264221392488469, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3765385150909424, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7202320098876953, + "num_tokens": 218368155.0, + "step": 8436 + }, + { + "epoch": 0.9265319569514606, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.028355360031128, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.7002831101417542, + "num_tokens": 218394595.0, + "step": 8437 + }, + { + "epoch": 0.9266417746540743, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2671263217926025, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7069387435913086, + "num_tokens": 218419929.0, + "step": 8438 + }, + { + "epoch": 0.9267515923566879, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3350160121917725, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7366012334823608, + "num_tokens": 218441566.0, + "step": 8439 + }, + { + "epoch": 0.9268614100593016, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.423727512359619, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7193409204483032, + "num_tokens": 218464451.0, + "step": 8440 + }, + { + "epoch": 0.9269712277619152, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0933430194854736, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7023906707763672, + "num_tokens": 218492548.0, + "step": 8441 + }, + { + "epoch": 0.9270810454645289, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.227553129196167, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7100573778152466, + "num_tokens": 218519220.0, + "step": 8442 + }, + { + "epoch": 0.9271908631671425, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.383826971054077, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.6999797821044922, + "num_tokens": 218541685.0, + "step": 8443 + }, + { + "epoch": 0.9273006808697563, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2731943130493164, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7262083888053894, + "num_tokens": 218565855.0, + "step": 8444 + }, + { + "epoch": 0.9274104985723699, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0968427658081055, + "learning_rate": 1e-06, + "loss": 1.0441, + "mean_token_accuracy": 0.6938934922218323, + "num_tokens": 218592633.0, + "step": 8445 + }, + { + "epoch": 0.9275203162749835, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.095886707305908, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7086433172225952, + "num_tokens": 218618417.0, + "step": 8446 + }, + { + "epoch": 0.9276301339775972, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2206575870513916, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7204631567001343, + "num_tokens": 218641999.0, + "step": 8447 + }, + { + "epoch": 0.9277399516802108, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.082906484603882, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.6859505772590637, + "num_tokens": 218670529.0, + "step": 8448 + }, + { + "epoch": 0.9278497693828245, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3877182006835938, + "learning_rate": 1e-06, + "loss": 1.076, + "mean_token_accuracy": 0.6849483251571655, + "num_tokens": 218695484.0, + "step": 8449 + }, + { + "epoch": 0.9279595870854381, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9833697080612183, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.7063775062561035, + "num_tokens": 218727725.0, + "step": 8450 + }, + { + "epoch": 0.9280694047880519, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.036417007446289, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7224045991897583, + "num_tokens": 218754505.0, + "step": 8451 + }, + { + "epoch": 0.9281792224906655, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.191607713699341, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7082715630531311, + "num_tokens": 218778201.0, + "step": 8452 + }, + { + "epoch": 0.9282890401932792, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.273134708404541, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7190104126930237, + "num_tokens": 218801071.0, + "step": 8453 + }, + { + "epoch": 0.9283988578958928, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.365565299987793, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7037118077278137, + "num_tokens": 218821290.0, + "step": 8454 + }, + { + "epoch": 0.9285086755985065, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0540342330932617, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7220049500465393, + "num_tokens": 218847790.0, + "step": 8455 + }, + { + "epoch": 0.9286184933011201, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.161867380142212, + "learning_rate": 1e-06, + "loss": 1.0385, + "mean_token_accuracy": 0.6812225580215454, + "num_tokens": 218872798.0, + "step": 8456 + }, + { + "epoch": 0.9287283110037338, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2291812896728516, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6963378190994263, + "num_tokens": 218896411.0, + "step": 8457 + }, + { + "epoch": 0.9288381287063474, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.256166696548462, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.701582133769989, + "num_tokens": 218919199.0, + "step": 8458 + }, + { + "epoch": 0.9289479464089612, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.121110200881958, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7074986100196838, + "num_tokens": 218947282.0, + "step": 8459 + }, + { + "epoch": 0.9290577641115748, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.220266819000244, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7068161964416504, + "num_tokens": 218970788.0, + "step": 8460 + }, + { + "epoch": 0.9291675818141885, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0006017684936523, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7177338600158691, + "num_tokens": 219001070.0, + "step": 8461 + }, + { + "epoch": 0.9292773995168021, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2078471183776855, + "learning_rate": 1e-06, + "loss": 1.0511, + "mean_token_accuracy": 0.6786963939666748, + "num_tokens": 219029358.0, + "step": 8462 + }, + { + "epoch": 0.9293872172194158, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1141326427459717, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7287287712097168, + "num_tokens": 219055155.0, + "step": 8463 + }, + { + "epoch": 0.9294970349220294, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8961166143417358, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.688705563545227, + "num_tokens": 219088784.0, + "step": 8464 + }, + { + "epoch": 0.929606852624643, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9046589136123657, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7228443026542664, + "num_tokens": 219119790.0, + "step": 8465 + }, + { + "epoch": 0.9297166703272568, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2094922065734863, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7123904824256897, + "num_tokens": 219142801.0, + "step": 8466 + }, + { + "epoch": 0.9298264880298704, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1320321559906006, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7320823669433594, + "num_tokens": 219168134.0, + "step": 8467 + }, + { + "epoch": 0.9299363057324841, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3001208305358887, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7316434383392334, + "num_tokens": 219191748.0, + "step": 8468 + }, + { + "epoch": 0.9300461234350977, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.4838757514953613, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6982288360595703, + "num_tokens": 219214834.0, + "step": 8469 + }, + { + "epoch": 0.9301559411377114, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1567459106445312, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7151494026184082, + "num_tokens": 219243265.0, + "step": 8470 + }, + { + "epoch": 0.930265758840325, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8306751251220703, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7173691391944885, + "num_tokens": 219275706.0, + "step": 8471 + }, + { + "epoch": 0.9303755765429387, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.038494348526001, + "learning_rate": 1e-06, + "loss": 1.0648, + "mean_token_accuracy": 0.6735840439796448, + "num_tokens": 219305391.0, + "step": 8472 + }, + { + "epoch": 0.9304853942455524, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1975598335266113, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6987919807434082, + "num_tokens": 219331399.0, + "step": 8473 + }, + { + "epoch": 0.9305952119481661, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0443716049194336, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7297248244285583, + "num_tokens": 219358967.0, + "step": 8474 + }, + { + "epoch": 0.9307050296507797, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8258898258209229, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7017574310302734, + "num_tokens": 219394839.0, + "step": 8475 + }, + { + "epoch": 0.9308148473533934, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.6028032302856445, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7191207408905029, + "num_tokens": 219413954.0, + "step": 8476 + }, + { + "epoch": 0.930924665056007, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1242611408233643, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7044399976730347, + "num_tokens": 219440917.0, + "step": 8477 + }, + { + "epoch": 0.9310344827586207, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.132704973220825, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6875566840171814, + "num_tokens": 219469897.0, + "step": 8478 + }, + { + "epoch": 0.9311443004612343, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.293492317199707, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7122279405593872, + "num_tokens": 219491251.0, + "step": 8479 + }, + { + "epoch": 0.9312541181638481, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.354565382003784, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7079472541809082, + "num_tokens": 219512466.0, + "step": 8480 + }, + { + "epoch": 0.9313639358664617, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1489245891571045, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7079894542694092, + "num_tokens": 219539230.0, + "step": 8481 + }, + { + "epoch": 0.9314737535690754, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.076345443725586, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7132962346076965, + "num_tokens": 219567470.0, + "step": 8482 + }, + { + "epoch": 0.931583571271689, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0425875186920166, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7125405073165894, + "num_tokens": 219595741.0, + "step": 8483 + }, + { + "epoch": 0.9316933889743026, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.197765827178955, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7240603566169739, + "num_tokens": 219620741.0, + "step": 8484 + }, + { + "epoch": 0.9318032066769163, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.142442464828491, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7024272680282593, + "num_tokens": 219648414.0, + "step": 8485 + }, + { + "epoch": 0.9319130243795299, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0499744415283203, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6964570879936218, + "num_tokens": 219675933.0, + "step": 8486 + }, + { + "epoch": 0.9320228420821436, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.162888526916504, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7002646327018738, + "num_tokens": 219702289.0, + "step": 8487 + }, + { + "epoch": 0.9321326597847573, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.298734426498413, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7004004120826721, + "num_tokens": 219724969.0, + "step": 8488 + }, + { + "epoch": 0.932242477487371, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.4582231044769287, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7266441583633423, + "num_tokens": 219745247.0, + "step": 8489 + }, + { + "epoch": 0.9323522951899846, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3942999839782715, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7204148769378662, + "num_tokens": 219767121.0, + "step": 8490 + }, + { + "epoch": 0.9324621128925983, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1026530265808105, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7071885466575623, + "num_tokens": 219793110.0, + "step": 8491 + }, + { + "epoch": 0.9325719305952119, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.18091082572937, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7193238139152527, + "num_tokens": 219817431.0, + "step": 8492 + }, + { + "epoch": 0.9326817482978256, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.272120714187622, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7128472328186035, + "num_tokens": 219840958.0, + "step": 8493 + }, + { + "epoch": 0.9327915660004392, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.365323066711426, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7044664025306702, + "num_tokens": 219864220.0, + "step": 8494 + }, + { + "epoch": 0.932901383703053, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.318852186203003, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.727789580821991, + "num_tokens": 219885350.0, + "step": 8495 + }, + { + "epoch": 0.9330112014056666, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1574959754943848, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7061617374420166, + "num_tokens": 219910183.0, + "step": 8496 + }, + { + "epoch": 0.9331210191082803, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3061835765838623, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7140491008758545, + "num_tokens": 219935143.0, + "step": 8497 + }, + { + "epoch": 0.9332308368108939, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1265528202056885, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6889164447784424, + "num_tokens": 219962271.0, + "step": 8498 + }, + { + "epoch": 0.9333406545135076, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9522207975387573, + "learning_rate": 1e-06, + "loss": 1.0558, + "mean_token_accuracy": 0.6860082745552063, + "num_tokens": 219996977.0, + "step": 8499 + }, + { + "epoch": 0.9334504722161212, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3422858715057373, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.709345817565918, + "num_tokens": 220019017.0, + "step": 8500 + }, + { + "epoch": 0.9335602899187349, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0786843299865723, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.721670925617218, + "num_tokens": 220044241.0, + "step": 8501 + }, + { + "epoch": 0.9336701076213486, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.453298330307007, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6981202960014343, + "num_tokens": 220066675.0, + "step": 8502 + }, + { + "epoch": 0.9337799253239623, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.4235100746154785, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7225770950317383, + "num_tokens": 220087579.0, + "step": 8503 + }, + { + "epoch": 0.9338897430265759, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.670849561691284, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7135793566703796, + "num_tokens": 220105762.0, + "step": 8504 + }, + { + "epoch": 0.9339995607291895, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.371166944503784, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7113724946975708, + "num_tokens": 220127377.0, + "step": 8505 + }, + { + "epoch": 0.9341093784318032, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.9979088306427002, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7182043790817261, + "num_tokens": 220156035.0, + "step": 8506 + }, + { + "epoch": 0.9342191961344168, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.975506067276001, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7110825181007385, + "num_tokens": 220186364.0, + "step": 8507 + }, + { + "epoch": 0.9343290138370305, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0127408504486084, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.6911374926567078, + "num_tokens": 220216923.0, + "step": 8508 + }, + { + "epoch": 0.9344388315396442, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.117452621459961, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.6844103336334229, + "num_tokens": 220243427.0, + "step": 8509 + }, + { + "epoch": 0.9345486492422579, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.267564296722412, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.6824349164962769, + "num_tokens": 220270601.0, + "step": 8510 + }, + { + "epoch": 0.9346584669448715, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1436409950256348, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.7019819021224976, + "num_tokens": 220299537.0, + "step": 8511 + }, + { + "epoch": 0.9347682846474852, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3308539390563965, + "learning_rate": 1e-06, + "loss": 1.0667, + "mean_token_accuracy": 0.679267168045044, + "num_tokens": 220325252.0, + "step": 8512 + }, + { + "epoch": 0.9348781023500988, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0131118297576904, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.712938666343689, + "num_tokens": 220351982.0, + "step": 8513 + }, + { + "epoch": 0.9349879200527125, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.971835970878601, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7053829431533813, + "num_tokens": 220381152.0, + "step": 8514 + }, + { + "epoch": 0.9350977377553261, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.5454509258270264, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7324997186660767, + "num_tokens": 220399949.0, + "step": 8515 + }, + { + "epoch": 0.9352075554579398, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 1.8941550254821777, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7108476758003235, + "num_tokens": 220432231.0, + "step": 8516 + }, + { + "epoch": 0.9353173731605535, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.2524607181549072, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7186543941497803, + "num_tokens": 220455363.0, + "step": 8517 + }, + { + "epoch": 0.9354271908631672, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.0042061805725098, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6965004205703735, + "num_tokens": 220485388.0, + "step": 8518 + }, + { + "epoch": 0.9355370085657808, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.14908766746521, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7003263831138611, + "num_tokens": 220510686.0, + "step": 8519 + }, + { + "epoch": 0.9356468262683945, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.1398255825042725, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7002270221710205, + "num_tokens": 220538470.0, + "step": 8520 + }, + { + "epoch": 0.9357566439710081, + "ewc_loss": 7.200241088867188e-05, + "grad_norm": 2.3914294242858887, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7292572855949402, + "num_tokens": 220559701.0, + "step": 8521 + }, + { + "epoch": 0.9358664616736218, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.06632661819458, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7285001277923584, + "num_tokens": 220587287.0, + "step": 8522 + }, + { + "epoch": 0.9359762793762354, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2127726078033447, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.7047814726829529, + "num_tokens": 220612885.0, + "step": 8523 + }, + { + "epoch": 0.9360860970788492, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0178985595703125, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6852507591247559, + "num_tokens": 220642379.0, + "step": 8524 + }, + { + "epoch": 0.9361959147814628, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0326662063598633, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7007787227630615, + "num_tokens": 220670345.0, + "step": 8525 + }, + { + "epoch": 0.9363057324840764, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2566347122192383, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6881053447723389, + "num_tokens": 220695748.0, + "step": 8526 + }, + { + "epoch": 0.9364155501866901, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0644233226776123, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.6980799436569214, + "num_tokens": 220725066.0, + "step": 8527 + }, + { + "epoch": 0.9365253678893037, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2884984016418457, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.6912509202957153, + "num_tokens": 220750281.0, + "step": 8528 + }, + { + "epoch": 0.9366351855919174, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.161283016204834, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7078261375427246, + "num_tokens": 220778300.0, + "step": 8529 + }, + { + "epoch": 0.936745003294531, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.170959234237671, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.708440899848938, + "num_tokens": 220803943.0, + "step": 8530 + }, + { + "epoch": 0.9368548209971448, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1401960849761963, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7236935496330261, + "num_tokens": 220827576.0, + "step": 8531 + }, + { + "epoch": 0.9369646386997584, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.535486936569214, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6969450116157532, + "num_tokens": 220849305.0, + "step": 8532 + }, + { + "epoch": 0.9370744564023721, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2422118186950684, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6924505233764648, + "num_tokens": 220875563.0, + "step": 8533 + }, + { + "epoch": 0.9371842741049857, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.261016607284546, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7113872766494751, + "num_tokens": 220898919.0, + "step": 8534 + }, + { + "epoch": 0.9372940918075994, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1226584911346436, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6951831579208374, + "num_tokens": 220925388.0, + "step": 8535 + }, + { + "epoch": 0.937403909510213, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4062652587890625, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7082138061523438, + "num_tokens": 220946587.0, + "step": 8536 + }, + { + "epoch": 0.9375137272128267, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2873854637145996, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.6984373927116394, + "num_tokens": 220970097.0, + "step": 8537 + }, + { + "epoch": 0.9376235449154404, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0010581016540527, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6963436603546143, + "num_tokens": 220999343.0, + "step": 8538 + }, + { + "epoch": 0.9377333626180541, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.054487705230713, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7061585187911987, + "num_tokens": 221027252.0, + "step": 8539 + }, + { + "epoch": 0.9378431803206677, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.343337297439575, + "learning_rate": 1e-06, + "loss": 1.0499, + "mean_token_accuracy": 0.702273964881897, + "num_tokens": 221052625.0, + "step": 8540 + }, + { + "epoch": 0.9379529980232814, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2867166996002197, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7074090242385864, + "num_tokens": 221076412.0, + "step": 8541 + }, + { + "epoch": 0.938062815725895, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2236435413360596, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7239564061164856, + "num_tokens": 221104149.0, + "step": 8542 + }, + { + "epoch": 0.9381726334285087, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.020291805267334, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7020115852355957, + "num_tokens": 221132434.0, + "step": 8543 + }, + { + "epoch": 0.9382824511311223, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1776626110076904, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7068607807159424, + "num_tokens": 221157135.0, + "step": 8544 + }, + { + "epoch": 0.9383922688337359, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.3663125038146973, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.723755955696106, + "num_tokens": 221177932.0, + "step": 8545 + }, + { + "epoch": 0.9385020865363497, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.3606956005096436, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7130298018455505, + "num_tokens": 221200677.0, + "step": 8546 + }, + { + "epoch": 0.9386119042389633, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.064868927001953, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7119987607002258, + "num_tokens": 221226510.0, + "step": 8547 + }, + { + "epoch": 0.938721721941577, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0768415927886963, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6829953193664551, + "num_tokens": 221254321.0, + "step": 8548 + }, + { + "epoch": 0.9388315396441906, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.3841304779052734, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7071066498756409, + "num_tokens": 221277301.0, + "step": 8549 + }, + { + "epoch": 0.9389413573468043, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.419278621673584, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7256811857223511, + "num_tokens": 221298457.0, + "step": 8550 + }, + { + "epoch": 0.9390511750494179, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1551687717437744, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7120413184165955, + "num_tokens": 221323999.0, + "step": 8551 + }, + { + "epoch": 0.9391609927520316, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.027346611022949, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7206735014915466, + "num_tokens": 221350861.0, + "step": 8552 + }, + { + "epoch": 0.9392708104546453, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.061135768890381, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.6998242139816284, + "num_tokens": 221379066.0, + "step": 8553 + }, + { + "epoch": 0.939380628157259, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.160733699798584, + "learning_rate": 1e-06, + "loss": 1.0899, + "mean_token_accuracy": 0.6751562356948853, + "num_tokens": 221405645.0, + "step": 8554 + }, + { + "epoch": 0.9394904458598726, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.924981951713562, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.707919180393219, + "num_tokens": 221434456.0, + "step": 8555 + }, + { + "epoch": 0.9396002635624863, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.3015506267547607, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7253284454345703, + "num_tokens": 221455517.0, + "step": 8556 + }, + { + "epoch": 0.9397100812650999, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4736897945404053, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7128640413284302, + "num_tokens": 221475373.0, + "step": 8557 + }, + { + "epoch": 0.9398198989677136, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.130380868911743, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7117915153503418, + "num_tokens": 221503028.0, + "step": 8558 + }, + { + "epoch": 0.9399297166703272, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.351574659347534, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7162950038909912, + "num_tokens": 221526144.0, + "step": 8559 + }, + { + "epoch": 0.940039534372941, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.501339912414551, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.694871723651886, + "num_tokens": 221547183.0, + "step": 8560 + }, + { + "epoch": 0.9401493520755546, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2946362495422363, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7499080896377563, + "num_tokens": 221569164.0, + "step": 8561 + }, + { + "epoch": 0.9402591697781683, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0535712242126465, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7053008079528809, + "num_tokens": 221597868.0, + "step": 8562 + }, + { + "epoch": 0.9403689874807819, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2129740715026855, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7059599161148071, + "num_tokens": 221624517.0, + "step": 8563 + }, + { + "epoch": 0.9404788051833955, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.9905848503112793, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.7033335566520691, + "num_tokens": 221654198.0, + "step": 8564 + }, + { + "epoch": 0.9405886228860092, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4308791160583496, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.712475061416626, + "num_tokens": 221676354.0, + "step": 8565 + }, + { + "epoch": 0.9406984405886228, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.9090137481689453, + "learning_rate": 1e-06, + "loss": 1.0579, + "mean_token_accuracy": 0.679892897605896, + "num_tokens": 221711475.0, + "step": 8566 + }, + { + "epoch": 0.9408082582912366, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1709203720092773, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.719489336013794, + "num_tokens": 221738072.0, + "step": 8567 + }, + { + "epoch": 0.9409180759938502, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1301491260528564, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6942741870880127, + "num_tokens": 221765201.0, + "step": 8568 + }, + { + "epoch": 0.9410278936964639, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.110241413116455, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7059317231178284, + "num_tokens": 221789983.0, + "step": 8569 + }, + { + "epoch": 0.9411377113990775, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0288264751434326, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7119866609573364, + "num_tokens": 221818675.0, + "step": 8570 + }, + { + "epoch": 0.9412475291016912, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.007040023803711, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6965938806533813, + "num_tokens": 221849966.0, + "step": 8571 + }, + { + "epoch": 0.9413573468043048, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1184804439544678, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.708793044090271, + "num_tokens": 221877967.0, + "step": 8572 + }, + { + "epoch": 0.9414671645069185, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2344062328338623, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.731480598449707, + "num_tokens": 221901832.0, + "step": 8573 + }, + { + "epoch": 0.9415769822095322, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.303321123123169, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7094475030899048, + "num_tokens": 221926364.0, + "step": 8574 + }, + { + "epoch": 0.9416867999121459, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2584071159362793, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7226963043212891, + "num_tokens": 221949997.0, + "step": 8575 + }, + { + "epoch": 0.9417966176147595, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.6312427520751953, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7148246765136719, + "num_tokens": 221972166.0, + "step": 8576 + }, + { + "epoch": 0.9419064353173732, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.038666248321533, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.7031835317611694, + "num_tokens": 222000577.0, + "step": 8577 + }, + { + "epoch": 0.9420162530199868, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1064486503601074, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7136340141296387, + "num_tokens": 222028566.0, + "step": 8578 + }, + { + "epoch": 0.9421260707226005, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2800381183624268, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7285594940185547, + "num_tokens": 222052475.0, + "step": 8579 + }, + { + "epoch": 0.9422358884252141, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.9799776077270508, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.6969826221466064, + "num_tokens": 222080988.0, + "step": 8580 + }, + { + "epoch": 0.9423457061278278, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.123948812484741, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.7044975757598877, + "num_tokens": 222108697.0, + "step": 8581 + }, + { + "epoch": 0.9424555238304415, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.7201590538024902, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7228503823280334, + "num_tokens": 222126981.0, + "step": 8582 + }, + { + "epoch": 0.9425653415330552, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0836899280548096, + "learning_rate": 1e-06, + "loss": 1.1121, + "mean_token_accuracy": 0.6709038019180298, + "num_tokens": 222155767.0, + "step": 8583 + }, + { + "epoch": 0.9426751592356688, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4910049438476562, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7000892162322998, + "num_tokens": 222177533.0, + "step": 8584 + }, + { + "epoch": 0.9427849769382824, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.1150832176208496, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7159863710403442, + "num_tokens": 222201844.0, + "step": 8585 + }, + { + "epoch": 0.9428947946408961, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.324483871459961, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7157778143882751, + "num_tokens": 222224495.0, + "step": 8586 + }, + { + "epoch": 0.9430046123435097, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.6197509765625, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7052173614501953, + "num_tokens": 222243560.0, + "step": 8587 + }, + { + "epoch": 0.9431144300461234, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.991607666015625, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7218846678733826, + "num_tokens": 222274129.0, + "step": 8588 + }, + { + "epoch": 0.9432242477487371, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.0227935314178467, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7192878723144531, + "num_tokens": 222305237.0, + "step": 8589 + }, + { + "epoch": 0.9433340654513508, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.405658483505249, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7029675245285034, + "num_tokens": 222328774.0, + "step": 8590 + }, + { + "epoch": 0.9434438831539644, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.112079381942749, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.723717212677002, + "num_tokens": 222352974.0, + "step": 8591 + }, + { + "epoch": 0.9435537008565781, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2931249141693115, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.704782247543335, + "num_tokens": 222375481.0, + "step": 8592 + }, + { + "epoch": 0.9436635185591917, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.219212293624878, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7056211829185486, + "num_tokens": 222399146.0, + "step": 8593 + }, + { + "epoch": 0.9437733362618054, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 1.847009539604187, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.6957203149795532, + "num_tokens": 222432584.0, + "step": 8594 + }, + { + "epoch": 0.943883153964419, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4401755332946777, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7048371434211731, + "num_tokens": 222454366.0, + "step": 8595 + }, + { + "epoch": 0.9439929716670328, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2858409881591797, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7102779150009155, + "num_tokens": 222479174.0, + "step": 8596 + }, + { + "epoch": 0.9441027893696464, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.237997531890869, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7252122163772583, + "num_tokens": 222503908.0, + "step": 8597 + }, + { + "epoch": 0.9442126070722601, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.640005588531494, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7431166768074036, + "num_tokens": 222520848.0, + "step": 8598 + }, + { + "epoch": 0.9443224247748737, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4870107173919678, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7308917045593262, + "num_tokens": 222538364.0, + "step": 8599 + }, + { + "epoch": 0.9444322424774874, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.098391056060791, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6934279203414917, + "num_tokens": 222565875.0, + "step": 8600 + }, + { + "epoch": 0.944542060180101, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2521653175354004, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.6840747594833374, + "num_tokens": 222591050.0, + "step": 8601 + }, + { + "epoch": 0.9446518778827147, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.156564950942993, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7159188985824585, + "num_tokens": 222616289.0, + "step": 8602 + }, + { + "epoch": 0.9447616955853284, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.046290159225464, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7051115036010742, + "num_tokens": 222642934.0, + "step": 8603 + }, + { + "epoch": 0.944871513287942, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.072434186935425, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7330469489097595, + "num_tokens": 222669971.0, + "step": 8604 + }, + { + "epoch": 0.9449813309905557, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.313032865524292, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7174369096755981, + "num_tokens": 222693886.0, + "step": 8605 + }, + { + "epoch": 0.9450911486931693, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.233947992324829, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7020726203918457, + "num_tokens": 222718751.0, + "step": 8606 + }, + { + "epoch": 0.945200966395783, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.4295284748077393, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7292957901954651, + "num_tokens": 222739723.0, + "step": 8607 + }, + { + "epoch": 0.9453107840983966, + "ewc_loss": 7.2479248046875e-05, + "grad_norm": 2.2478926181793213, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7157753705978394, + "num_tokens": 222764431.0, + "step": 8608 + }, + { + "epoch": 0.9454206018010103, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.3560361862182617, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.720640242099762, + "num_tokens": 222786067.0, + "step": 8609 + }, + { + "epoch": 0.9455304195036239, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.18866229057312, + "learning_rate": 1e-06, + "loss": 1.0359, + "mean_token_accuracy": 0.6892940998077393, + "num_tokens": 222811849.0, + "step": 8610 + }, + { + "epoch": 0.9456402372062377, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.253800868988037, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6906626224517822, + "num_tokens": 222837098.0, + "step": 8611 + }, + { + "epoch": 0.9457500549088513, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.130951404571533, + "learning_rate": 1e-06, + "loss": 1.0605, + "mean_token_accuracy": 0.6925946474075317, + "num_tokens": 222867209.0, + "step": 8612 + }, + { + "epoch": 0.945859872611465, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.224574089050293, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6858125925064087, + "num_tokens": 222891024.0, + "step": 8613 + }, + { + "epoch": 0.9459696903140786, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.193498373031616, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7159068584442139, + "num_tokens": 222914495.0, + "step": 8614 + }, + { + "epoch": 0.9460795080166923, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.5086898803710938, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7212187051773071, + "num_tokens": 222933415.0, + "step": 8615 + }, + { + "epoch": 0.9461893257193059, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.2965424060821533, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7084774374961853, + "num_tokens": 222956258.0, + "step": 8616 + }, + { + "epoch": 0.9462991434219196, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.43766713142395, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7185791730880737, + "num_tokens": 222976109.0, + "step": 8617 + }, + { + "epoch": 0.9464089611245333, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.188263177871704, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.7022377848625183, + "num_tokens": 223002419.0, + "step": 8618 + }, + { + "epoch": 0.946518778827147, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.106330633163452, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7024039030075073, + "num_tokens": 223029426.0, + "step": 8619 + }, + { + "epoch": 0.9466285965297606, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.12329363822937, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7247628569602966, + "num_tokens": 223055601.0, + "step": 8620 + }, + { + "epoch": 0.9467384142323743, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.0269031524658203, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6905688047409058, + "num_tokens": 223086401.0, + "step": 8621 + }, + { + "epoch": 0.9468482319349879, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.190908670425415, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7072776556015015, + "num_tokens": 223112709.0, + "step": 8622 + }, + { + "epoch": 0.9469580496376016, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.38142991065979, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7127141952514648, + "num_tokens": 223132726.0, + "step": 8623 + }, + { + "epoch": 0.9470678673402152, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.6092426776885986, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.72624272108078, + "num_tokens": 223151442.0, + "step": 8624 + }, + { + "epoch": 0.947177685042829, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.2476589679718018, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7062503695487976, + "num_tokens": 223174241.0, + "step": 8625 + }, + { + "epoch": 0.9472875027454426, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.354867935180664, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7202386856079102, + "num_tokens": 223196993.0, + "step": 8626 + }, + { + "epoch": 0.9473973204480562, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 1.9043028354644775, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7006887197494507, + "num_tokens": 223226771.0, + "step": 8627 + }, + { + "epoch": 0.9475071381506699, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.2817182540893555, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7160284519195557, + "num_tokens": 223249487.0, + "step": 8628 + }, + { + "epoch": 0.9476169558532835, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.128767967224121, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6991590261459351, + "num_tokens": 223275629.0, + "step": 8629 + }, + { + "epoch": 0.9477267735558972, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.1177287101745605, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.6901102662086487, + "num_tokens": 223304025.0, + "step": 8630 + }, + { + "epoch": 0.9478365912585108, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.0689425468444824, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7062224745750427, + "num_tokens": 223333077.0, + "step": 8631 + }, + { + "epoch": 0.9479464089611246, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.3902275562286377, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7018018364906311, + "num_tokens": 223356647.0, + "step": 8632 + }, + { + "epoch": 0.9480562266637382, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.7973313331604004, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7298482656478882, + "num_tokens": 223374548.0, + "step": 8633 + }, + { + "epoch": 0.9481660443663519, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.2501556873321533, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7199326753616333, + "num_tokens": 223400800.0, + "step": 8634 + }, + { + "epoch": 0.9482758620689655, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.6857798099517822, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.729753315448761, + "num_tokens": 223419768.0, + "step": 8635 + }, + { + "epoch": 0.9483856797715792, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.051421642303467, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.714848518371582, + "num_tokens": 223446443.0, + "step": 8636 + }, + { + "epoch": 0.9484954974741928, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.1016652584075928, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7186160087585449, + "num_tokens": 223473510.0, + "step": 8637 + }, + { + "epoch": 0.9486053151768065, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.1564817428588867, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7361774444580078, + "num_tokens": 223503095.0, + "step": 8638 + }, + { + "epoch": 0.9487151328794201, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.1040024757385254, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6987378597259521, + "num_tokens": 223532652.0, + "step": 8639 + }, + { + "epoch": 0.9488249505820339, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.1080877780914307, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7165297269821167, + "num_tokens": 223560980.0, + "step": 8640 + }, + { + "epoch": 0.9489347682846475, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.009270429611206, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7288007736206055, + "num_tokens": 223586348.0, + "step": 8641 + }, + { + "epoch": 0.9490445859872612, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.478471279144287, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7110958695411682, + "num_tokens": 223607708.0, + "step": 8642 + }, + { + "epoch": 0.9491544036898748, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.475260019302368, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7315982580184937, + "num_tokens": 223628312.0, + "step": 8643 + }, + { + "epoch": 0.9492642213924884, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.10781192779541, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7079715728759766, + "num_tokens": 223656073.0, + "step": 8644 + }, + { + "epoch": 0.9493740390951021, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.1677637100219727, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7132138013839722, + "num_tokens": 223681909.0, + "step": 8645 + }, + { + "epoch": 0.9494838567977157, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.233330249786377, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6944633722305298, + "num_tokens": 223706150.0, + "step": 8646 + }, + { + "epoch": 0.9495936745003295, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.30104660987854, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7043477892875671, + "num_tokens": 223730524.0, + "step": 8647 + }, + { + "epoch": 0.9497034922029431, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.6488068103790283, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7371137142181396, + "num_tokens": 223748705.0, + "step": 8648 + }, + { + "epoch": 0.9498133099055568, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 1.8951590061187744, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.7015488147735596, + "num_tokens": 223784072.0, + "step": 8649 + }, + { + "epoch": 0.9499231276081704, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.0988361835479736, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7380621433258057, + "num_tokens": 223811086.0, + "step": 8650 + }, + { + "epoch": 0.9500329453107841, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.296597957611084, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7154203653335571, + "num_tokens": 223835135.0, + "step": 8651 + }, + { + "epoch": 0.9501427630133977, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.219637632369995, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6896138787269592, + "num_tokens": 223861787.0, + "step": 8652 + }, + { + "epoch": 0.9502525807160114, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.057377576828003, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6876842379570007, + "num_tokens": 223891162.0, + "step": 8653 + }, + { + "epoch": 0.9503623984186251, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.403472661972046, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.718633770942688, + "num_tokens": 223912433.0, + "step": 8654 + }, + { + "epoch": 0.9504722161212388, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.2956416606903076, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6884640455245972, + "num_tokens": 223938108.0, + "step": 8655 + }, + { + "epoch": 0.9505820338238524, + "ewc_loss": 7.295608520507812e-05, + "grad_norm": 2.059283494949341, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6880374550819397, + "num_tokens": 223965637.0, + "step": 8656 + }, + { + "epoch": 0.9506918515264661, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.557389259338379, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7146767973899841, + "num_tokens": 223985296.0, + "step": 8657 + }, + { + "epoch": 0.9508016692290797, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.0959932804107666, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6985673308372498, + "num_tokens": 224013684.0, + "step": 8658 + }, + { + "epoch": 0.9509114869316934, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.0120744705200195, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.702325165271759, + "num_tokens": 224041925.0, + "step": 8659 + }, + { + "epoch": 0.951021304634307, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.4300200939178467, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7052381038665771, + "num_tokens": 224062807.0, + "step": 8660 + }, + { + "epoch": 0.9511311223369208, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.101848840713501, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.7067191004753113, + "num_tokens": 224090444.0, + "step": 8661 + }, + { + "epoch": 0.9512409400395344, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.3459794521331787, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6937724351882935, + "num_tokens": 224115116.0, + "step": 8662 + }, + { + "epoch": 0.9513507577421481, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.0217907428741455, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.6918972730636597, + "num_tokens": 224144631.0, + "step": 8663 + }, + { + "epoch": 0.9514605754447617, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.051816940307617, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7097599506378174, + "num_tokens": 224173130.0, + "step": 8664 + }, + { + "epoch": 0.9515703931473753, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.29530668258667, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7293355464935303, + "num_tokens": 224196040.0, + "step": 8665 + }, + { + "epoch": 0.951680210849989, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.0655906200408936, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.736115038394928, + "num_tokens": 224221596.0, + "step": 8666 + }, + { + "epoch": 0.9517900285526026, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.205524444580078, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7143927812576294, + "num_tokens": 224247333.0, + "step": 8667 + }, + { + "epoch": 0.9518998462552163, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.144139528274536, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7202687859535217, + "num_tokens": 224274178.0, + "step": 8668 + }, + { + "epoch": 0.95200966395783, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 1.9961152076721191, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.6964006423950195, + "num_tokens": 224304474.0, + "step": 8669 + }, + { + "epoch": 0.9521194816604437, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.314528226852417, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6875138282775879, + "num_tokens": 224329214.0, + "step": 8670 + }, + { + "epoch": 0.9522292993630573, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.197113275527954, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.7080810070037842, + "num_tokens": 224353752.0, + "step": 8671 + }, + { + "epoch": 0.952339117065671, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.713935613632202, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7377524375915527, + "num_tokens": 224369261.0, + "step": 8672 + }, + { + "epoch": 0.9524489347682846, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.0942542552948, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7072257995605469, + "num_tokens": 224397273.0, + "step": 8673 + }, + { + "epoch": 0.9525587524708983, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.3585498332977295, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7107460498809814, + "num_tokens": 224419111.0, + "step": 8674 + }, + { + "epoch": 0.9526685701735119, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.135471820831299, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7142265439033508, + "num_tokens": 224445100.0, + "step": 8675 + }, + { + "epoch": 0.9527783878761257, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.232408046722412, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7293552160263062, + "num_tokens": 224469047.0, + "step": 8676 + }, + { + "epoch": 0.9528882055787393, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.289364814758301, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7200595736503601, + "num_tokens": 224491089.0, + "step": 8677 + }, + { + "epoch": 0.952998023281353, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.195176362991333, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7061063647270203, + "num_tokens": 224515134.0, + "step": 8678 + }, + { + "epoch": 0.9531078409839666, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 1.9854085445404053, + "learning_rate": 1e-06, + "loss": 1.0726, + "mean_token_accuracy": 0.6784523129463196, + "num_tokens": 224548089.0, + "step": 8679 + }, + { + "epoch": 0.9532176586865803, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.204897880554199, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7097922563552856, + "num_tokens": 224571705.0, + "step": 8680 + }, + { + "epoch": 0.9533274763891939, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.202899932861328, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6906168460845947, + "num_tokens": 224596889.0, + "step": 8681 + }, + { + "epoch": 0.9534372940918076, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 1.8124934434890747, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7156095504760742, + "num_tokens": 224632269.0, + "step": 8682 + }, + { + "epoch": 0.9535471117944213, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.048478126525879, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.6990931034088135, + "num_tokens": 224659968.0, + "step": 8683 + }, + { + "epoch": 0.953656929497035, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 1.975603699684143, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7160429954528809, + "num_tokens": 224689653.0, + "step": 8684 + }, + { + "epoch": 0.9537667471996486, + "ewc_loss": 7.390975952148438e-05, + "grad_norm": 2.005096912384033, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6967892050743103, + "num_tokens": 224717652.0, + "step": 8685 + }, + { + "epoch": 0.9538765649022622, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.9175044298171997, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6917852163314819, + "num_tokens": 224751602.0, + "step": 8686 + }, + { + "epoch": 0.9539863826048759, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2446670532226562, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6955904364585876, + "num_tokens": 224775812.0, + "step": 8687 + }, + { + "epoch": 0.9540962003074895, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2418291568756104, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7243305444717407, + "num_tokens": 224798433.0, + "step": 8688 + }, + { + "epoch": 0.9542060180101032, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0985100269317627, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.6958437561988831, + "num_tokens": 224826605.0, + "step": 8689 + }, + { + "epoch": 0.9543158357127169, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.170445203781128, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7262232899665833, + "num_tokens": 224850608.0, + "step": 8690 + }, + { + "epoch": 0.9544256534153306, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.4873390197753906, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7091503143310547, + "num_tokens": 224872229.0, + "step": 8691 + }, + { + "epoch": 0.9545354711179442, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.097367763519287, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7222834825515747, + "num_tokens": 224899062.0, + "step": 8692 + }, + { + "epoch": 0.9546452888205579, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.4354922771453857, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7149649262428284, + "num_tokens": 224920953.0, + "step": 8693 + }, + { + "epoch": 0.9547551065231715, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2890431880950928, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7026393413543701, + "num_tokens": 224945827.0, + "step": 8694 + }, + { + "epoch": 0.9548649242257852, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.1104798316955566, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7039583921432495, + "num_tokens": 224975422.0, + "step": 8695 + }, + { + "epoch": 0.9549747419283988, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.221989631652832, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6978362202644348, + "num_tokens": 225001592.0, + "step": 8696 + }, + { + "epoch": 0.9550845596310125, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.5442726612091064, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7205349802970886, + "num_tokens": 225022802.0, + "step": 8697 + }, + { + "epoch": 0.9551943773336262, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2533531188964844, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6871005296707153, + "num_tokens": 225048842.0, + "step": 8698 + }, + { + "epoch": 0.9553041950362399, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.9616426229476929, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7092928886413574, + "num_tokens": 225080619.0, + "step": 8699 + }, + { + "epoch": 0.9554140127388535, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.932613492012024, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7108830213546753, + "num_tokens": 225110529.0, + "step": 8700 + }, + { + "epoch": 0.9555238304414672, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.1189587116241455, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7368038892745972, + "num_tokens": 225136463.0, + "step": 8701 + }, + { + "epoch": 0.9556336481440808, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.212310314178467, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.687638521194458, + "num_tokens": 225161023.0, + "step": 8702 + }, + { + "epoch": 0.9557434658466945, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.1365504264831543, + "learning_rate": 1e-06, + "loss": 1.1021, + "mean_token_accuracy": 0.6732166409492493, + "num_tokens": 225188822.0, + "step": 8703 + }, + { + "epoch": 0.9558532835493081, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.9990495443344116, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7056138515472412, + "num_tokens": 225219426.0, + "step": 8704 + }, + { + "epoch": 0.9559631012519219, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.4677586555480957, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7122582197189331, + "num_tokens": 225241096.0, + "step": 8705 + }, + { + "epoch": 0.9560729189545355, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2292778491973877, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7070870995521545, + "num_tokens": 225264747.0, + "step": 8706 + }, + { + "epoch": 0.9561827366571491, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2328503131866455, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7057464718818665, + "num_tokens": 225287683.0, + "step": 8707 + }, + { + "epoch": 0.9562925543597628, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.056252956390381, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6909029483795166, + "num_tokens": 225316574.0, + "step": 8708 + }, + { + "epoch": 0.9564023720623764, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.120809316635132, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7102466821670532, + "num_tokens": 225343257.0, + "step": 8709 + }, + { + "epoch": 0.9565121897649901, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2872862815856934, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7403424978256226, + "num_tokens": 225366372.0, + "step": 8710 + }, + { + "epoch": 0.9566220074676037, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0236458778381348, + "learning_rate": 1e-06, + "loss": 1.063, + "mean_token_accuracy": 0.6892532110214233, + "num_tokens": 225397796.0, + "step": 8711 + }, + { + "epoch": 0.9567318251702175, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0152010917663574, + "learning_rate": 1e-06, + "loss": 1.1176, + "mean_token_accuracy": 0.6783764362335205, + "num_tokens": 225431624.0, + "step": 8712 + }, + { + "epoch": 0.9568416428728311, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0878195762634277, + "learning_rate": 1e-06, + "loss": 1.0822, + "mean_token_accuracy": 0.680050253868103, + "num_tokens": 225459819.0, + "step": 8713 + }, + { + "epoch": 0.9569514605754448, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.1332497596740723, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7079799175262451, + "num_tokens": 225487993.0, + "step": 8714 + }, + { + "epoch": 0.9570612782780584, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0850744247436523, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7196112275123596, + "num_tokens": 225513334.0, + "step": 8715 + }, + { + "epoch": 0.9571710959806721, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.231754779815674, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7112760543823242, + "num_tokens": 225537623.0, + "step": 8716 + }, + { + "epoch": 0.9572809136832857, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.5603601932525635, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7075501084327698, + "num_tokens": 225556885.0, + "step": 8717 + }, + { + "epoch": 0.9573907313858994, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.3047876358032227, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.6999037861824036, + "num_tokens": 225582847.0, + "step": 8718 + }, + { + "epoch": 0.9575005490885131, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.6887400150299072, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7357529401779175, + "num_tokens": 225598872.0, + "step": 8719 + }, + { + "epoch": 0.9576103667911268, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.232448101043701, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6989290714263916, + "num_tokens": 225624823.0, + "step": 8720 + }, + { + "epoch": 0.9577201844937404, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.226346731185913, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7017084956169128, + "num_tokens": 225649755.0, + "step": 8721 + }, + { + "epoch": 0.9578300021963541, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.8207738399505615, + "learning_rate": 1e-06, + "loss": 1.0251, + "mean_token_accuracy": 0.6879541873931885, + "num_tokens": 225683143.0, + "step": 8722 + }, + { + "epoch": 0.9579398198989677, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.867612361907959, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7224416732788086, + "num_tokens": 225714652.0, + "step": 8723 + }, + { + "epoch": 0.9580496376015813, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.2393276691436768, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7201269865036011, + "num_tokens": 225737649.0, + "step": 8724 + }, + { + "epoch": 0.958159455304195, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.1677584648132324, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7178777456283569, + "num_tokens": 225762807.0, + "step": 8725 + }, + { + "epoch": 0.9582692730068088, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.05566668510437, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.6885964870452881, + "num_tokens": 225790806.0, + "step": 8726 + }, + { + "epoch": 0.9583790907094224, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0861048698425293, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7155748009681702, + "num_tokens": 225818475.0, + "step": 8727 + }, + { + "epoch": 0.958488908412036, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.3980727195739746, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7130805253982544, + "num_tokens": 225838758.0, + "step": 8728 + }, + { + "epoch": 0.9585987261146497, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.022139072418213, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7091928720474243, + "num_tokens": 225869813.0, + "step": 8729 + }, + { + "epoch": 0.9587085438172633, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.212412118911743, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7198518514633179, + "num_tokens": 225893794.0, + "step": 8730 + }, + { + "epoch": 0.958818361519877, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.0951247215270996, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.699607789516449, + "num_tokens": 225920531.0, + "step": 8731 + }, + { + "epoch": 0.9589281792224906, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 1.9545397758483887, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7121621966362, + "num_tokens": 225951936.0, + "step": 8732 + }, + { + "epoch": 0.9590379969251043, + "ewc_loss": 7.43865966796875e-05, + "grad_norm": 2.3599348068237305, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7146883010864258, + "num_tokens": 225974222.0, + "step": 8733 + }, + { + "epoch": 0.959147814627718, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.275275945663452, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7259913086891174, + "num_tokens": 225997213.0, + "step": 8734 + }, + { + "epoch": 0.9592576323303317, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.4382214546203613, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7035974264144897, + "num_tokens": 226017402.0, + "step": 8735 + }, + { + "epoch": 0.9593674500329453, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1266286373138428, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6857696771621704, + "num_tokens": 226046124.0, + "step": 8736 + }, + { + "epoch": 0.959477267735559, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1927313804626465, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.700441837310791, + "num_tokens": 226072832.0, + "step": 8737 + }, + { + "epoch": 0.9595870854381726, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.013735294342041, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.713686466217041, + "num_tokens": 226099350.0, + "step": 8738 + }, + { + "epoch": 0.9596969031407863, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1568703651428223, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7099491357803345, + "num_tokens": 226123899.0, + "step": 8739 + }, + { + "epoch": 0.9598067208433999, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0344419479370117, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7117111086845398, + "num_tokens": 226151450.0, + "step": 8740 + }, + { + "epoch": 0.9599165385460137, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9998239278793335, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7188825607299805, + "num_tokens": 226180002.0, + "step": 8741 + }, + { + "epoch": 0.9600263562486273, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9924670457839966, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6955898404121399, + "num_tokens": 226207642.0, + "step": 8742 + }, + { + "epoch": 0.960136173951241, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.293274402618408, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7077714800834656, + "num_tokens": 226233048.0, + "step": 8743 + }, + { + "epoch": 0.9602459916538546, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.27843976020813, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7253535985946655, + "num_tokens": 226254190.0, + "step": 8744 + }, + { + "epoch": 0.9603558093564682, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1258320808410645, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7045445442199707, + "num_tokens": 226280390.0, + "step": 8745 + }, + { + "epoch": 0.9604656270590819, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0486764907836914, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.7047765254974365, + "num_tokens": 226308489.0, + "step": 8746 + }, + { + "epoch": 0.9605754447616955, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1286587715148926, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7153524160385132, + "num_tokens": 226334100.0, + "step": 8747 + }, + { + "epoch": 0.9606852624643093, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.4218289852142334, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7227181792259216, + "num_tokens": 226355150.0, + "step": 8748 + }, + { + "epoch": 0.9607950801669229, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0753912925720215, + "learning_rate": 1e-06, + "loss": 1.0625, + "mean_token_accuracy": 0.6841491460800171, + "num_tokens": 226385231.0, + "step": 8749 + }, + { + "epoch": 0.9609048978695366, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9350450038909912, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7115763425827026, + "num_tokens": 226416209.0, + "step": 8750 + }, + { + "epoch": 0.9610147155721502, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.4683871269226074, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7137538194656372, + "num_tokens": 226435635.0, + "step": 8751 + }, + { + "epoch": 0.9611245332747639, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9027035236358643, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7027574777603149, + "num_tokens": 226468132.0, + "step": 8752 + }, + { + "epoch": 0.9612343509773775, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1632909774780273, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7258338928222656, + "num_tokens": 226492763.0, + "step": 8753 + }, + { + "epoch": 0.9613441686799912, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2103822231292725, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7081612348556519, + "num_tokens": 226518240.0, + "step": 8754 + }, + { + "epoch": 0.9614539863826049, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.440875768661499, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7118723392486572, + "num_tokens": 226538071.0, + "step": 8755 + }, + { + "epoch": 0.9615638040852186, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.8291419744491577, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7070303559303284, + "num_tokens": 226567951.0, + "step": 8756 + }, + { + "epoch": 0.9616736217878322, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.6238670349121094, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7125742435455322, + "num_tokens": 226585283.0, + "step": 8757 + }, + { + "epoch": 0.9617834394904459, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.994725227355957, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7030678391456604, + "num_tokens": 226614720.0, + "step": 8758 + }, + { + "epoch": 0.9618932571930595, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.4619014263153076, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7110615372657776, + "num_tokens": 226636638.0, + "step": 8759 + }, + { + "epoch": 0.9620030748956732, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0651698112487793, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6962165832519531, + "num_tokens": 226664693.0, + "step": 8760 + }, + { + "epoch": 0.9621128925982868, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2071826457977295, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6783267259597778, + "num_tokens": 226691569.0, + "step": 8761 + }, + { + "epoch": 0.9622227103009005, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.600580930709839, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7252288460731506, + "num_tokens": 226708968.0, + "step": 8762 + }, + { + "epoch": 0.9623325280035142, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9725379943847656, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6938745975494385, + "num_tokens": 226739371.0, + "step": 8763 + }, + { + "epoch": 0.9624423457061279, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1432290077209473, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.698029637336731, + "num_tokens": 226765112.0, + "step": 8764 + }, + { + "epoch": 0.9625521634087415, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1964967250823975, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7049073576927185, + "num_tokens": 226790665.0, + "step": 8765 + }, + { + "epoch": 0.9626619811113551, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0034425258636475, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7071772217750549, + "num_tokens": 226820430.0, + "step": 8766 + }, + { + "epoch": 0.9627717988139688, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9674274921417236, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6953328847885132, + "num_tokens": 226852716.0, + "step": 8767 + }, + { + "epoch": 0.9628816165165824, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0493886470794678, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7165683507919312, + "num_tokens": 226880918.0, + "step": 8768 + }, + { + "epoch": 0.9629914342191961, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2195885181427, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7051827311515808, + "num_tokens": 226904758.0, + "step": 8769 + }, + { + "epoch": 0.9631012519218098, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1352665424346924, + "learning_rate": 1e-06, + "loss": 1.0199, + "mean_token_accuracy": 0.6929800510406494, + "num_tokens": 226933464.0, + "step": 8770 + }, + { + "epoch": 0.9632110696244235, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0676987171173096, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7164018750190735, + "num_tokens": 226961442.0, + "step": 8771 + }, + { + "epoch": 0.9633208873270371, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1891684532165527, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7014356255531311, + "num_tokens": 226987718.0, + "step": 8772 + }, + { + "epoch": 0.9634307050296508, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.194688320159912, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7008189558982849, + "num_tokens": 227014707.0, + "step": 8773 + }, + { + "epoch": 0.9635405227322644, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.083218812942505, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6973372101783752, + "num_tokens": 227042658.0, + "step": 8774 + }, + { + "epoch": 0.9636503404348781, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2549474239349365, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7200063467025757, + "num_tokens": 227065980.0, + "step": 8775 + }, + { + "epoch": 0.9637601581374917, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1561524868011475, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6875865459442139, + "num_tokens": 227092798.0, + "step": 8776 + }, + { + "epoch": 0.9638699758401055, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0964715480804443, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7001171112060547, + "num_tokens": 227119037.0, + "step": 8777 + }, + { + "epoch": 0.9639797935427191, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 1.9874053001403809, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7139289379119873, + "num_tokens": 227147385.0, + "step": 8778 + }, + { + "epoch": 0.9640896112453328, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0376768112182617, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7013007402420044, + "num_tokens": 227175846.0, + "step": 8779 + }, + { + "epoch": 0.9641994289479464, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1036794185638428, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6945656538009644, + "num_tokens": 227203360.0, + "step": 8780 + }, + { + "epoch": 0.9643092466505601, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.190624475479126, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.700046956539154, + "num_tokens": 227228358.0, + "step": 8781 + }, + { + "epoch": 0.9644190643531737, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0790469646453857, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7244799137115479, + "num_tokens": 227255492.0, + "step": 8782 + }, + { + "epoch": 0.9645288820557874, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.4181649684906006, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7137529850006104, + "num_tokens": 227276939.0, + "step": 8783 + }, + { + "epoch": 0.9646386997584011, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0070242881774902, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6972466707229614, + "num_tokens": 227305598.0, + "step": 8784 + }, + { + "epoch": 0.9647485174610148, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2606029510498047, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.7019461393356323, + "num_tokens": 227329581.0, + "step": 8785 + }, + { + "epoch": 0.9648583351636284, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.1740880012512207, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.710208535194397, + "num_tokens": 227354844.0, + "step": 8786 + }, + { + "epoch": 0.964968152866242, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.0364935398101807, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7005516290664673, + "num_tokens": 227383559.0, + "step": 8787 + }, + { + "epoch": 0.9650779705688557, + "ewc_loss": 7.486343383789062e-05, + "grad_norm": 2.2095985412597656, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.716853141784668, + "num_tokens": 227407240.0, + "step": 8788 + }, + { + "epoch": 0.9651877882714693, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.938594102859497, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.7083202600479126, + "num_tokens": 227438972.0, + "step": 8789 + }, + { + "epoch": 0.965297605974083, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.052961826324463, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.7048604488372803, + "num_tokens": 227466755.0, + "step": 8790 + }, + { + "epoch": 0.9654074236766966, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.4717748165130615, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7353838682174683, + "num_tokens": 227486042.0, + "step": 8791 + }, + { + "epoch": 0.9655172413793104, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.0390841960906982, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7327773571014404, + "num_tokens": 227513569.0, + "step": 8792 + }, + { + "epoch": 0.965627059081924, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.4349730014801025, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7022944688796997, + "num_tokens": 227536632.0, + "step": 8793 + }, + { + "epoch": 0.9657368767845377, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.0622456073760986, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6912633180618286, + "num_tokens": 227565842.0, + "step": 8794 + }, + { + "epoch": 0.9658466944871513, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.944298267364502, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7043496370315552, + "num_tokens": 227595085.0, + "step": 8795 + }, + { + "epoch": 0.965956512189765, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.16641902923584, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7124896049499512, + "num_tokens": 227619531.0, + "step": 8796 + }, + { + "epoch": 0.9660663298923786, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2416234016418457, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.6942300796508789, + "num_tokens": 227643118.0, + "step": 8797 + }, + { + "epoch": 0.9661761475949923, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.1324808597564697, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7128535509109497, + "num_tokens": 227668000.0, + "step": 8798 + }, + { + "epoch": 0.966285965297606, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2862179279327393, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7282158136367798, + "num_tokens": 227690524.0, + "step": 8799 + }, + { + "epoch": 0.9663957830002197, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2081496715545654, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7174813151359558, + "num_tokens": 227714620.0, + "step": 8800 + }, + { + "epoch": 0.9665056007028333, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.128805160522461, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7250070571899414, + "num_tokens": 227739464.0, + "step": 8801 + }, + { + "epoch": 0.966615418405447, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.193405866622925, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.701491117477417, + "num_tokens": 227764732.0, + "step": 8802 + }, + { + "epoch": 0.9667252361080606, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.9410141706466675, + "learning_rate": 1e-06, + "loss": 1.032, + "mean_token_accuracy": 0.6841956377029419, + "num_tokens": 227795769.0, + "step": 8803 + }, + { + "epoch": 0.9668350538106742, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2678375244140625, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7157771587371826, + "num_tokens": 227820704.0, + "step": 8804 + }, + { + "epoch": 0.9669448715132879, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.8914170265197754, + "learning_rate": 1e-06, + "loss": 1.0622, + "mean_token_accuracy": 0.6914469003677368, + "num_tokens": 227857518.0, + "step": 8805 + }, + { + "epoch": 0.9670546892159017, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.1430752277374268, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7205381393432617, + "num_tokens": 227881946.0, + "step": 8806 + }, + { + "epoch": 0.9671645069185153, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.055121660232544, + "learning_rate": 1e-06, + "loss": 1.0454, + "mean_token_accuracy": 0.6869087815284729, + "num_tokens": 227912468.0, + "step": 8807 + }, + { + "epoch": 0.967274324621129, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.3100383281707764, + "learning_rate": 1e-06, + "loss": 1.0673, + "mean_token_accuracy": 0.6822491884231567, + "num_tokens": 227938208.0, + "step": 8808 + }, + { + "epoch": 0.9673841423237426, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.197690963745117, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.6945530772209167, + "num_tokens": 227964581.0, + "step": 8809 + }, + { + "epoch": 0.9674939600263562, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.0625994205474854, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7063907384872437, + "num_tokens": 227995051.0, + "step": 8810 + }, + { + "epoch": 0.9676037777289699, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.765475273132324, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7051217555999756, + "num_tokens": 228011822.0, + "step": 8811 + }, + { + "epoch": 0.9677135954315835, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.102652072906494, + "learning_rate": 1e-06, + "loss": 1.078, + "mean_token_accuracy": 0.6734697818756104, + "num_tokens": 228041097.0, + "step": 8812 + }, + { + "epoch": 0.9678234131341973, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.8932374715805054, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7079054117202759, + "num_tokens": 228073745.0, + "step": 8813 + }, + { + "epoch": 0.9679332308368109, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.3506991863250732, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7152847051620483, + "num_tokens": 228096488.0, + "step": 8814 + }, + { + "epoch": 0.9680430485394246, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.0113794803619385, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.696161150932312, + "num_tokens": 228126843.0, + "step": 8815 + }, + { + "epoch": 0.9681528662420382, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.284916400909424, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7255688905715942, + "num_tokens": 228148818.0, + "step": 8816 + }, + { + "epoch": 0.9682626839446519, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.254836082458496, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7292274236679077, + "num_tokens": 228171522.0, + "step": 8817 + }, + { + "epoch": 0.9683725016472655, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.284914970397949, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6900244951248169, + "num_tokens": 228197254.0, + "step": 8818 + }, + { + "epoch": 0.9684823193498792, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2868547439575195, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7024168372154236, + "num_tokens": 228219818.0, + "step": 8819 + }, + { + "epoch": 0.9685921370524928, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.3852152824401855, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.717025637626648, + "num_tokens": 228239719.0, + "step": 8820 + }, + { + "epoch": 0.9687019547551066, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.233626365661621, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7257516384124756, + "num_tokens": 228262713.0, + "step": 8821 + }, + { + "epoch": 0.9688117724577202, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.1422886848449707, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7074110507965088, + "num_tokens": 228289415.0, + "step": 8822 + }, + { + "epoch": 0.9689215901603339, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.087103843688965, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7073276042938232, + "num_tokens": 228315429.0, + "step": 8823 + }, + { + "epoch": 0.9690314078629475, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.177319288253784, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7417573928833008, + "num_tokens": 228339247.0, + "step": 8824 + }, + { + "epoch": 0.9691412255655611, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.169051170349121, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7236416935920715, + "num_tokens": 228364715.0, + "step": 8825 + }, + { + "epoch": 0.9692510432681748, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.435809850692749, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7265448570251465, + "num_tokens": 228386350.0, + "step": 8826 + }, + { + "epoch": 0.9693608609707884, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.960174322128296, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7072007656097412, + "num_tokens": 228418504.0, + "step": 8827 + }, + { + "epoch": 0.9694706786734022, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.24135684967041, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7169948220252991, + "num_tokens": 228443214.0, + "step": 8828 + }, + { + "epoch": 0.9695804963760158, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.350600242614746, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7166423201560974, + "num_tokens": 228465758.0, + "step": 8829 + }, + { + "epoch": 0.9696903140786295, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 4.116939544677734, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7181859016418457, + "num_tokens": 228492918.0, + "step": 8830 + }, + { + "epoch": 0.9698001317812431, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.2509779930114746, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6966973543167114, + "num_tokens": 228517881.0, + "step": 8831 + }, + { + "epoch": 0.9699099494838568, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 1.9609568119049072, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7137424349784851, + "num_tokens": 228548073.0, + "step": 8832 + }, + { + "epoch": 0.9700197671864704, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.421815872192383, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7180588245391846, + "num_tokens": 228568379.0, + "step": 8833 + }, + { + "epoch": 0.9701295848890841, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.1100237369537354, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7085987329483032, + "num_tokens": 228595878.0, + "step": 8834 + }, + { + "epoch": 0.9702394025916978, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.106651544570923, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.6999375224113464, + "num_tokens": 228626734.0, + "step": 8835 + }, + { + "epoch": 0.9703492202943115, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.1559865474700928, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6906885504722595, + "num_tokens": 228657659.0, + "step": 8836 + }, + { + "epoch": 0.9704590379969251, + "ewc_loss": 7.534027099609375e-05, + "grad_norm": 2.3557288646698, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7104417681694031, + "num_tokens": 228679738.0, + "step": 8837 + }, + { + "epoch": 0.9705688556995388, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.104152202606201, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7098979949951172, + "num_tokens": 228706159.0, + "step": 8838 + }, + { + "epoch": 0.9706786734021524, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.432900905609131, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.710329532623291, + "num_tokens": 228727542.0, + "step": 8839 + }, + { + "epoch": 0.9707884911047661, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.1790642738342285, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.6960999965667725, + "num_tokens": 228751295.0, + "step": 8840 + }, + { + "epoch": 0.9708983088073797, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.280876874923706, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.7001084089279175, + "num_tokens": 228778417.0, + "step": 8841 + }, + { + "epoch": 0.9710081265099935, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 1.856018304824829, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6976026296615601, + "num_tokens": 228812038.0, + "step": 8842 + }, + { + "epoch": 0.9711179442126071, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0713586807250977, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7079362869262695, + "num_tokens": 228839443.0, + "step": 8843 + }, + { + "epoch": 0.9712277619152208, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.16605281829834, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7203733921051025, + "num_tokens": 228863670.0, + "step": 8844 + }, + { + "epoch": 0.9713375796178344, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 1.8223665952682495, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7053729891777039, + "num_tokens": 228899048.0, + "step": 8845 + }, + { + "epoch": 0.971447397320448, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.180063009262085, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6830487251281738, + "num_tokens": 228925711.0, + "step": 8846 + }, + { + "epoch": 0.9715572150230617, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.3485238552093506, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7101246118545532, + "num_tokens": 228948934.0, + "step": 8847 + }, + { + "epoch": 0.9716670327256753, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.043004035949707, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7109331488609314, + "num_tokens": 228975185.0, + "step": 8848 + }, + { + "epoch": 0.971776850428289, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.152583360671997, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7292722463607788, + "num_tokens": 229001724.0, + "step": 8849 + }, + { + "epoch": 0.9718866681309027, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.553417444229126, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7282371520996094, + "num_tokens": 229023802.0, + "step": 8850 + }, + { + "epoch": 0.9719964858335164, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 1.994778037071228, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7226430773735046, + "num_tokens": 229053214.0, + "step": 8851 + }, + { + "epoch": 0.97210630353613, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 1.9776360988616943, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7119815349578857, + "num_tokens": 229082556.0, + "step": 8852 + }, + { + "epoch": 0.9722161212387437, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.4365453720092773, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7015297412872314, + "num_tokens": 229102678.0, + "step": 8853 + }, + { + "epoch": 0.9723259389413573, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.332275629043579, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7094582319259644, + "num_tokens": 229127059.0, + "step": 8854 + }, + { + "epoch": 0.972435756643971, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.257624387741089, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.7024858593940735, + "num_tokens": 229151902.0, + "step": 8855 + }, + { + "epoch": 0.9725455743465846, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.287525177001953, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7206000089645386, + "num_tokens": 229174950.0, + "step": 8856 + }, + { + "epoch": 0.9726553920491984, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.3646461963653564, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7194393277168274, + "num_tokens": 229196120.0, + "step": 8857 + }, + { + "epoch": 0.972765209751812, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.161144733428955, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.6998454332351685, + "num_tokens": 229223840.0, + "step": 8858 + }, + { + "epoch": 0.9728750274544257, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.4864590167999268, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7148852348327637, + "num_tokens": 229243153.0, + "step": 8859 + }, + { + "epoch": 0.9729848451570393, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.16237473487854, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7016317844390869, + "num_tokens": 229270777.0, + "step": 8860 + }, + { + "epoch": 0.973094662859653, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.358046770095825, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7137161493301392, + "num_tokens": 229292593.0, + "step": 8861 + }, + { + "epoch": 0.9732044805622666, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.2037222385406494, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7127776741981506, + "num_tokens": 229317275.0, + "step": 8862 + }, + { + "epoch": 0.9733142982648803, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.161722183227539, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6923356652259827, + "num_tokens": 229343841.0, + "step": 8863 + }, + { + "epoch": 0.973424115967494, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.317849636077881, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.6957798004150391, + "num_tokens": 229368220.0, + "step": 8864 + }, + { + "epoch": 0.9735339336701077, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.444805145263672, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6920309662818909, + "num_tokens": 229389388.0, + "step": 8865 + }, + { + "epoch": 0.9736437513727213, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0391154289245605, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6974074840545654, + "num_tokens": 229419230.0, + "step": 8866 + }, + { + "epoch": 0.973753569075335, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0022153854370117, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6927207708358765, + "num_tokens": 229448049.0, + "step": 8867 + }, + { + "epoch": 0.9738633867779486, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.2635645866394043, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6971514225006104, + "num_tokens": 229472943.0, + "step": 8868 + }, + { + "epoch": 0.9739732044805622, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.4494593143463135, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7245177030563354, + "num_tokens": 229492958.0, + "step": 8869 + }, + { + "epoch": 0.9740830221831759, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 1.9895603656768799, + "learning_rate": 1e-06, + "loss": 0.7874, + "mean_token_accuracy": 0.753980278968811, + "num_tokens": 229518904.0, + "step": 8870 + }, + { + "epoch": 0.9741928398857896, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0684499740600586, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7052042484283447, + "num_tokens": 229546570.0, + "step": 8871 + }, + { + "epoch": 0.9743026575884033, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.361004114151001, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7168165445327759, + "num_tokens": 229567965.0, + "step": 8872 + }, + { + "epoch": 0.9744124752910169, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0289785861968994, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7119373679161072, + "num_tokens": 229596198.0, + "step": 8873 + }, + { + "epoch": 0.9745222929936306, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.0222737789154053, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7052301168441772, + "num_tokens": 229626224.0, + "step": 8874 + }, + { + "epoch": 0.9746321106962442, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.247767448425293, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.690861701965332, + "num_tokens": 229652706.0, + "step": 8875 + }, + { + "epoch": 0.9747419283988579, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.1071665287017822, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7078806161880493, + "num_tokens": 229679203.0, + "step": 8876 + }, + { + "epoch": 0.9748517461014715, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.3361024856567383, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7006703615188599, + "num_tokens": 229702096.0, + "step": 8877 + }, + { + "epoch": 0.9749615638040853, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.4870622158050537, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.698650598526001, + "num_tokens": 229721871.0, + "step": 8878 + }, + { + "epoch": 0.9750713815066989, + "ewc_loss": 7.581710815429688e-05, + "grad_norm": 2.36940598487854, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6865139007568359, + "num_tokens": 229745519.0, + "step": 8879 + }, + { + "epoch": 0.9751811992093126, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1429965496063232, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7174750566482544, + "num_tokens": 229771220.0, + "step": 8880 + }, + { + "epoch": 0.9752910169119262, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3147425651550293, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6919119358062744, + "num_tokens": 229795634.0, + "step": 8881 + }, + { + "epoch": 0.9754008346145399, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.411714792251587, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7366065979003906, + "num_tokens": 229815447.0, + "step": 8882 + }, + { + "epoch": 0.9755106523171535, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9102028608322144, + "learning_rate": 1e-06, + "loss": 1.1201, + "mean_token_accuracy": 0.6641838550567627, + "num_tokens": 229846624.0, + "step": 8883 + }, + { + "epoch": 0.9756204700197671, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.242401361465454, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7104442715644836, + "num_tokens": 229870063.0, + "step": 8884 + }, + { + "epoch": 0.9757302877223808, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9693949222564697, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7069291472434998, + "num_tokens": 229898350.0, + "step": 8885 + }, + { + "epoch": 0.9758401054249946, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3679757118225098, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.6937289237976074, + "num_tokens": 229921538.0, + "step": 8886 + }, + { + "epoch": 0.9759499231276082, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2153074741363525, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6973894834518433, + "num_tokens": 229945119.0, + "step": 8887 + }, + { + "epoch": 0.9760597408302218, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3286049365997314, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.6900099515914917, + "num_tokens": 229968920.0, + "step": 8888 + }, + { + "epoch": 0.9761695585328355, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.8763409852981567, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7096179723739624, + "num_tokens": 230003802.0, + "step": 8889 + }, + { + "epoch": 0.9762793762354491, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.196119546890259, + "learning_rate": 1e-06, + "loss": 1.0472, + "mean_token_accuracy": 0.6990588903427124, + "num_tokens": 230030227.0, + "step": 8890 + }, + { + "epoch": 0.9763891939380628, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.453051805496216, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7156517505645752, + "num_tokens": 230053413.0, + "step": 8891 + }, + { + "epoch": 0.9764990116406764, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3641364574432373, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7047790884971619, + "num_tokens": 230076540.0, + "step": 8892 + }, + { + "epoch": 0.9766088293432902, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.146409511566162, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7109977006912231, + "num_tokens": 230102079.0, + "step": 8893 + }, + { + "epoch": 0.9767186470459038, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.090592622756958, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6970257759094238, + "num_tokens": 230129033.0, + "step": 8894 + }, + { + "epoch": 0.9768284647485175, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.038520574569702, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6910347938537598, + "num_tokens": 230158669.0, + "step": 8895 + }, + { + "epoch": 0.9769382824511311, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.4581847190856934, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.722826361656189, + "num_tokens": 230178979.0, + "step": 8896 + }, + { + "epoch": 0.9770481001537448, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.170626163482666, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.713589072227478, + "num_tokens": 230203677.0, + "step": 8897 + }, + { + "epoch": 0.9771579178563584, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.435822010040283, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7405616641044617, + "num_tokens": 230223685.0, + "step": 8898 + }, + { + "epoch": 0.9772677355589721, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3813178539276123, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7273508310317993, + "num_tokens": 230243836.0, + "step": 8899 + }, + { + "epoch": 0.9773775532615858, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2228569984436035, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6941868662834167, + "num_tokens": 230267068.0, + "step": 8900 + }, + { + "epoch": 0.9774873709641995, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1252732276916504, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6972376108169556, + "num_tokens": 230293163.0, + "step": 8901 + }, + { + "epoch": 0.9775971886668131, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9603265523910522, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6832959055900574, + "num_tokens": 230325348.0, + "step": 8902 + }, + { + "epoch": 0.9777070063694268, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.5323126316070557, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7330260276794434, + "num_tokens": 230345616.0, + "step": 8903 + }, + { + "epoch": 0.9778168240720404, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.4024922847747803, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7161411643028259, + "num_tokens": 230366751.0, + "step": 8904 + }, + { + "epoch": 0.977926641774654, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1252429485321045, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6851896643638611, + "num_tokens": 230395863.0, + "step": 8905 + }, + { + "epoch": 0.9780364594772677, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.7239208221435547, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7272806167602539, + "num_tokens": 230414203.0, + "step": 8906 + }, + { + "epoch": 0.9781462771798815, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.154466390609741, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7132922410964966, + "num_tokens": 230440005.0, + "step": 8907 + }, + { + "epoch": 0.9782560948824951, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3300654888153076, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7017869353294373, + "num_tokens": 230464148.0, + "step": 8908 + }, + { + "epoch": 0.9783659125851087, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3741278648376465, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7118769884109497, + "num_tokens": 230484594.0, + "step": 8909 + }, + { + "epoch": 0.9784757302877224, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0955734252929688, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7095073461532593, + "num_tokens": 230510260.0, + "step": 8910 + }, + { + "epoch": 0.978585547990336, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0352089405059814, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6962854862213135, + "num_tokens": 230539461.0, + "step": 8911 + }, + { + "epoch": 0.9786953656929497, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1586873531341553, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7084485292434692, + "num_tokens": 230566495.0, + "step": 8912 + }, + { + "epoch": 0.9788051833955633, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.8156754970550537, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6937283277511597, + "num_tokens": 230601093.0, + "step": 8913 + }, + { + "epoch": 0.978915001098177, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.923915147781372, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7086800336837769, + "num_tokens": 230635869.0, + "step": 8914 + }, + { + "epoch": 0.9790248188007907, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0445749759674072, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.694585919380188, + "num_tokens": 230663962.0, + "step": 8915 + }, + { + "epoch": 0.9791346365034044, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9510657787322998, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.689822793006897, + "num_tokens": 230696912.0, + "step": 8916 + }, + { + "epoch": 0.979244454206018, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.058948040008545, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7021884322166443, + "num_tokens": 230725370.0, + "step": 8917 + }, + { + "epoch": 0.9793542719086317, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1846940517425537, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7074050903320312, + "num_tokens": 230749997.0, + "step": 8918 + }, + { + "epoch": 0.9794640896112453, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1458017826080322, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7385571599006653, + "num_tokens": 230774837.0, + "step": 8919 + }, + { + "epoch": 0.979573907313859, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.300544261932373, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7224919199943542, + "num_tokens": 230799299.0, + "step": 8920 + }, + { + "epoch": 0.9796837250164726, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.337056875228882, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7041211128234863, + "num_tokens": 230822987.0, + "step": 8921 + }, + { + "epoch": 0.9797935427190864, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.5075340270996094, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7068185210227966, + "num_tokens": 230843790.0, + "step": 8922 + }, + { + "epoch": 0.9799033604217, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.242732286453247, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7217259407043457, + "num_tokens": 230866674.0, + "step": 8923 + }, + { + "epoch": 0.9800131781243137, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1359899044036865, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6961033344268799, + "num_tokens": 230896171.0, + "step": 8924 + }, + { + "epoch": 0.9801229958269273, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9451842308044434, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.692477822303772, + "num_tokens": 230927449.0, + "step": 8925 + }, + { + "epoch": 0.980232813529541, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.387490749359131, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7145041227340698, + "num_tokens": 230949584.0, + "step": 8926 + }, + { + "epoch": 0.9803426312321546, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3775057792663574, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7229514122009277, + "num_tokens": 230971160.0, + "step": 8927 + }, + { + "epoch": 0.9804524489347682, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0890390872955322, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7133429050445557, + "num_tokens": 230998463.0, + "step": 8928 + }, + { + "epoch": 0.980562266637382, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.992260217666626, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7312830686569214, + "num_tokens": 231027874.0, + "step": 8929 + }, + { + "epoch": 0.9806720843399956, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.100752830505371, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7038038969039917, + "num_tokens": 231054479.0, + "step": 8930 + }, + { + "epoch": 0.9807819020426093, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0394608974456787, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7011173963546753, + "num_tokens": 231084304.0, + "step": 8931 + }, + { + "epoch": 0.9808917197452229, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1985793113708496, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7109360694885254, + "num_tokens": 231109727.0, + "step": 8932 + }, + { + "epoch": 0.9810015374478366, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.18723726272583, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.696769118309021, + "num_tokens": 231138001.0, + "step": 8933 + }, + { + "epoch": 0.9811113551504502, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.4591727256774902, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.7051818370819092, + "num_tokens": 231161019.0, + "step": 8934 + }, + { + "epoch": 0.9812211728530639, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3883349895477295, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7205914258956909, + "num_tokens": 231182904.0, + "step": 8935 + }, + { + "epoch": 0.9813309905556776, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3738157749176025, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7134305238723755, + "num_tokens": 231205834.0, + "step": 8936 + }, + { + "epoch": 0.9814408082582913, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.351771593093872, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7174587249755859, + "num_tokens": 231229297.0, + "step": 8937 + }, + { + "epoch": 0.9815506259609049, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0846827030181885, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7165698409080505, + "num_tokens": 231253732.0, + "step": 8938 + }, + { + "epoch": 0.9816604436635186, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1425938606262207, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6999715566635132, + "num_tokens": 231281803.0, + "step": 8939 + }, + { + "epoch": 0.9817702613661322, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9494884014129639, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.688323974609375, + "num_tokens": 231312159.0, + "step": 8940 + }, + { + "epoch": 0.9818800790687459, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.102485179901123, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7178959846496582, + "num_tokens": 231339780.0, + "step": 8941 + }, + { + "epoch": 0.9819898967713595, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1875715255737305, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7271895408630371, + "num_tokens": 231365522.0, + "step": 8942 + }, + { + "epoch": 0.9820997144739732, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2067391872406006, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6965978741645813, + "num_tokens": 231392658.0, + "step": 8943 + }, + { + "epoch": 0.9822095321765869, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9946132898330688, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6983479857444763, + "num_tokens": 231421138.0, + "step": 8944 + }, + { + "epoch": 0.9823193498792006, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9928776025772095, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7011431455612183, + "num_tokens": 231449858.0, + "step": 8945 + }, + { + "epoch": 0.9824291675818142, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.3840484619140625, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7008984088897705, + "num_tokens": 231471215.0, + "step": 8946 + }, + { + "epoch": 0.9825389852844278, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2299158573150635, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7523353099822998, + "num_tokens": 231493172.0, + "step": 8947 + }, + { + "epoch": 0.9826488029870415, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.297851085662842, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.7132309079170227, + "num_tokens": 231517220.0, + "step": 8948 + }, + { + "epoch": 0.9827586206896551, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.043027400970459, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6947749853134155, + "num_tokens": 231546201.0, + "step": 8949 + }, + { + "epoch": 0.9828684383922688, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.262064218521118, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7180715799331665, + "num_tokens": 231569750.0, + "step": 8950 + }, + { + "epoch": 0.9829782560948825, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.429730176925659, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7025700211524963, + "num_tokens": 231590161.0, + "step": 8951 + }, + { + "epoch": 0.9830880737974962, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9959754943847656, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7040556073188782, + "num_tokens": 231621120.0, + "step": 8952 + }, + { + "epoch": 0.9831978915001098, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.992701768875122, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7055208086967468, + "num_tokens": 231648905.0, + "step": 8953 + }, + { + "epoch": 0.9833077092027235, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2467801570892334, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.688542902469635, + "num_tokens": 231676121.0, + "step": 8954 + }, + { + "epoch": 0.9834175269053371, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.093479633331299, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7114160656929016, + "num_tokens": 231703666.0, + "step": 8955 + }, + { + "epoch": 0.9835273446079508, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2618958950042725, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.679179310798645, + "num_tokens": 231729719.0, + "step": 8956 + }, + { + "epoch": 0.9836371623105644, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.154090404510498, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.6974107027053833, + "num_tokens": 231756062.0, + "step": 8957 + }, + { + "epoch": 0.9837469800131782, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9909818172454834, + "learning_rate": 1e-06, + "loss": 1.0977, + "mean_token_accuracy": 0.6677188873291016, + "num_tokens": 231787721.0, + "step": 8958 + }, + { + "epoch": 0.9838567977157918, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1415090560913086, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6992318034172058, + "num_tokens": 231813870.0, + "step": 8959 + }, + { + "epoch": 0.9839666154184055, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1052093505859375, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7113130688667297, + "num_tokens": 231840917.0, + "step": 8960 + }, + { + "epoch": 0.9840764331210191, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.333925247192383, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7170631289482117, + "num_tokens": 231862843.0, + "step": 8961 + }, + { + "epoch": 0.9841862508236328, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.053748607635498, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7141196727752686, + "num_tokens": 231889043.0, + "step": 8962 + }, + { + "epoch": 0.9842960685262464, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.1093087196350098, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7067102193832397, + "num_tokens": 231916725.0, + "step": 8963 + }, + { + "epoch": 0.98440588622886, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.02011775970459, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7206791639328003, + "num_tokens": 231946104.0, + "step": 8964 + }, + { + "epoch": 0.9845157039314738, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.4303882122039795, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7063027620315552, + "num_tokens": 231967785.0, + "step": 8965 + }, + { + "epoch": 0.9846255216340875, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.0257925987243652, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7039848566055298, + "num_tokens": 231994741.0, + "step": 8966 + }, + { + "epoch": 0.9847353393367011, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.4454574584960938, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.695501446723938, + "num_tokens": 232016422.0, + "step": 8967 + }, + { + "epoch": 0.9848451570393147, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.258432626724243, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.7045244574546814, + "num_tokens": 232040771.0, + "step": 8968 + }, + { + "epoch": 0.9849549747419284, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.097034215927124, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7185662984848022, + "num_tokens": 232066750.0, + "step": 8969 + }, + { + "epoch": 0.985064792444542, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 2.2062573432922363, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7093793153762817, + "num_tokens": 232090449.0, + "step": 8970 + }, + { + "epoch": 0.9851746101471557, + "ewc_loss": 7.62939453125e-05, + "grad_norm": 1.9043282270431519, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6904305815696716, + "num_tokens": 232122063.0, + "step": 8971 + }, + { + "epoch": 0.9852844278497693, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.0795516967773438, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7156789302825928, + "num_tokens": 232149959.0, + "step": 8972 + }, + { + "epoch": 0.9853942455523831, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.5136373043060303, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7284901142120361, + "num_tokens": 232169056.0, + "step": 8973 + }, + { + "epoch": 0.9855040632549967, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.1953284740448, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7089815139770508, + "num_tokens": 232194796.0, + "step": 8974 + }, + { + "epoch": 0.9856138809576104, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.0534677505493164, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.711200475692749, + "num_tokens": 232224152.0, + "step": 8975 + }, + { + "epoch": 0.985723698660224, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.2447030544281006, + "learning_rate": 1e-06, + "loss": 1.042, + "mean_token_accuracy": 0.6897157430648804, + "num_tokens": 232248823.0, + "step": 8976 + }, + { + "epoch": 0.9858335163628377, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 1.9445688724517822, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.696615993976593, + "num_tokens": 232278960.0, + "step": 8977 + }, + { + "epoch": 0.9859433340654513, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.439272880554199, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7310031056404114, + "num_tokens": 232299002.0, + "step": 8978 + }, + { + "epoch": 0.986053151768065, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.071178674697876, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.728247344493866, + "num_tokens": 232323783.0, + "step": 8979 + }, + { + "epoch": 0.9861629694706787, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.265421152114868, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.731234073638916, + "num_tokens": 232345943.0, + "step": 8980 + }, + { + "epoch": 0.9862727871732924, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.2322964668273926, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.6998537182807922, + "num_tokens": 232370383.0, + "step": 8981 + }, + { + "epoch": 0.986382604875906, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.0647988319396973, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.691228449344635, + "num_tokens": 232399776.0, + "step": 8982 + }, + { + "epoch": 0.9864924225785197, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 1.9524670839309692, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.704890251159668, + "num_tokens": 232429891.0, + "step": 8983 + }, + { + "epoch": 0.9866022402811333, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.291288137435913, + "learning_rate": 1e-06, + "loss": 1.0745, + "mean_token_accuracy": 0.6775150299072266, + "num_tokens": 232456867.0, + "step": 8984 + }, + { + "epoch": 0.986712057983747, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.192445755004883, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7324985265731812, + "num_tokens": 232481328.0, + "step": 8985 + }, + { + "epoch": 0.9868218756863606, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.2165606021881104, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.715682864189148, + "num_tokens": 232504646.0, + "step": 8986 + }, + { + "epoch": 0.9869316933889744, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 1.9966143369674683, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6987912654876709, + "num_tokens": 232535282.0, + "step": 8987 + }, + { + "epoch": 0.987041511091588, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.1917452812194824, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7098388671875, + "num_tokens": 232561180.0, + "step": 8988 + }, + { + "epoch": 0.9871513287942016, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.123647689819336, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.7090945839881897, + "num_tokens": 232588832.0, + "step": 8989 + }, + { + "epoch": 0.9872611464968153, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0791921615600586, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6973116993904114, + "num_tokens": 232616528.0, + "step": 8990 + }, + { + "epoch": 0.9873709641994289, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.190948963165283, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.701121985912323, + "num_tokens": 232643620.0, + "step": 8991 + }, + { + "epoch": 0.9874807819020426, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.0263864994049072, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.708755373954773, + "num_tokens": 232672758.0, + "step": 8992 + }, + { + "epoch": 0.9875905996046562, + "ewc_loss": 7.677078247070312e-05, + "grad_norm": 2.2100348472595215, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7033634185791016, + "num_tokens": 232698818.0, + "step": 8993 + }, + { + "epoch": 0.98770041730727, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1438567638397217, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.702110767364502, + "num_tokens": 232723684.0, + "step": 8994 + }, + { + "epoch": 0.9878102350098836, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1651148796081543, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7196933031082153, + "num_tokens": 232748578.0, + "step": 8995 + }, + { + "epoch": 0.9879200527124973, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0429582595825195, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7115596532821655, + "num_tokens": 232777194.0, + "step": 8996 + }, + { + "epoch": 0.9880298704151109, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.9611352682113647, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6974567174911499, + "num_tokens": 232806509.0, + "step": 8997 + }, + { + "epoch": 0.9881396881177246, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3596887588500977, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.6956514120101929, + "num_tokens": 232829602.0, + "step": 8998 + }, + { + "epoch": 0.9882495058203382, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3573927879333496, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7099916338920593, + "num_tokens": 232853101.0, + "step": 8999 + }, + { + "epoch": 0.9883593235229519, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.266435384750366, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7167996168136597, + "num_tokens": 232876847.0, + "step": 9000 + }, + { + "epoch": 0.9884691412255655, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.426938772201538, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7286180257797241, + "num_tokens": 232897886.0, + "step": 9001 + }, + { + "epoch": 0.9885789589281793, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1934871673583984, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6953619122505188, + "num_tokens": 232923874.0, + "step": 9002 + }, + { + "epoch": 0.9886887766307929, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.204055070877075, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7104328870773315, + "num_tokens": 232948175.0, + "step": 9003 + }, + { + "epoch": 0.9887985943334066, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3598005771636963, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.6983739733695984, + "num_tokens": 232970566.0, + "step": 9004 + }, + { + "epoch": 0.9889084120360202, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.365262746810913, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6946545839309692, + "num_tokens": 232993237.0, + "step": 9005 + }, + { + "epoch": 0.9890182297386338, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.881759762763977, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.6921612024307251, + "num_tokens": 233027525.0, + "step": 9006 + }, + { + "epoch": 0.9891280474412475, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.9285025596618652, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7030588388442993, + "num_tokens": 233061174.0, + "step": 9007 + }, + { + "epoch": 0.9892378651438611, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.229700803756714, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6973645687103271, + "num_tokens": 233087169.0, + "step": 9008 + }, + { + "epoch": 0.9893476828464749, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1641018390655518, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6991768479347229, + "num_tokens": 233113169.0, + "step": 9009 + }, + { + "epoch": 0.9894575005490885, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.9182426929473877, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7046820521354675, + "num_tokens": 233145830.0, + "step": 9010 + }, + { + "epoch": 0.9895673182517022, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.090569257736206, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6960205435752869, + "num_tokens": 233172642.0, + "step": 9011 + }, + { + "epoch": 0.9896771359543158, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.2874338626861572, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7250967025756836, + "num_tokens": 233194751.0, + "step": 9012 + }, + { + "epoch": 0.9897869536569295, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.8977757692337036, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.704208493232727, + "num_tokens": 233225997.0, + "step": 9013 + }, + { + "epoch": 0.9898967713595431, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.09521222114563, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.705852210521698, + "num_tokens": 233252009.0, + "step": 9014 + }, + { + "epoch": 0.9900065890621568, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.105309247970581, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7227809429168701, + "num_tokens": 233278741.0, + "step": 9015 + }, + { + "epoch": 0.9901164067647705, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3396236896514893, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7193320989608765, + "num_tokens": 233298654.0, + "step": 9016 + }, + { + "epoch": 0.9902262244673842, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.013972282409668, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7100573182106018, + "num_tokens": 233327394.0, + "step": 9017 + }, + { + "epoch": 0.9903360421699978, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.152831554412842, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7109533548355103, + "num_tokens": 233352462.0, + "step": 9018 + }, + { + "epoch": 0.9904458598726115, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.35488224029541, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7178663611412048, + "num_tokens": 233374329.0, + "step": 9019 + }, + { + "epoch": 0.9905556775752251, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0657975673675537, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7186352014541626, + "num_tokens": 233399704.0, + "step": 9020 + }, + { + "epoch": 0.9906654952778388, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3391263484954834, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.70456463098526, + "num_tokens": 233422103.0, + "step": 9021 + }, + { + "epoch": 0.9907753129804524, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3912816047668457, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7089856863021851, + "num_tokens": 233444041.0, + "step": 9022 + }, + { + "epoch": 0.9908851306830662, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0679972171783447, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7157822847366333, + "num_tokens": 233472497.0, + "step": 9023 + }, + { + "epoch": 0.9909949483856798, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3082258701324463, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7083420157432556, + "num_tokens": 233495761.0, + "step": 9024 + }, + { + "epoch": 0.9911047660882935, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0996694564819336, + "learning_rate": 1e-06, + "loss": 1.0605, + "mean_token_accuracy": 0.6796519756317139, + "num_tokens": 233526528.0, + "step": 9025 + }, + { + "epoch": 0.9912145837909071, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2262678146362305, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7122111320495605, + "num_tokens": 233549552.0, + "step": 9026 + }, + { + "epoch": 0.9913244014935207, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1511528491973877, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7075754404067993, + "num_tokens": 233575314.0, + "step": 9027 + }, + { + "epoch": 0.9914342191961344, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1315512657165527, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7036744952201843, + "num_tokens": 233600910.0, + "step": 9028 + }, + { + "epoch": 0.991544036898748, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.2357521057128906, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.700912356376648, + "num_tokens": 233624970.0, + "step": 9029 + }, + { + "epoch": 0.9916538546013618, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.964009404182434, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.7005106210708618, + "num_tokens": 233656091.0, + "step": 9030 + }, + { + "epoch": 0.9917636723039754, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3106672763824463, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.714282214641571, + "num_tokens": 233678512.0, + "step": 9031 + }, + { + "epoch": 0.9918734900065891, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1418585777282715, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7096536159515381, + "num_tokens": 233704187.0, + "step": 9032 + }, + { + "epoch": 0.9919833077092027, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.1240999698638916, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7037632465362549, + "num_tokens": 233731448.0, + "step": 9033 + }, + { + "epoch": 0.9920931254118164, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.24841046333313, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7126302719116211, + "num_tokens": 233755151.0, + "step": 9034 + }, + { + "epoch": 0.99220294311443, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.8225269317626953, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7126030325889587, + "num_tokens": 233789917.0, + "step": 9035 + }, + { + "epoch": 0.9923127608170437, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.3361265659332275, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7137671113014221, + "num_tokens": 233812965.0, + "step": 9036 + }, + { + "epoch": 0.9924225785196573, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.055849313735962, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7072433233261108, + "num_tokens": 233842025.0, + "step": 9037 + }, + { + "epoch": 0.9925323962222711, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0484001636505127, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7186486124992371, + "num_tokens": 233870113.0, + "step": 9038 + }, + { + "epoch": 0.9926422139248847, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.096436023712158, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.710200309753418, + "num_tokens": 233897948.0, + "step": 9039 + }, + { + "epoch": 0.9927520316274984, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.213451623916626, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6959471702575684, + "num_tokens": 233922903.0, + "step": 9040 + }, + { + "epoch": 0.992861849330112, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.110062837600708, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7334986925125122, + "num_tokens": 233949903.0, + "step": 9041 + }, + { + "epoch": 0.9929716670327257, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.0256130695343018, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7004680633544922, + "num_tokens": 233980714.0, + "step": 9042 + }, + { + "epoch": 0.9930814847353393, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 1.9130884408950806, + "learning_rate": 1e-06, + "loss": 1.0586, + "mean_token_accuracy": 0.6868646144866943, + "num_tokens": 234011634.0, + "step": 9043 + }, + { + "epoch": 0.993191302437953, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.321462631225586, + "learning_rate": 1e-06, + "loss": 1.0375, + "mean_token_accuracy": 0.6878840923309326, + "num_tokens": 234035125.0, + "step": 9044 + }, + { + "epoch": 0.9933011201405667, + "ewc_loss": 7.724761962890625e-05, + "grad_norm": 2.356501579284668, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.7003161907196045, + "num_tokens": 234057417.0, + "step": 9045 + }, + { + "epoch": 0.9934109378431804, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.5065758228302, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6967324018478394, + "num_tokens": 234078743.0, + "step": 9046 + }, + { + "epoch": 0.993520755545794, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1161277294158936, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7297791242599487, + "num_tokens": 234105817.0, + "step": 9047 + }, + { + "epoch": 0.9936305732484076, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.062267780303955, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6934747099876404, + "num_tokens": 234132511.0, + "step": 9048 + }, + { + "epoch": 0.9937403909510213, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1973345279693604, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7228823900222778, + "num_tokens": 234155713.0, + "step": 9049 + }, + { + "epoch": 0.9938502086536349, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.953757643699646, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7122764587402344, + "num_tokens": 234184946.0, + "step": 9050 + }, + { + "epoch": 0.9939600263562486, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3402810096740723, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7140092849731445, + "num_tokens": 234206926.0, + "step": 9051 + }, + { + "epoch": 0.9940698440588623, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.167543888092041, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7162480354309082, + "num_tokens": 234231917.0, + "step": 9052 + }, + { + "epoch": 0.994179661761476, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1498687267303467, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6976066827774048, + "num_tokens": 234257913.0, + "step": 9053 + }, + { + "epoch": 0.9942894794640896, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3575360774993896, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.724963366985321, + "num_tokens": 234281030.0, + "step": 9054 + }, + { + "epoch": 0.9943992971667033, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2204370498657227, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7136334180831909, + "num_tokens": 234304409.0, + "step": 9055 + }, + { + "epoch": 0.9945091148693169, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9806946516036987, + "learning_rate": 1e-06, + "loss": 1.0738, + "mean_token_accuracy": 0.6731841564178467, + "num_tokens": 234336413.0, + "step": 9056 + }, + { + "epoch": 0.9946189325719306, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.194920063018799, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7171042561531067, + "num_tokens": 234360815.0, + "step": 9057 + }, + { + "epoch": 0.9947287502745442, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1731960773468018, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7146457433700562, + "num_tokens": 234389038.0, + "step": 9058 + }, + { + "epoch": 0.994838567977158, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3904595375061035, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7004491090774536, + "num_tokens": 234410875.0, + "step": 9059 + }, + { + "epoch": 0.9949483856797716, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1748573780059814, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.723423421382904, + "num_tokens": 234436468.0, + "step": 9060 + }, + { + "epoch": 0.9950582033823853, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.148813486099243, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7112036943435669, + "num_tokens": 234462935.0, + "step": 9061 + }, + { + "epoch": 0.9951680210849989, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9009054899215698, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7112074494361877, + "num_tokens": 234496078.0, + "step": 9062 + }, + { + "epoch": 0.9952778387876126, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.15448260307312, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6904411315917969, + "num_tokens": 234524273.0, + "step": 9063 + }, + { + "epoch": 0.9953876564902262, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2077903747558594, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7209487557411194, + "num_tokens": 234549620.0, + "step": 9064 + }, + { + "epoch": 0.9954974741928398, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0445754528045654, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7169029712677002, + "num_tokens": 234576957.0, + "step": 9065 + }, + { + "epoch": 0.9956072918954535, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.198429584503174, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7010981440544128, + "num_tokens": 234601466.0, + "step": 9066 + }, + { + "epoch": 0.9957171095980673, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.931761622428894, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7291930913925171, + "num_tokens": 234632198.0, + "step": 9067 + }, + { + "epoch": 0.9958269273006809, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.963740348815918, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7134535908699036, + "num_tokens": 234662027.0, + "step": 9068 + }, + { + "epoch": 0.9959367450032945, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.037585973739624, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7126131057739258, + "num_tokens": 234688572.0, + "step": 9069 + }, + { + "epoch": 0.9960465627059082, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.185094118118286, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6880398988723755, + "num_tokens": 234716877.0, + "step": 9070 + }, + { + "epoch": 0.9961563804085218, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.966340184211731, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7016136646270752, + "num_tokens": 234746332.0, + "step": 9071 + }, + { + "epoch": 0.9962661981111355, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2594361305236816, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7139507532119751, + "num_tokens": 234769472.0, + "step": 9072 + }, + { + "epoch": 0.9963760158137491, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1869990825653076, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7029882669448853, + "num_tokens": 234794215.0, + "step": 9073 + }, + { + "epoch": 0.9964858335163629, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.981591820716858, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7146790027618408, + "num_tokens": 234821761.0, + "step": 9074 + }, + { + "epoch": 0.9965956512189765, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.139575958251953, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7158528566360474, + "num_tokens": 234848284.0, + "step": 9075 + }, + { + "epoch": 0.9967054689215902, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.60062313079834, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.716270387172699, + "num_tokens": 234866900.0, + "step": 9076 + }, + { + "epoch": 0.9968152866242038, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2521321773529053, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7200800180435181, + "num_tokens": 234890696.0, + "step": 9077 + }, + { + "epoch": 0.9969251043268175, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.6076626777648926, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7155474424362183, + "num_tokens": 234909194.0, + "step": 9078 + }, + { + "epoch": 0.9970349220294311, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9826894998550415, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.7001546025276184, + "num_tokens": 234938499.0, + "step": 9079 + }, + { + "epoch": 0.9971447397320448, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.688243865966797, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7065953016281128, + "num_tokens": 234956871.0, + "step": 9080 + }, + { + "epoch": 0.9972545574346585, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1409473419189453, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7230430245399475, + "num_tokens": 234983142.0, + "step": 9081 + }, + { + "epoch": 0.9973643751372722, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9177485704421997, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.702512264251709, + "num_tokens": 235015256.0, + "step": 9082 + }, + { + "epoch": 0.9974741928398858, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1159398555755615, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7272292375564575, + "num_tokens": 235039088.0, + "step": 9083 + }, + { + "epoch": 0.9975840105424995, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4851584434509277, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.7040279507637024, + "num_tokens": 235059458.0, + "step": 9084 + }, + { + "epoch": 0.9976938282451131, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.294434070587158, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7221870422363281, + "num_tokens": 235080777.0, + "step": 9085 + }, + { + "epoch": 0.9978036459477267, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0507726669311523, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7307009100914001, + "num_tokens": 235106971.0, + "step": 9086 + }, + { + "epoch": 0.9979134636503404, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.6188528537750244, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7161188125610352, + "num_tokens": 235126459.0, + "step": 9087 + }, + { + "epoch": 0.9980232813529542, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0088655948638916, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7053523063659668, + "num_tokens": 235154795.0, + "step": 9088 + }, + { + "epoch": 0.9981330990555678, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.396364688873291, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.69992595911026, + "num_tokens": 235177941.0, + "step": 9089 + }, + { + "epoch": 0.9982429167581814, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0842947959899902, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.6995928287506104, + "num_tokens": 235205466.0, + "step": 9090 + }, + { + "epoch": 0.9983527344607951, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3172032833099365, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7038456201553345, + "num_tokens": 235228890.0, + "step": 9091 + }, + { + "epoch": 0.9984625521634087, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.392988443374634, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7214678525924683, + "num_tokens": 235250592.0, + "step": 9092 + }, + { + "epoch": 0.9985723698660224, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0017855167388916, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7192431688308716, + "num_tokens": 235278580.0, + "step": 9093 + }, + { + "epoch": 0.998682187568636, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2648024559020996, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7222630977630615, + "num_tokens": 235301987.0, + "step": 9094 + }, + { + "epoch": 0.9987920052712497, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2815990447998047, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6995396614074707, + "num_tokens": 235328062.0, + "step": 9095 + }, + { + "epoch": 0.9989018229738634, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3341875076293945, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7284945249557495, + "num_tokens": 235350562.0, + "step": 9096 + }, + { + "epoch": 0.9990116406764771, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9935492277145386, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7079698443412781, + "num_tokens": 235379626.0, + "step": 9097 + }, + { + "epoch": 0.9991214583790907, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.245018243789673, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7110093832015991, + "num_tokens": 235405941.0, + "step": 9098 + }, + { + "epoch": 0.9992312760817044, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0916244983673096, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7088395357131958, + "num_tokens": 235432729.0, + "step": 9099 + }, + { + "epoch": 0.999341093784318, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9793542623519897, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.7022194862365723, + "num_tokens": 235463599.0, + "step": 9100 + }, + { + "epoch": 0.9994509114869317, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2931883335113525, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7081968784332275, + "num_tokens": 235486921.0, + "step": 9101 + }, + { + "epoch": 0.9995607291895453, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.8778610229492188, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6931626796722412, + "num_tokens": 235519812.0, + "step": 9102 + }, + { + "epoch": 0.9996705468921591, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4284212589263916, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7269026041030884, + "num_tokens": 235543456.0, + "step": 9103 + }, + { + "epoch": 0.9997803645947727, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0602540969848633, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6956310272216797, + "num_tokens": 235572016.0, + "step": 9104 + }, + { + "epoch": 0.9998901822973864, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.386622667312622, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.720072865486145, + "num_tokens": 235594912.0, + "step": 9105 + }, + { + "epoch": 1.0, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.34403395652771, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7073003053665161, + "num_tokens": 235617245.0, + "step": 9106 + }, + { + "epoch": 1.0001098177026138, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0890705585479736, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7015252113342285, + "num_tokens": 235647042.0, + "step": 9107 + }, + { + "epoch": 1.0002196354052273, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0169425010681152, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7186342477798462, + "num_tokens": 235675620.0, + "step": 9108 + }, + { + "epoch": 1.000329453107841, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3598928451538086, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7139196395874023, + "num_tokens": 235699190.0, + "step": 9109 + }, + { + "epoch": 1.0004392708104546, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.868664026260376, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7153533697128296, + "num_tokens": 235729568.0, + "step": 9110 + }, + { + "epoch": 1.0005490885130683, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.133013963699341, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7261967658996582, + "num_tokens": 235753975.0, + "step": 9111 + }, + { + "epoch": 1.0006589062156819, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3019275665283203, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7275508642196655, + "num_tokens": 235774114.0, + "step": 9112 + }, + { + "epoch": 1.0007687239182956, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.141806125640869, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6957064867019653, + "num_tokens": 235799505.0, + "step": 9113 + }, + { + "epoch": 1.0008785416209094, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.111530065536499, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7195385694503784, + "num_tokens": 235828000.0, + "step": 9114 + }, + { + "epoch": 1.000988359323523, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4225168228149414, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7458585500717163, + "num_tokens": 235851026.0, + "step": 9115 + }, + { + "epoch": 1.0010981770261367, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2561357021331787, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.685488224029541, + "num_tokens": 235880806.0, + "step": 9116 + }, + { + "epoch": 1.0012079947287502, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.1788790225982666, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7025350332260132, + "num_tokens": 235913442.0, + "step": 9117 + }, + { + "epoch": 1.001317812431364, + "ewc_loss": 7.82012939453125e-05, + "grad_norm": 2.3193230628967285, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7303279638290405, + "num_tokens": 235936145.0, + "step": 9118 + }, + { + "epoch": 1.0014276301339775, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.4066004753112793, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7220230102539062, + "num_tokens": 235958222.0, + "step": 9119 + }, + { + "epoch": 1.0015374478365913, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3643710613250732, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7301634550094604, + "num_tokens": 235980934.0, + "step": 9120 + }, + { + "epoch": 1.001647265539205, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.286726236343384, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7082527875900269, + "num_tokens": 236005418.0, + "step": 9121 + }, + { + "epoch": 1.0017570832418186, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.204103946685791, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7244277596473694, + "num_tokens": 236031486.0, + "step": 9122 + }, + { + "epoch": 1.0018669009444323, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2884984016418457, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6956775188446045, + "num_tokens": 236058222.0, + "step": 9123 + }, + { + "epoch": 1.0019767186470458, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.222501516342163, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7342345118522644, + "num_tokens": 236082865.0, + "step": 9124 + }, + { + "epoch": 1.0020865363496596, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9092249870300293, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6951675415039062, + "num_tokens": 236118019.0, + "step": 9125 + }, + { + "epoch": 1.0021963540522731, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3464407920837402, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7177712917327881, + "num_tokens": 236140259.0, + "step": 9126 + }, + { + "epoch": 1.002306171754887, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9596314430236816, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7056630849838257, + "num_tokens": 236171193.0, + "step": 9127 + }, + { + "epoch": 1.0024159894575007, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.17101788520813, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7155424356460571, + "num_tokens": 236198553.0, + "step": 9128 + }, + { + "epoch": 1.0025258071601142, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2825570106506348, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.6999112963676453, + "num_tokens": 236225340.0, + "step": 9129 + }, + { + "epoch": 1.002635624862728, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.32680082321167, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7205838561058044, + "num_tokens": 236249433.0, + "step": 9130 + }, + { + "epoch": 1.0027454425653415, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0522425174713135, + "learning_rate": 1e-06, + "loss": 1.03, + "mean_token_accuracy": 0.6901458501815796, + "num_tokens": 236278592.0, + "step": 9131 + }, + { + "epoch": 1.0028552602679552, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.6007754802703857, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7377744317054749, + "num_tokens": 236296550.0, + "step": 9132 + }, + { + "epoch": 1.0029650779705688, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0584237575531006, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7128525972366333, + "num_tokens": 236325169.0, + "step": 9133 + }, + { + "epoch": 1.0030748956731825, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.0122110843658447, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6967008709907532, + "num_tokens": 236355195.0, + "step": 9134 + }, + { + "epoch": 1.0031847133757963, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.8807390928268433, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.696540892124176, + "num_tokens": 236392651.0, + "step": 9135 + }, + { + "epoch": 1.0032945310784098, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.8704657554626465, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7089563608169556, + "num_tokens": 236429123.0, + "step": 9136 + }, + { + "epoch": 1.0034043487810236, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2140133380889893, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7244015336036682, + "num_tokens": 236458505.0, + "step": 9137 + }, + { + "epoch": 1.0035141664836371, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2818984985351562, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7238710522651672, + "num_tokens": 236483365.0, + "step": 9138 + }, + { + "epoch": 1.0036239841862509, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.171804189682007, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7190999984741211, + "num_tokens": 236507918.0, + "step": 9139 + }, + { + "epoch": 1.0037338018888644, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.788399577140808, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7065150737762451, + "num_tokens": 236542471.0, + "step": 9140 + }, + { + "epoch": 1.0038436195914782, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.2757060527801514, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7287731170654297, + "num_tokens": 236566275.0, + "step": 9141 + }, + { + "epoch": 1.0039534372940917, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.3660922050476074, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7243814468383789, + "num_tokens": 236588724.0, + "step": 9142 + }, + { + "epoch": 1.0040632549967055, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 1.9882657527923584, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6937559843063354, + "num_tokens": 236619468.0, + "step": 9143 + }, + { + "epoch": 1.0041730726993192, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4575259685516357, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7163153886795044, + "num_tokens": 236641105.0, + "step": 9144 + }, + { + "epoch": 1.0042828904019327, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4274046421051025, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.723476231098175, + "num_tokens": 236664417.0, + "step": 9145 + }, + { + "epoch": 1.0043927081045465, + "ewc_loss": 7.772445678710938e-05, + "grad_norm": 2.4354748725891113, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.722664475440979, + "num_tokens": 236685394.0, + "step": 9146 + }, + { + "epoch": 1.00450252580716, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3835082054138184, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7365493774414062, + "num_tokens": 236709099.0, + "step": 9147 + }, + { + "epoch": 1.0046123435097738, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.0713634490966797, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.7035636901855469, + "num_tokens": 236736328.0, + "step": 9148 + }, + { + "epoch": 1.0047221612123873, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1657419204711914, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7250015735626221, + "num_tokens": 236762359.0, + "step": 9149 + }, + { + "epoch": 1.004831978915001, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1333506107330322, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7030388116836548, + "num_tokens": 236788664.0, + "step": 9150 + }, + { + "epoch": 1.0049417966176148, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.272735118865967, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7259711027145386, + "num_tokens": 236811440.0, + "step": 9151 + }, + { + "epoch": 1.0050516143202284, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 1.9657589197158813, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7018911838531494, + "num_tokens": 236844805.0, + "step": 9152 + }, + { + "epoch": 1.0051614320228421, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.23578143119812, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7287642359733582, + "num_tokens": 236869318.0, + "step": 9153 + }, + { + "epoch": 1.0052712497254557, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 1.8819754123687744, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7407160997390747, + "num_tokens": 236901039.0, + "step": 9154 + }, + { + "epoch": 1.0053810674280694, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1848092079162598, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7002780437469482, + "num_tokens": 236928639.0, + "step": 9155 + }, + { + "epoch": 1.005490885130683, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.2595832347869873, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6985186338424683, + "num_tokens": 236955654.0, + "step": 9156 + }, + { + "epoch": 1.0056007028332967, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.2275896072387695, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.729767918586731, + "num_tokens": 236980378.0, + "step": 9157 + }, + { + "epoch": 1.0057105205359105, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1854190826416016, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7045198678970337, + "num_tokens": 237008385.0, + "step": 9158 + }, + { + "epoch": 1.005820338238524, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.185593843460083, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.70978844165802, + "num_tokens": 237034555.0, + "step": 9159 + }, + { + "epoch": 1.0059301559411378, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.095357894897461, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7059136033058167, + "num_tokens": 237066324.0, + "step": 9160 + }, + { + "epoch": 1.0060399736437513, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.0208346843719482, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7061175107955933, + "num_tokens": 237094989.0, + "step": 9161 + }, + { + "epoch": 1.006149791346365, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3014309406280518, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.6961483955383301, + "num_tokens": 237120765.0, + "step": 9162 + }, + { + "epoch": 1.0062596090489786, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 1.9932364225387573, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.708901047706604, + "num_tokens": 237151511.0, + "step": 9163 + }, + { + "epoch": 1.0063694267515924, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3566529750823975, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7118378281593323, + "num_tokens": 237176336.0, + "step": 9164 + }, + { + "epoch": 1.0064792444542061, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1106116771698, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6926122903823853, + "num_tokens": 237207576.0, + "step": 9165 + }, + { + "epoch": 1.0065890621568196, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.145622491836548, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7114192843437195, + "num_tokens": 237235137.0, + "step": 9166 + }, + { + "epoch": 1.0066988798594334, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.5084197521209717, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7237517833709717, + "num_tokens": 237255124.0, + "step": 9167 + }, + { + "epoch": 1.006808697562047, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.486536979675293, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.69587242603302, + "num_tokens": 237278364.0, + "step": 9168 + }, + { + "epoch": 1.0069185152646607, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3085620403289795, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7317407727241516, + "num_tokens": 237302500.0, + "step": 9169 + }, + { + "epoch": 1.0070283329672742, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.2154359817504883, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7315241694450378, + "num_tokens": 237327908.0, + "step": 9170 + }, + { + "epoch": 1.007138150669888, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1895928382873535, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7074184417724609, + "num_tokens": 237355523.0, + "step": 9171 + }, + { + "epoch": 1.0072479683725017, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1492207050323486, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.711763858795166, + "num_tokens": 237381925.0, + "step": 9172 + }, + { + "epoch": 1.0073577860751153, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.16898512840271, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7154147624969482, + "num_tokens": 237408436.0, + "step": 9173 + }, + { + "epoch": 1.007467603777729, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.249215841293335, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7195535898208618, + "num_tokens": 237434340.0, + "step": 9174 + }, + { + "epoch": 1.0075774214803426, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.146430730819702, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7422652840614319, + "num_tokens": 237459680.0, + "step": 9175 + }, + { + "epoch": 1.0076872391829563, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 1.9599272012710571, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.6924872398376465, + "num_tokens": 237492224.0, + "step": 9176 + }, + { + "epoch": 1.0077970568855699, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1570520401000977, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7451282143592834, + "num_tokens": 237516302.0, + "step": 9177 + }, + { + "epoch": 1.0079068745881836, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.2659363746643066, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7133055925369263, + "num_tokens": 237540954.0, + "step": 9178 + }, + { + "epoch": 1.0080166922907974, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.2275631427764893, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7134695649147034, + "num_tokens": 237565679.0, + "step": 9179 + }, + { + "epoch": 1.008126509993411, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.0211832523345947, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7042202949523926, + "num_tokens": 237592971.0, + "step": 9180 + }, + { + "epoch": 1.0082363276960247, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.169523000717163, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7211791276931763, + "num_tokens": 237618973.0, + "step": 9181 + }, + { + "epoch": 1.0083461453986382, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.0769095420837402, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7086993455886841, + "num_tokens": 237649304.0, + "step": 9182 + }, + { + "epoch": 1.008455963101252, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.384627103805542, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6803106665611267, + "num_tokens": 237676775.0, + "step": 9183 + }, + { + "epoch": 1.0085657808038655, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.010906934738159, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7048614025115967, + "num_tokens": 237707669.0, + "step": 9184 + }, + { + "epoch": 1.0086755985064793, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3114709854125977, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7314629554748535, + "num_tokens": 237732335.0, + "step": 9185 + }, + { + "epoch": 1.008785416209093, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.4273412227630615, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7153894305229187, + "num_tokens": 237755742.0, + "step": 9186 + }, + { + "epoch": 1.0088952339117065, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3346917629241943, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7180580496788025, + "num_tokens": 237779404.0, + "step": 9187 + }, + { + "epoch": 1.0090050516143203, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.4543464183807373, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7276906967163086, + "num_tokens": 237801424.0, + "step": 9188 + }, + { + "epoch": 1.0091148693169338, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.4245307445526123, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7135491967201233, + "num_tokens": 237824471.0, + "step": 9189 + }, + { + "epoch": 1.0092246870195476, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 1.997828722000122, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6916265487670898, + "num_tokens": 237856973.0, + "step": 9190 + }, + { + "epoch": 1.0093345047221611, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.1093947887420654, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7085369825363159, + "num_tokens": 237886112.0, + "step": 9191 + }, + { + "epoch": 1.0094443224247749, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.3190267086029053, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.6872552037239075, + "num_tokens": 237912563.0, + "step": 9192 + }, + { + "epoch": 1.0095541401273886, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.5220284461975098, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7279377579689026, + "num_tokens": 237931329.0, + "step": 9193 + }, + { + "epoch": 1.0096639578300022, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.849025011062622, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7252610325813293, + "num_tokens": 237949637.0, + "step": 9194 + }, + { + "epoch": 1.009773775532616, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.682365894317627, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7412233948707581, + "num_tokens": 237968165.0, + "step": 9195 + }, + { + "epoch": 1.0098835932352295, + "ewc_loss": 7.867813110351562e-05, + "grad_norm": 2.296605348587036, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7084973454475403, + "num_tokens": 237993937.0, + "step": 9196 + }, + { + "epoch": 1.0099934109378432, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.272246837615967, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7201763391494751, + "num_tokens": 238017976.0, + "step": 9197 + }, + { + "epoch": 1.0101032286404568, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 1.9373151063919067, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.69617760181427, + "num_tokens": 238049863.0, + "step": 9198 + }, + { + "epoch": 1.0102130463430705, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.084092140197754, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7294003367424011, + "num_tokens": 238079151.0, + "step": 9199 + }, + { + "epoch": 1.010322864045684, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.2598488330841064, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7272497415542603, + "num_tokens": 238103439.0, + "step": 9200 + }, + { + "epoch": 1.0104326817482978, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.162923812866211, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7045891284942627, + "num_tokens": 238132132.0, + "step": 9201 + }, + { + "epoch": 1.0105424994509116, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.5587167739868164, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7332421541213989, + "num_tokens": 238152749.0, + "step": 9202 + }, + { + "epoch": 1.010652317153525, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.0371973514556885, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7282993197441101, + "num_tokens": 238180616.0, + "step": 9203 + }, + { + "epoch": 1.0107621348561389, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.4146788120269775, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7104517221450806, + "num_tokens": 238205218.0, + "step": 9204 + }, + { + "epoch": 1.0108719525587524, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.4418411254882812, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.719050407409668, + "num_tokens": 238228805.0, + "step": 9205 + }, + { + "epoch": 1.0109817702613662, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.0784332752227783, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7047514915466309, + "num_tokens": 238258713.0, + "step": 9206 + }, + { + "epoch": 1.0110915879639797, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.4623358249664307, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.756607711315155, + "num_tokens": 238277651.0, + "step": 9207 + }, + { + "epoch": 1.0112014056665934, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.092285394668579, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.6994726061820984, + "num_tokens": 238306098.0, + "step": 9208 + }, + { + "epoch": 1.0113112233692072, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.399784803390503, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7334561944007874, + "num_tokens": 238327071.0, + "step": 9209 + }, + { + "epoch": 1.0114210410718207, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.2051377296447754, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7050833106040955, + "num_tokens": 238353322.0, + "step": 9210 + }, + { + "epoch": 1.0115308587744345, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.032503843307495, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7229679822921753, + "num_tokens": 238382018.0, + "step": 9211 + }, + { + "epoch": 1.011640676477048, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.4893152713775635, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7334996461868286, + "num_tokens": 238401017.0, + "step": 9212 + }, + { + "epoch": 1.0117504941796618, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.211672067642212, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7201142311096191, + "num_tokens": 238428954.0, + "step": 9213 + }, + { + "epoch": 1.0118603118822753, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.1406960487365723, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7357598543167114, + "num_tokens": 238455499.0, + "step": 9214 + }, + { + "epoch": 1.011970129584889, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.157029628753662, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7180997729301453, + "num_tokens": 238481219.0, + "step": 9215 + }, + { + "epoch": 1.0120799472875028, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.135172128677368, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7383054494857788, + "num_tokens": 238506894.0, + "step": 9216 + }, + { + "epoch": 1.0121897649901164, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.260848045349121, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7231263518333435, + "num_tokens": 238534255.0, + "step": 9217 + }, + { + "epoch": 1.0122995826927301, + "ewc_loss": 7.915496826171875e-05, + "grad_norm": 2.5151660442352295, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7239835262298584, + "num_tokens": 238555269.0, + "step": 9218 + }, + { + "epoch": 1.0124094003953437, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 1.9377487897872925, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7065504193305969, + "num_tokens": 238588256.0, + "step": 9219 + }, + { + "epoch": 1.0125192180979574, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 1.875536561012268, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7054171562194824, + "num_tokens": 238622800.0, + "step": 9220 + }, + { + "epoch": 1.012629035800571, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.4062435626983643, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7267690896987915, + "num_tokens": 238645351.0, + "step": 9221 + }, + { + "epoch": 1.0127388535031847, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.3534743785858154, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7044885158538818, + "num_tokens": 238668786.0, + "step": 9222 + }, + { + "epoch": 1.0128486712057985, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.1436848640441895, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7275152206420898, + "num_tokens": 238694597.0, + "step": 9223 + }, + { + "epoch": 1.012958488908412, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.275172233581543, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7291735410690308, + "num_tokens": 238719112.0, + "step": 9224 + }, + { + "epoch": 1.0130683066110258, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 1.8452128171920776, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7222370505332947, + "num_tokens": 238751808.0, + "step": 9225 + }, + { + "epoch": 1.0131781243136393, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.3058900833129883, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7248988151550293, + "num_tokens": 238776031.0, + "step": 9226 + }, + { + "epoch": 1.013287942016253, + "ewc_loss": 7.963180541992188e-05, + "grad_norm": 2.4979355335235596, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7159653902053833, + "num_tokens": 238796867.0, + "step": 9227 + }, + { + "epoch": 1.0133977597188666, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2989351749420166, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7096831798553467, + "num_tokens": 238822315.0, + "step": 9228 + }, + { + "epoch": 1.0135075774214803, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2495479583740234, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7101670503616333, + "num_tokens": 238847142.0, + "step": 9229 + }, + { + "epoch": 1.013617395124094, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.079428195953369, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.707467794418335, + "num_tokens": 238877845.0, + "step": 9230 + }, + { + "epoch": 1.0137272128267076, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.162977695465088, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7010836601257324, + "num_tokens": 238904482.0, + "step": 9231 + }, + { + "epoch": 1.0138370305293214, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.355508327484131, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7038900852203369, + "num_tokens": 238928874.0, + "step": 9232 + }, + { + "epoch": 1.013946848231935, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.372138023376465, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.706697940826416, + "num_tokens": 238955467.0, + "step": 9233 + }, + { + "epoch": 1.0140566659345487, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.4268815517425537, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7384806871414185, + "num_tokens": 238977613.0, + "step": 9234 + }, + { + "epoch": 1.0141664836371622, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.4076619148254395, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.704675555229187, + "num_tokens": 239002498.0, + "step": 9235 + }, + { + "epoch": 1.014276301339776, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 1.9672514200210571, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7026548385620117, + "num_tokens": 239038191.0, + "step": 9236 + }, + { + "epoch": 1.0143861190423897, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.066082239151001, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6900502443313599, + "num_tokens": 239067476.0, + "step": 9237 + }, + { + "epoch": 1.0144959367450033, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.227886199951172, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7193599939346313, + "num_tokens": 239094249.0, + "step": 9238 + }, + { + "epoch": 1.014605754447617, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2671396732330322, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7154159545898438, + "num_tokens": 239119765.0, + "step": 9239 + }, + { + "epoch": 1.0147155721502306, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.058089256286621, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7145562767982483, + "num_tokens": 239147445.0, + "step": 9240 + }, + { + "epoch": 1.0148253898528443, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 3.279423713684082, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7269498705863953, + "num_tokens": 239169710.0, + "step": 9241 + }, + { + "epoch": 1.0149352075554579, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.4273905754089355, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.731336236000061, + "num_tokens": 239191689.0, + "step": 9242 + }, + { + "epoch": 1.0150450252580716, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.1421029567718506, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.6994255185127258, + "num_tokens": 239220203.0, + "step": 9243 + }, + { + "epoch": 1.0151548429606854, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.241002082824707, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.739303708076477, + "num_tokens": 239247003.0, + "step": 9244 + }, + { + "epoch": 1.015264660663299, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.496816396713257, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7341321706771851, + "num_tokens": 239266673.0, + "step": 9245 + }, + { + "epoch": 1.0153744783659127, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.284360647201538, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.709784209728241, + "num_tokens": 239291722.0, + "step": 9246 + }, + { + "epoch": 1.0154842960685262, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.3033101558685303, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7225905656814575, + "num_tokens": 239315226.0, + "step": 9247 + }, + { + "epoch": 1.01559411377114, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2709553241729736, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7286850810050964, + "num_tokens": 239341129.0, + "step": 9248 + }, + { + "epoch": 1.0157039314737535, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.1405351161956787, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7159493565559387, + "num_tokens": 239366717.0, + "step": 9249 + }, + { + "epoch": 1.0158137491763672, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.213360548019409, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7140886783599854, + "num_tokens": 239392989.0, + "step": 9250 + }, + { + "epoch": 1.015923566878981, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.1216776371002197, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.703342080116272, + "num_tokens": 239424703.0, + "step": 9251 + }, + { + "epoch": 1.0160333845815945, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2107467651367188, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.6931183338165283, + "num_tokens": 239454364.0, + "step": 9252 + }, + { + "epoch": 1.0161432022842083, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.257502555847168, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6940528154373169, + "num_tokens": 239482975.0, + "step": 9253 + }, + { + "epoch": 1.0162530199868218, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.117741107940674, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6939384937286377, + "num_tokens": 239515619.0, + "step": 9254 + }, + { + "epoch": 1.0163628376894356, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.4517927169799805, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7074167728424072, + "num_tokens": 239537698.0, + "step": 9255 + }, + { + "epoch": 1.0164726553920491, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.3355555534362793, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.723414421081543, + "num_tokens": 239562114.0, + "step": 9256 + }, + { + "epoch": 1.0165824730946629, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.043638229370117, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.6994590759277344, + "num_tokens": 239590906.0, + "step": 9257 + }, + { + "epoch": 1.0166922907972764, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.188375949859619, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7345467805862427, + "num_tokens": 239615166.0, + "step": 9258 + }, + { + "epoch": 1.0168021084998902, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.114816427230835, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7108478546142578, + "num_tokens": 239641273.0, + "step": 9259 + }, + { + "epoch": 1.016911926202504, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.090186357498169, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7097028493881226, + "num_tokens": 239669932.0, + "step": 9260 + }, + { + "epoch": 1.0170217439051175, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.398876190185547, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7232843637466431, + "num_tokens": 239691431.0, + "step": 9261 + }, + { + "epoch": 1.0171315616077312, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.553053855895996, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7122853994369507, + "num_tokens": 239714438.0, + "step": 9262 + }, + { + "epoch": 1.0172413793103448, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.0357608795166016, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7193734645843506, + "num_tokens": 239741908.0, + "step": 9263 + }, + { + "epoch": 1.0173511970129585, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.111785411834717, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7124500274658203, + "num_tokens": 239772040.0, + "step": 9264 + }, + { + "epoch": 1.017461014715572, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.137826919555664, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7342153787612915, + "num_tokens": 239798973.0, + "step": 9265 + }, + { + "epoch": 1.0175708324181858, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.1738672256469727, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7280547618865967, + "num_tokens": 239826141.0, + "step": 9266 + }, + { + "epoch": 1.0176806501207996, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2175304889678955, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7037910223007202, + "num_tokens": 239852736.0, + "step": 9267 + }, + { + "epoch": 1.017790467823413, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.539133071899414, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7375107407569885, + "num_tokens": 239872701.0, + "step": 9268 + }, + { + "epoch": 1.0179002855260268, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.0374255180358887, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.7022629976272583, + "num_tokens": 239904145.0, + "step": 9269 + }, + { + "epoch": 1.0180101032286404, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.198256015777588, + "learning_rate": 1e-06, + "loss": 0.7817, + "mean_token_accuracy": 0.7476263642311096, + "num_tokens": 239928765.0, + "step": 9270 + }, + { + "epoch": 1.0181199209312541, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2612626552581787, + "learning_rate": 1e-06, + "loss": 1.0043, + "mean_token_accuracy": 0.6960415840148926, + "num_tokens": 239953182.0, + "step": 9271 + }, + { + "epoch": 1.0182297386338677, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.5085370540618896, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7203103303909302, + "num_tokens": 239975609.0, + "step": 9272 + }, + { + "epoch": 1.0183395563364814, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 1.972874641418457, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7231481671333313, + "num_tokens": 240006532.0, + "step": 9273 + }, + { + "epoch": 1.0184493740390952, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 1.8674625158309937, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7295796871185303, + "num_tokens": 240037463.0, + "step": 9274 + }, + { + "epoch": 1.0185591917417087, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.0571670532226562, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6978633999824524, + "num_tokens": 240068775.0, + "step": 9275 + }, + { + "epoch": 1.0186690094443225, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.046358346939087, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7056201100349426, + "num_tokens": 240098457.0, + "step": 9276 + }, + { + "epoch": 1.018778827146936, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.217552661895752, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7048996090888977, + "num_tokens": 240126114.0, + "step": 9277 + }, + { + "epoch": 1.0188886448495498, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.337388515472412, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7267106771469116, + "num_tokens": 240149000.0, + "step": 9278 + }, + { + "epoch": 1.0189984625521633, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 1.8629398345947266, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6907702684402466, + "num_tokens": 240184417.0, + "step": 9279 + }, + { + "epoch": 1.019108280254777, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.107802629470825, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7099955081939697, + "num_tokens": 240213251.0, + "step": 9280 + }, + { + "epoch": 1.0192180979573908, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.2221317291259766, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6912184953689575, + "num_tokens": 240243902.0, + "step": 9281 + }, + { + "epoch": 1.0193279156600044, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.132230520248413, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7178976535797119, + "num_tokens": 240271686.0, + "step": 9282 + }, + { + "epoch": 1.0194377333626181, + "ewc_loss": 8.0108642578125e-05, + "grad_norm": 2.268016815185547, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7011364102363586, + "num_tokens": 240298105.0, + "step": 9283 + }, + { + "epoch": 1.0195475510652316, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1563639640808105, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7218666076660156, + "num_tokens": 240323329.0, + "step": 9284 + }, + { + "epoch": 1.0196573687678454, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.241389513015747, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7145101428031921, + "num_tokens": 240348312.0, + "step": 9285 + }, + { + "epoch": 1.019767186470459, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1347012519836426, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7133055329322815, + "num_tokens": 240373960.0, + "step": 9286 + }, + { + "epoch": 1.0198770041730727, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.4672720432281494, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7248504161834717, + "num_tokens": 240395472.0, + "step": 9287 + }, + { + "epoch": 1.0199868218756865, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 1.9927234649658203, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7006165981292725, + "num_tokens": 240427402.0, + "step": 9288 + }, + { + "epoch": 1.0200966395783, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 1.9089571237564087, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7256616353988647, + "num_tokens": 240461485.0, + "step": 9289 + }, + { + "epoch": 1.0202064572809137, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2376108169555664, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7200218439102173, + "num_tokens": 240488154.0, + "step": 9290 + }, + { + "epoch": 1.0203162749835273, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.32132625579834, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7183084487915039, + "num_tokens": 240515041.0, + "step": 9291 + }, + { + "epoch": 1.020426092686141, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.0212087631225586, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7221231460571289, + "num_tokens": 240543212.0, + "step": 9292 + }, + { + "epoch": 1.0205359103887546, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 1.929597020149231, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7261605262756348, + "num_tokens": 240577429.0, + "step": 9293 + }, + { + "epoch": 1.0206457280913683, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.128776788711548, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7142704129219055, + "num_tokens": 240606022.0, + "step": 9294 + }, + { + "epoch": 1.020755545793982, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.4913547039031982, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7118443250656128, + "num_tokens": 240627360.0, + "step": 9295 + }, + { + "epoch": 1.0208653634965956, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.147752523422241, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7075425386428833, + "num_tokens": 240655953.0, + "step": 9296 + }, + { + "epoch": 1.0209751811992094, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.145749568939209, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7185431718826294, + "num_tokens": 240681901.0, + "step": 9297 + }, + { + "epoch": 1.021084998901823, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1214754581451416, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7124941945075989, + "num_tokens": 240708896.0, + "step": 9298 + }, + { + "epoch": 1.0211948166044367, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.169362783432007, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7300823926925659, + "num_tokens": 240734080.0, + "step": 9299 + }, + { + "epoch": 1.0213046343070502, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.461702823638916, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7076733112335205, + "num_tokens": 240757776.0, + "step": 9300 + }, + { + "epoch": 1.021414452009664, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.4939005374908447, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7233819961547852, + "num_tokens": 240779967.0, + "step": 9301 + }, + { + "epoch": 1.0215242697122777, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1191537380218506, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7080026865005493, + "num_tokens": 240811222.0, + "step": 9302 + }, + { + "epoch": 1.0216340874148913, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1601927280426025, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7163419127464294, + "num_tokens": 240839514.0, + "step": 9303 + }, + { + "epoch": 1.021743905117505, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.075286626815796, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6991255879402161, + "num_tokens": 240869492.0, + "step": 9304 + }, + { + "epoch": 1.0218537228201185, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.000312566757202, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.696699321269989, + "num_tokens": 240903330.0, + "step": 9305 + }, + { + "epoch": 1.0219635405227323, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.0815062522888184, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7495344877243042, + "num_tokens": 240931670.0, + "step": 9306 + }, + { + "epoch": 1.0220733582253458, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2447452545166016, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7321542501449585, + "num_tokens": 240955765.0, + "step": 9307 + }, + { + "epoch": 1.0221831759279596, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.102346420288086, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7189856767654419, + "num_tokens": 240984698.0, + "step": 9308 + }, + { + "epoch": 1.0222929936305734, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2804338932037354, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7144386768341064, + "num_tokens": 241011486.0, + "step": 9309 + }, + { + "epoch": 1.022402811333187, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2324793338775635, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7066835761070251, + "num_tokens": 241038036.0, + "step": 9310 + }, + { + "epoch": 1.0225126290358006, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.3829574584960938, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7194802761077881, + "num_tokens": 241062846.0, + "step": 9311 + }, + { + "epoch": 1.0226224467384142, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.254034996032715, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.728003740310669, + "num_tokens": 241089714.0, + "step": 9312 + }, + { + "epoch": 1.022732264441028, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.0167975425720215, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.7005324959754944, + "num_tokens": 241122272.0, + "step": 9313 + }, + { + "epoch": 1.0228420821436415, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.5303635597229004, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7233123183250427, + "num_tokens": 241143105.0, + "step": 9314 + }, + { + "epoch": 1.0229518998462552, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1585233211517334, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7046691179275513, + "num_tokens": 241170861.0, + "step": 9315 + }, + { + "epoch": 1.023061717548869, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.3980793952941895, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7253071069717407, + "num_tokens": 241193185.0, + "step": 9316 + }, + { + "epoch": 1.0231715352514825, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.0238325595855713, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7219328284263611, + "num_tokens": 241222262.0, + "step": 9317 + }, + { + "epoch": 1.0232813529540963, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.431608200073242, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7114822864532471, + "num_tokens": 241245067.0, + "step": 9318 + }, + { + "epoch": 1.0233911706567098, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2037711143493652, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7146636247634888, + "num_tokens": 241272467.0, + "step": 9319 + }, + { + "epoch": 1.0235009883593236, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.4413974285125732, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7136900424957275, + "num_tokens": 241297534.0, + "step": 9320 + }, + { + "epoch": 1.023610806061937, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.0719568729400635, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7117048501968384, + "num_tokens": 241327611.0, + "step": 9321 + }, + { + "epoch": 1.0237206237645509, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1437838077545166, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7120687961578369, + "num_tokens": 241356755.0, + "step": 9322 + }, + { + "epoch": 1.0238304414671644, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.445388078689575, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7182592749595642, + "num_tokens": 241378369.0, + "step": 9323 + }, + { + "epoch": 1.0239402591697782, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.2596089839935303, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7163118124008179, + "num_tokens": 241404176.0, + "step": 9324 + }, + { + "epoch": 1.024050076872392, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.322723150253296, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7244588136672974, + "num_tokens": 241429475.0, + "step": 9325 + }, + { + "epoch": 1.0241598945750054, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.5129518508911133, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7174276113510132, + "num_tokens": 241450642.0, + "step": 9326 + }, + { + "epoch": 1.0242697122776192, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 1.9994524717330933, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7343772649765015, + "num_tokens": 241480796.0, + "step": 9327 + }, + { + "epoch": 1.0243795299802327, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 3.0645182132720947, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7325516939163208, + "num_tokens": 241498114.0, + "step": 9328 + }, + { + "epoch": 1.0244893476828465, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.589553117752075, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.730164647102356, + "num_tokens": 241519122.0, + "step": 9329 + }, + { + "epoch": 1.02459916538546, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.00528621673584, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7067571878433228, + "num_tokens": 241550097.0, + "step": 9330 + }, + { + "epoch": 1.0247089830880738, + "ewc_loss": 8.058547973632812e-05, + "grad_norm": 2.1347620487213135, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6939699053764343, + "num_tokens": 241578237.0, + "step": 9331 + }, + { + "epoch": 1.0248188007906875, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.372568368911743, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7372003793716431, + "num_tokens": 241599538.0, + "step": 9332 + }, + { + "epoch": 1.024928618493301, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.352731227874756, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7237153053283691, + "num_tokens": 241623433.0, + "step": 9333 + }, + { + "epoch": 1.0250384361959148, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2640602588653564, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.722915768623352, + "num_tokens": 241647005.0, + "step": 9334 + }, + { + "epoch": 1.0251482538985284, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.0939900875091553, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7052209377288818, + "num_tokens": 241676177.0, + "step": 9335 + }, + { + "epoch": 1.0252580716011421, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.114304542541504, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.725752592086792, + "num_tokens": 241705183.0, + "step": 9336 + }, + { + "epoch": 1.0253678893037557, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.537100076675415, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7079535722732544, + "num_tokens": 241728831.0, + "step": 9337 + }, + { + "epoch": 1.0254777070063694, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 1.9750912189483643, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7308462858200073, + "num_tokens": 241757350.0, + "step": 9338 + }, + { + "epoch": 1.0255875247089832, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 1.968132734298706, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7089864015579224, + "num_tokens": 241792208.0, + "step": 9339 + }, + { + "epoch": 1.0256973424115967, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1641976833343506, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.722077488899231, + "num_tokens": 241820190.0, + "step": 9340 + }, + { + "epoch": 1.0258071601142105, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3196895122528076, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7216124534606934, + "num_tokens": 241844640.0, + "step": 9341 + }, + { + "epoch": 1.025916977816824, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.4760165214538574, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.732775866985321, + "num_tokens": 241866014.0, + "step": 9342 + }, + { + "epoch": 1.0260267955194378, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.081535577774048, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6945293545722961, + "num_tokens": 241896370.0, + "step": 9343 + }, + { + "epoch": 1.0261366132220513, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2203028202056885, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7062211036682129, + "num_tokens": 241921445.0, + "step": 9344 + }, + { + "epoch": 1.026246430924665, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.421849489212036, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7281749844551086, + "num_tokens": 241942105.0, + "step": 9345 + }, + { + "epoch": 1.0263562486272788, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.076338052749634, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7003848552703857, + "num_tokens": 241974155.0, + "step": 9346 + }, + { + "epoch": 1.0264660663298923, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1832010746002197, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7142773866653442, + "num_tokens": 242000788.0, + "step": 9347 + }, + { + "epoch": 1.026575884032506, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.0952577590942383, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7265843152999878, + "num_tokens": 242028371.0, + "step": 9348 + }, + { + "epoch": 1.0266857017351196, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.125333786010742, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7266539335250854, + "num_tokens": 242056593.0, + "step": 9349 + }, + { + "epoch": 1.0267955194377334, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1380057334899902, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7383073568344116, + "num_tokens": 242083641.0, + "step": 9350 + }, + { + "epoch": 1.026905337140347, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1282036304473877, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7318058013916016, + "num_tokens": 242109187.0, + "step": 9351 + }, + { + "epoch": 1.0270151548429607, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3589813709259033, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7169783115386963, + "num_tokens": 242131308.0, + "step": 9352 + }, + { + "epoch": 1.0271249725455744, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2891294956207275, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7169191837310791, + "num_tokens": 242158139.0, + "step": 9353 + }, + { + "epoch": 1.027234790248188, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1789960861206055, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.706613302230835, + "num_tokens": 242185517.0, + "step": 9354 + }, + { + "epoch": 1.0273446079508017, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.20192551612854, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7156752347946167, + "num_tokens": 242212944.0, + "step": 9355 + }, + { + "epoch": 1.0274544256534153, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.291642427444458, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7172408103942871, + "num_tokens": 242237921.0, + "step": 9356 + }, + { + "epoch": 1.027564243356029, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.20992112159729, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.70741206407547, + "num_tokens": 242264084.0, + "step": 9357 + }, + { + "epoch": 1.0276740610586426, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.225216865539551, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6966186761856079, + "num_tokens": 242289515.0, + "step": 9358 + }, + { + "epoch": 1.0277838787612563, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.374661684036255, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7314693927764893, + "num_tokens": 242312676.0, + "step": 9359 + }, + { + "epoch": 1.02789369646387, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.139380931854248, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7086862325668335, + "num_tokens": 242341108.0, + "step": 9360 + }, + { + "epoch": 1.0280035141664836, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3382771015167236, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7169727087020874, + "num_tokens": 242364261.0, + "step": 9361 + }, + { + "epoch": 1.0281133318690974, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.555001974105835, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7205087542533875, + "num_tokens": 242385200.0, + "step": 9362 + }, + { + "epoch": 1.028223149571711, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 1.9011266231536865, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.701144814491272, + "num_tokens": 242421340.0, + "step": 9363 + }, + { + "epoch": 1.0283329672743247, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1237902641296387, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7042967081069946, + "num_tokens": 242449861.0, + "step": 9364 + }, + { + "epoch": 1.0284427849769382, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1788814067840576, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7366870641708374, + "num_tokens": 242476627.0, + "step": 9365 + }, + { + "epoch": 1.028552602679552, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3307740688323975, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7078408002853394, + "num_tokens": 242502244.0, + "step": 9366 + }, + { + "epoch": 1.0286624203821657, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2934868335723877, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7160339951515198, + "num_tokens": 242528568.0, + "step": 9367 + }, + { + "epoch": 1.0287722380847792, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.155386447906494, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7321920394897461, + "num_tokens": 242554618.0, + "step": 9368 + }, + { + "epoch": 1.028882055787393, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2642223834991455, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7151750326156616, + "num_tokens": 242576525.0, + "step": 9369 + }, + { + "epoch": 1.0289918734900065, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.079538106918335, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.6980294585227966, + "num_tokens": 242604868.0, + "step": 9370 + }, + { + "epoch": 1.0291016911926203, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.303542375564575, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.727962076663971, + "num_tokens": 242627934.0, + "step": 9371 + }, + { + "epoch": 1.0292115088952338, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.5072295665740967, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7288438081741333, + "num_tokens": 242648332.0, + "step": 9372 + }, + { + "epoch": 1.0293213265978476, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1601040363311768, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7244561314582825, + "num_tokens": 242675591.0, + "step": 9373 + }, + { + "epoch": 1.0294311443004613, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3240807056427, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7107940912246704, + "num_tokens": 242699493.0, + "step": 9374 + }, + { + "epoch": 1.0295409620030749, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.105815887451172, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6836570501327515, + "num_tokens": 242726766.0, + "step": 9375 + }, + { + "epoch": 1.0296507797056886, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 1.8945965766906738, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6935313940048218, + "num_tokens": 242761048.0, + "step": 9376 + }, + { + "epoch": 1.0297605974083022, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.4994688034057617, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7006028294563293, + "num_tokens": 242785961.0, + "step": 9377 + }, + { + "epoch": 1.029870415110916, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1992557048797607, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7105047702789307, + "num_tokens": 242813003.0, + "step": 9378 + }, + { + "epoch": 1.0299802328135295, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.0930395126342773, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6965266466140747, + "num_tokens": 242844550.0, + "step": 9379 + }, + { + "epoch": 1.0300900505161432, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.525057315826416, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7115439176559448, + "num_tokens": 242868020.0, + "step": 9380 + }, + { + "epoch": 1.030199868218757, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.4424657821655273, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7073662281036377, + "num_tokens": 242890243.0, + "step": 9381 + }, + { + "epoch": 1.0303096859213705, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1945934295654297, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7200475335121155, + "num_tokens": 242915586.0, + "step": 9382 + }, + { + "epoch": 1.0304195036239843, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1135051250457764, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7078625559806824, + "num_tokens": 242943042.0, + "step": 9383 + }, + { + "epoch": 1.0305293213265978, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.208937644958496, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7218111753463745, + "num_tokens": 242968983.0, + "step": 9384 + }, + { + "epoch": 1.0306391390292116, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.187757968902588, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7183750867843628, + "num_tokens": 242999537.0, + "step": 9385 + }, + { + "epoch": 1.030748956731825, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.0899195671081543, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7074629664421082, + "num_tokens": 243029698.0, + "step": 9386 + }, + { + "epoch": 1.0308587744344389, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.221890687942505, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7223498225212097, + "num_tokens": 243057058.0, + "step": 9387 + }, + { + "epoch": 1.0309685921370524, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.107523202896118, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7217892408370972, + "num_tokens": 243083376.0, + "step": 9388 + }, + { + "epoch": 1.0310784098396661, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.4121737480163574, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7164762020111084, + "num_tokens": 243106844.0, + "step": 9389 + }, + { + "epoch": 1.03118822754228, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 1.9497545957565308, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.710943341255188, + "num_tokens": 243139308.0, + "step": 9390 + }, + { + "epoch": 1.0312980452448934, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.164090633392334, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7082101106643677, + "num_tokens": 243165509.0, + "step": 9391 + }, + { + "epoch": 1.0314078629475072, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.3124725818634033, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7294226884841919, + "num_tokens": 243187638.0, + "step": 9392 + }, + { + "epoch": 1.0315176806501207, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.5497395992279053, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7327289581298828, + "num_tokens": 243207999.0, + "step": 9393 + }, + { + "epoch": 1.0316274983527345, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1871330738067627, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7065148949623108, + "num_tokens": 243236555.0, + "step": 9394 + }, + { + "epoch": 1.031737316055348, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.160783052444458, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7361961603164673, + "num_tokens": 243262821.0, + "step": 9395 + }, + { + "epoch": 1.0318471337579618, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.127328395843506, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7103004455566406, + "num_tokens": 243292156.0, + "step": 9396 + }, + { + "epoch": 1.0319569514605755, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.2932896614074707, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7141798138618469, + "num_tokens": 243317821.0, + "step": 9397 + }, + { + "epoch": 1.032066769163189, + "ewc_loss": 8.106231689453125e-05, + "grad_norm": 2.1945760250091553, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7049850225448608, + "num_tokens": 243345859.0, + "step": 9398 + }, + { + "epoch": 1.0321765868658028, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.2711143493652344, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7052100896835327, + "num_tokens": 243372271.0, + "step": 9399 + }, + { + "epoch": 1.0322864045684164, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.2158114910125732, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.6966738700866699, + "num_tokens": 243399628.0, + "step": 9400 + }, + { + "epoch": 1.0323962222710301, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.6813230514526367, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7382330298423767, + "num_tokens": 243419002.0, + "step": 9401 + }, + { + "epoch": 1.0325060399736437, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.084615707397461, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7083923816680908, + "num_tokens": 243449824.0, + "step": 9402 + }, + { + "epoch": 1.0326158576762574, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.3137195110321045, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7147555947303772, + "num_tokens": 243474777.0, + "step": 9403 + }, + { + "epoch": 1.0327256753788712, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 1.977927803993225, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7083508968353271, + "num_tokens": 243508762.0, + "step": 9404 + }, + { + "epoch": 1.0328354930814847, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.61731219291687, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7353426218032837, + "num_tokens": 243528461.0, + "step": 9405 + }, + { + "epoch": 1.0329453107840985, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.2635085582733154, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7138229608535767, + "num_tokens": 243556317.0, + "step": 9406 + }, + { + "epoch": 1.033055128486712, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.1369826793670654, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7187028527259827, + "num_tokens": 243585250.0, + "step": 9407 + }, + { + "epoch": 1.0331649461893258, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.473768711090088, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7371187210083008, + "num_tokens": 243606557.0, + "step": 9408 + }, + { + "epoch": 1.0332747638919393, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.1503639221191406, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7106308341026306, + "num_tokens": 243632472.0, + "step": 9409 + }, + { + "epoch": 1.033384581594553, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.1224474906921387, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.728157639503479, + "num_tokens": 243660567.0, + "step": 9410 + }, + { + "epoch": 1.0334943992971668, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.528193473815918, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7230445742607117, + "num_tokens": 243681101.0, + "step": 9411 + }, + { + "epoch": 1.0336042169997803, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.112826347351074, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7106435298919678, + "num_tokens": 243710120.0, + "step": 9412 + }, + { + "epoch": 1.033714034702394, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.5068764686584473, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7348464727401733, + "num_tokens": 243731564.0, + "step": 9413 + }, + { + "epoch": 1.0338238524050076, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.527787208557129, + "learning_rate": 1e-06, + "loss": 0.8003, + "mean_token_accuracy": 0.7509854435920715, + "num_tokens": 243752083.0, + "step": 9414 + }, + { + "epoch": 1.0339336701076214, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.329559564590454, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.7530651092529297, + "num_tokens": 243776098.0, + "step": 9415 + }, + { + "epoch": 1.034043487810235, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.4178059101104736, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.73664391040802, + "num_tokens": 243798831.0, + "step": 9416 + }, + { + "epoch": 1.0341533055128487, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.0881600379943848, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7091801166534424, + "num_tokens": 243826736.0, + "step": 9417 + }, + { + "epoch": 1.0342631232154624, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.3211450576782227, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7234216928482056, + "num_tokens": 243850406.0, + "step": 9418 + }, + { + "epoch": 1.034372940918076, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.688089370727539, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7475512623786926, + "num_tokens": 243866665.0, + "step": 9419 + }, + { + "epoch": 1.0344827586206897, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.2942347526550293, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7239947319030762, + "num_tokens": 243891260.0, + "step": 9420 + }, + { + "epoch": 1.0345925763233033, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.4183638095855713, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7251083850860596, + "num_tokens": 243912863.0, + "step": 9421 + }, + { + "epoch": 1.034702394025917, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 1.987654447555542, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7217697501182556, + "num_tokens": 243941500.0, + "step": 9422 + }, + { + "epoch": 1.0348122117285306, + "ewc_loss": 8.153915405273438e-05, + "grad_norm": 2.356499671936035, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7028533816337585, + "num_tokens": 243965604.0, + "step": 9423 + }, + { + "epoch": 1.0349220294311443, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2704684734344482, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7257219552993774, + "num_tokens": 243989569.0, + "step": 9424 + }, + { + "epoch": 1.035031847133758, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.0782487392425537, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7287464141845703, + "num_tokens": 244018389.0, + "step": 9425 + }, + { + "epoch": 1.0351416648363716, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2939229011535645, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.718104898929596, + "num_tokens": 244042053.0, + "step": 9426 + }, + { + "epoch": 1.0352514825389854, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.0413358211517334, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7164764404296875, + "num_tokens": 244072302.0, + "step": 9427 + }, + { + "epoch": 1.035361300241599, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.446629524230957, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7066550254821777, + "num_tokens": 244095758.0, + "step": 9428 + }, + { + "epoch": 1.0354711179442126, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.085667371749878, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7203848361968994, + "num_tokens": 244124923.0, + "step": 9429 + }, + { + "epoch": 1.0355809356468262, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.372252941131592, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7142257690429688, + "num_tokens": 244148357.0, + "step": 9430 + }, + { + "epoch": 1.03569075334944, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.1955318450927734, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7181445360183716, + "num_tokens": 244175301.0, + "step": 9431 + }, + { + "epoch": 1.0358005710520537, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2205491065979004, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7212462425231934, + "num_tokens": 244201386.0, + "step": 9432 + }, + { + "epoch": 1.0359103887546672, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.214813232421875, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7103112936019897, + "num_tokens": 244228633.0, + "step": 9433 + }, + { + "epoch": 1.036020206457281, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.212078809738159, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7134226560592651, + "num_tokens": 244256930.0, + "step": 9434 + }, + { + "epoch": 1.0361300241598945, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.376814603805542, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7489109635353088, + "num_tokens": 244278992.0, + "step": 9435 + }, + { + "epoch": 1.0362398418625083, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.1039187908172607, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7178546190261841, + "num_tokens": 244307073.0, + "step": 9436 + }, + { + "epoch": 1.0363496595651218, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2600302696228027, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7152329087257385, + "num_tokens": 244333029.0, + "step": 9437 + }, + { + "epoch": 1.0364594772677356, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.4147326946258545, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7395310401916504, + "num_tokens": 244354581.0, + "step": 9438 + }, + { + "epoch": 1.036569294970349, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.641171932220459, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7145484685897827, + "num_tokens": 244376640.0, + "step": 9439 + }, + { + "epoch": 1.0366791126729629, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.196324586868286, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6978058815002441, + "num_tokens": 244403385.0, + "step": 9440 + }, + { + "epoch": 1.0367889303755766, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2462971210479736, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7069896459579468, + "num_tokens": 244431391.0, + "step": 9441 + }, + { + "epoch": 1.0368987480781902, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.1062660217285156, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7079341411590576, + "num_tokens": 244458610.0, + "step": 9442 + }, + { + "epoch": 1.037008565780804, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.0298163890838623, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7065476775169373, + "num_tokens": 244490252.0, + "step": 9443 + }, + { + "epoch": 1.0371183834834174, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.192690372467041, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7274730205535889, + "num_tokens": 244516220.0, + "step": 9444 + }, + { + "epoch": 1.0372282011860312, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2068707942962646, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7173148393630981, + "num_tokens": 244544253.0, + "step": 9445 + }, + { + "epoch": 1.037338018888645, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.3652584552764893, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7204141616821289, + "num_tokens": 244567126.0, + "step": 9446 + }, + { + "epoch": 1.0374478365912585, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.1178526878356934, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7219313383102417, + "num_tokens": 244595282.0, + "step": 9447 + }, + { + "epoch": 1.0375576542938723, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.572963237762451, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7031756639480591, + "num_tokens": 244617865.0, + "step": 9448 + }, + { + "epoch": 1.0376674719964858, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.1018598079681396, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7160049080848694, + "num_tokens": 244645296.0, + "step": 9449 + }, + { + "epoch": 1.0377772896990995, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.683363437652588, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7389994859695435, + "num_tokens": 244663788.0, + "step": 9450 + }, + { + "epoch": 1.037887107401713, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.484503984451294, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7181568145751953, + "num_tokens": 244685999.0, + "step": 9451 + }, + { + "epoch": 1.0379969251043268, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.2179911136627197, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7032744288444519, + "num_tokens": 244713331.0, + "step": 9452 + }, + { + "epoch": 1.0381067428069404, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.1153652667999268, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7275701761245728, + "num_tokens": 244740153.0, + "step": 9453 + }, + { + "epoch": 1.0382165605095541, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.3772833347320557, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7219527959823608, + "num_tokens": 244763602.0, + "step": 9454 + }, + { + "epoch": 1.0383263782121679, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 2.0218029022216797, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7269629240036011, + "num_tokens": 244793724.0, + "step": 9455 + }, + { + "epoch": 1.0384361959147814, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.2041027545928955, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.737456202507019, + "num_tokens": 244818177.0, + "step": 9456 + }, + { + "epoch": 1.0385460136173952, + "ewc_loss": 8.20159912109375e-05, + "grad_norm": 1.9662387371063232, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7281082272529602, + "num_tokens": 244848557.0, + "step": 9457 + }, + { + "epoch": 1.0386558313200087, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.2314040660858154, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.727524995803833, + "num_tokens": 244874269.0, + "step": 9458 + }, + { + "epoch": 1.0387656490226225, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.4725635051727295, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7273972034454346, + "num_tokens": 244894552.0, + "step": 9459 + }, + { + "epoch": 1.038875466725236, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.1939449310302734, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7333115339279175, + "num_tokens": 244919365.0, + "step": 9460 + }, + { + "epoch": 1.0389852844278498, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.2626118659973145, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7126853466033936, + "num_tokens": 244943450.0, + "step": 9461 + }, + { + "epoch": 1.0390951021304635, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.364651679992676, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7159938216209412, + "num_tokens": 244966856.0, + "step": 9462 + }, + { + "epoch": 1.039204919833077, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.0810868740081787, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6973952054977417, + "num_tokens": 244998059.0, + "step": 9463 + }, + { + "epoch": 1.0393147375356908, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.2815616130828857, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7087259292602539, + "num_tokens": 245022171.0, + "step": 9464 + }, + { + "epoch": 1.0394245552383043, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.2130889892578125, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7273954153060913, + "num_tokens": 245049263.0, + "step": 9465 + }, + { + "epoch": 1.039534372940918, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.3022079467773438, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7385351657867432, + "num_tokens": 245073793.0, + "step": 9466 + }, + { + "epoch": 1.0396441906435316, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 1.9895433187484741, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7184140682220459, + "num_tokens": 245103029.0, + "step": 9467 + }, + { + "epoch": 1.0397540083461454, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.275519371032715, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7164628505706787, + "num_tokens": 245128495.0, + "step": 9468 + }, + { + "epoch": 1.0398638260487592, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.0956239700317383, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7244519591331482, + "num_tokens": 245156033.0, + "step": 9469 + }, + { + "epoch": 1.0399736437513727, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.653022050857544, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7382925748825073, + "num_tokens": 245174744.0, + "step": 9470 + }, + { + "epoch": 1.0400834614539864, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.375805616378784, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.71192467212677, + "num_tokens": 245200011.0, + "step": 9471 + }, + { + "epoch": 1.0401932791566, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.1529734134674072, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7273194193840027, + "num_tokens": 245227165.0, + "step": 9472 + }, + { + "epoch": 1.0403030968592137, + "ewc_loss": 8.249282836914062e-05, + "grad_norm": 2.029989004135132, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7110907435417175, + "num_tokens": 245257985.0, + "step": 9473 + }, + { + "epoch": 1.0404129145618273, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3857483863830566, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7497159242630005, + "num_tokens": 245279876.0, + "step": 9474 + }, + { + "epoch": 1.040522732264441, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3124465942382812, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7031753063201904, + "num_tokens": 245305272.0, + "step": 9475 + }, + { + "epoch": 1.0406325499670548, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3655920028686523, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7141125202178955, + "num_tokens": 245328051.0, + "step": 9476 + }, + { + "epoch": 1.0407423676696683, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 1.9764435291290283, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7148960828781128, + "num_tokens": 245358157.0, + "step": 9477 + }, + { + "epoch": 1.040852185372282, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.1078569889068604, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7070850729942322, + "num_tokens": 245384741.0, + "step": 9478 + }, + { + "epoch": 1.0409620030748956, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.2465097904205322, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7082265019416809, + "num_tokens": 245409463.0, + "step": 9479 + }, + { + "epoch": 1.0410718207775094, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.1137328147888184, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7340176105499268, + "num_tokens": 245438723.0, + "step": 9480 + }, + { + "epoch": 1.041181638480123, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.502016067504883, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7320587635040283, + "num_tokens": 245458902.0, + "step": 9481 + }, + { + "epoch": 1.0412914561827367, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.1252315044403076, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7050976753234863, + "num_tokens": 245485163.0, + "step": 9482 + }, + { + "epoch": 1.0414012738853504, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.0882678031921387, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7303207516670227, + "num_tokens": 245511526.0, + "step": 9483 + }, + { + "epoch": 1.041511091587964, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.4255640506744385, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7247496843338013, + "num_tokens": 245533940.0, + "step": 9484 + }, + { + "epoch": 1.0416209092905777, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.463223695755005, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7497937679290771, + "num_tokens": 245553366.0, + "step": 9485 + }, + { + "epoch": 1.0417307269931912, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.5471484661102295, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7356104850769043, + "num_tokens": 245573331.0, + "step": 9486 + }, + { + "epoch": 1.041840544695805, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.4959464073181152, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7419871687889099, + "num_tokens": 245595441.0, + "step": 9487 + }, + { + "epoch": 1.0419503623984185, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.5021049976348877, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7279057502746582, + "num_tokens": 245616215.0, + "step": 9488 + }, + { + "epoch": 1.0420601801010323, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.219943046569824, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7144886255264282, + "num_tokens": 245641474.0, + "step": 9489 + }, + { + "epoch": 1.042169997803646, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.2966158390045166, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7133856415748596, + "num_tokens": 245666626.0, + "step": 9490 + }, + { + "epoch": 1.0422798155062596, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.102458953857422, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7234644889831543, + "num_tokens": 245693609.0, + "step": 9491 + }, + { + "epoch": 1.0423896332088733, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.436694622039795, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.719616174697876, + "num_tokens": 245715909.0, + "step": 9492 + }, + { + "epoch": 1.0424994509114869, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3035972118377686, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7259789705276489, + "num_tokens": 245743609.0, + "step": 9493 + }, + { + "epoch": 1.0426092686141006, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.543302059173584, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7210976481437683, + "num_tokens": 245765807.0, + "step": 9494 + }, + { + "epoch": 1.0427190863167142, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.057100296020508, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7017240524291992, + "num_tokens": 245798404.0, + "step": 9495 + }, + { + "epoch": 1.042828904019328, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.4039347171783447, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7150225639343262, + "num_tokens": 245820725.0, + "step": 9496 + }, + { + "epoch": 1.0429387217219417, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.293381929397583, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7102564573287964, + "num_tokens": 245844551.0, + "step": 9497 + }, + { + "epoch": 1.0430485394245552, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.146026372909546, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7327005863189697, + "num_tokens": 245870890.0, + "step": 9498 + }, + { + "epoch": 1.043158357127169, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.2062482833862305, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7112857103347778, + "num_tokens": 245896044.0, + "step": 9499 + }, + { + "epoch": 1.0432681748297825, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.264967679977417, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7006178498268127, + "num_tokens": 245923223.0, + "step": 9500 + }, + { + "epoch": 1.0433779925323963, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.207383394241333, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7069932222366333, + "num_tokens": 245950175.0, + "step": 9501 + }, + { + "epoch": 1.0434878102350098, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.285599946975708, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7274776697158813, + "num_tokens": 245974500.0, + "step": 9502 + }, + { + "epoch": 1.0435976279376236, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 1.9731230735778809, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.7030763030052185, + "num_tokens": 246005891.0, + "step": 9503 + }, + { + "epoch": 1.043707445640237, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3044211864471436, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6885298490524292, + "num_tokens": 246030961.0, + "step": 9504 + }, + { + "epoch": 1.0438172633428509, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3056581020355225, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7094839811325073, + "num_tokens": 246054915.0, + "step": 9505 + }, + { + "epoch": 1.0439270810454646, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.222196102142334, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7265139818191528, + "num_tokens": 246081890.0, + "step": 9506 + }, + { + "epoch": 1.0440368987480781, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.1112279891967773, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.722083330154419, + "num_tokens": 246110456.0, + "step": 9507 + }, + { + "epoch": 1.044146716450692, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.0672216415405273, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.6979125738143921, + "num_tokens": 246139574.0, + "step": 9508 + }, + { + "epoch": 1.0442565341533054, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.2786009311676025, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7044981718063354, + "num_tokens": 246165445.0, + "step": 9509 + }, + { + "epoch": 1.0443663518559192, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.47786545753479, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7344058752059937, + "num_tokens": 246185870.0, + "step": 9510 + }, + { + "epoch": 1.0444761695585327, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.023375988006592, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6994398832321167, + "num_tokens": 246216822.0, + "step": 9511 + }, + { + "epoch": 1.0445859872611465, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.4629831314086914, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7318878173828125, + "num_tokens": 246239351.0, + "step": 9512 + }, + { + "epoch": 1.0446958049637602, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.4018595218658447, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7345547676086426, + "num_tokens": 246262271.0, + "step": 9513 + }, + { + "epoch": 1.0448056226663738, + "ewc_loss": 8.344650268554688e-05, + "grad_norm": 2.3909764289855957, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7151650786399841, + "num_tokens": 246286562.0, + "step": 9514 + }, + { + "epoch": 1.0449154403689875, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.444427013397217, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7140663862228394, + "num_tokens": 246310265.0, + "step": 9515 + }, + { + "epoch": 1.045025258071601, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.358623743057251, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7020201683044434, + "num_tokens": 246335195.0, + "step": 9516 + }, + { + "epoch": 1.0451350757742148, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.34028959274292, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7274808287620544, + "num_tokens": 246360099.0, + "step": 9517 + }, + { + "epoch": 1.0452448934768284, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.3113231658935547, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6856691837310791, + "num_tokens": 246385857.0, + "step": 9518 + }, + { + "epoch": 1.0453547111794421, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.200906991958618, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7378271818161011, + "num_tokens": 246411918.0, + "step": 9519 + }, + { + "epoch": 1.0454645288820559, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.369994878768921, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7435252666473389, + "num_tokens": 246433859.0, + "step": 9520 + }, + { + "epoch": 1.0455743465846694, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.472032308578491, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7331042289733887, + "num_tokens": 246455155.0, + "step": 9521 + }, + { + "epoch": 1.0456841642872832, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.0967650413513184, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7168717384338379, + "num_tokens": 246482223.0, + "step": 9522 + }, + { + "epoch": 1.0457939819898967, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.2703750133514404, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7050197124481201, + "num_tokens": 246507535.0, + "step": 9523 + }, + { + "epoch": 1.0459037996925105, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.35056209564209, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7171056270599365, + "num_tokens": 246530389.0, + "step": 9524 + }, + { + "epoch": 1.046013617395124, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.1322784423828125, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7112271785736084, + "num_tokens": 246558410.0, + "step": 9525 + }, + { + "epoch": 1.0461234350977378, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.044088840484619, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7321556806564331, + "num_tokens": 246586971.0, + "step": 9526 + }, + { + "epoch": 1.0462332528003515, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.2245049476623535, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7205427885055542, + "num_tokens": 246611921.0, + "step": 9527 + }, + { + "epoch": 1.046343070502965, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.301319122314453, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7220642566680908, + "num_tokens": 246637183.0, + "step": 9528 + }, + { + "epoch": 1.0464528882055788, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.2227697372436523, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.711580753326416, + "num_tokens": 246661123.0, + "step": 9529 + }, + { + "epoch": 1.0465627059081923, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.262824773788452, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7298744916915894, + "num_tokens": 246684376.0, + "step": 9530 + }, + { + "epoch": 1.046672523610806, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.5831456184387207, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7259731888771057, + "num_tokens": 246703163.0, + "step": 9531 + }, + { + "epoch": 1.0467823413134196, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.217560052871704, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7312373518943787, + "num_tokens": 246727436.0, + "step": 9532 + }, + { + "epoch": 1.0468921590160334, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.116682291030884, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7237571477890015, + "num_tokens": 246754963.0, + "step": 9533 + }, + { + "epoch": 1.0470019767186471, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.406358003616333, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.732795238494873, + "num_tokens": 246777203.0, + "step": 9534 + }, + { + "epoch": 1.0471117944212607, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.411957263946533, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7469923496246338, + "num_tokens": 246800178.0, + "step": 9535 + }, + { + "epoch": 1.0472216121238744, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.180968761444092, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7076992392539978, + "num_tokens": 246827037.0, + "step": 9536 + }, + { + "epoch": 1.047331429826488, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.4679043292999268, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7486375570297241, + "num_tokens": 246848866.0, + "step": 9537 + }, + { + "epoch": 1.0474412475291017, + "ewc_loss": 8.392333984375e-05, + "grad_norm": 2.4767544269561768, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.723758339881897, + "num_tokens": 246869487.0, + "step": 9538 + }, + { + "epoch": 1.0475510652317153, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.320134162902832, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7007308006286621, + "num_tokens": 246897217.0, + "step": 9539 + }, + { + "epoch": 1.047660882934329, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.07956862449646, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7031005024909973, + "num_tokens": 246927339.0, + "step": 9540 + }, + { + "epoch": 1.0477707006369428, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.002612590789795, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7199259996414185, + "num_tokens": 246957855.0, + "step": 9541 + }, + { + "epoch": 1.0478805183395563, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.376107692718506, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7312638759613037, + "num_tokens": 246979698.0, + "step": 9542 + }, + { + "epoch": 1.04799033604217, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.1242899894714355, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.741422712802887, + "num_tokens": 247007272.0, + "step": 9543 + }, + { + "epoch": 1.0481001537447836, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.4093005657196045, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7372255325317383, + "num_tokens": 247030032.0, + "step": 9544 + }, + { + "epoch": 1.0482099714473974, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.303682804107666, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7258815765380859, + "num_tokens": 247053901.0, + "step": 9545 + }, + { + "epoch": 1.048319789150011, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.4792048931121826, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7142602801322937, + "num_tokens": 247076976.0, + "step": 9546 + }, + { + "epoch": 1.0484296068526247, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.5528385639190674, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7374856472015381, + "num_tokens": 247097690.0, + "step": 9547 + }, + { + "epoch": 1.0485394245552384, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.3048994541168213, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7380033135414124, + "num_tokens": 247121076.0, + "step": 9548 + }, + { + "epoch": 1.048649242257852, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.2354769706726074, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7317277193069458, + "num_tokens": 247145963.0, + "step": 9549 + }, + { + "epoch": 1.0487590599604657, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.1897926330566406, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7105852961540222, + "num_tokens": 247172368.0, + "step": 9550 + }, + { + "epoch": 1.0488688776630792, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.0975775718688965, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.6979444026947021, + "num_tokens": 247201180.0, + "step": 9551 + }, + { + "epoch": 1.048978695365693, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.043022871017456, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.704513430595398, + "num_tokens": 247232563.0, + "step": 9552 + }, + { + "epoch": 1.0490885130683065, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.292750120162964, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.697402834892273, + "num_tokens": 247257790.0, + "step": 9553 + }, + { + "epoch": 1.0491983307709203, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.3756210803985596, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7211141586303711, + "num_tokens": 247279700.0, + "step": 9554 + }, + { + "epoch": 1.0493081484735338, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.360008716583252, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7374100685119629, + "num_tokens": 247303222.0, + "step": 9555 + }, + { + "epoch": 1.0494179661761476, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.3299546241760254, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7381747364997864, + "num_tokens": 247325782.0, + "step": 9556 + }, + { + "epoch": 1.0495277838787613, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.092200756072998, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.6977447271347046, + "num_tokens": 247356404.0, + "step": 9557 + }, + { + "epoch": 1.0496376015813749, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 1.9563990831375122, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7014814615249634, + "num_tokens": 247389537.0, + "step": 9558 + }, + { + "epoch": 1.0497474192839886, + "ewc_loss": 8.487701416015625e-05, + "grad_norm": 2.413290023803711, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7120386362075806, + "num_tokens": 247412214.0, + "step": 9559 + }, + { + "epoch": 1.0498572369866022, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.434967041015625, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.7486932873725891, + "num_tokens": 247432745.0, + "step": 9560 + }, + { + "epoch": 1.049967054689216, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2441861629486084, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7193681001663208, + "num_tokens": 247457197.0, + "step": 9561 + }, + { + "epoch": 1.0500768723918297, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.570061683654785, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7013167142868042, + "num_tokens": 247479184.0, + "step": 9562 + }, + { + "epoch": 1.0501866900944432, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.4227211475372314, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7370262145996094, + "num_tokens": 247502682.0, + "step": 9563 + }, + { + "epoch": 1.050296507797057, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.249577760696411, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7062718868255615, + "num_tokens": 247531199.0, + "step": 9564 + }, + { + "epoch": 1.0504063254996705, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2924344539642334, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6990718841552734, + "num_tokens": 247560239.0, + "step": 9565 + }, + { + "epoch": 1.0505161432022843, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.3648152351379395, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7036373615264893, + "num_tokens": 247584598.0, + "step": 9566 + }, + { + "epoch": 1.0506259609048978, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.1741881370544434, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7256360054016113, + "num_tokens": 247610185.0, + "step": 9567 + }, + { + "epoch": 1.0507357786075116, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.4131970405578613, + "learning_rate": 1e-06, + "loss": 0.773, + "mean_token_accuracy": 0.7508841753005981, + "num_tokens": 247632298.0, + "step": 9568 + }, + { + "epoch": 1.050845596310125, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.0614733695983887, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7311087846755981, + "num_tokens": 247659916.0, + "step": 9569 + }, + { + "epoch": 1.0509554140127388, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.087899684906006, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7244828343391418, + "num_tokens": 247688480.0, + "step": 9570 + }, + { + "epoch": 1.0510652317153526, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.1740500926971436, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7157109379768372, + "num_tokens": 247716386.0, + "step": 9571 + }, + { + "epoch": 1.0511750494179661, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.274869203567505, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7146782875061035, + "num_tokens": 247741248.0, + "step": 9572 + }, + { + "epoch": 1.05128486712058, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2373297214508057, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7201122641563416, + "num_tokens": 247767135.0, + "step": 9573 + }, + { + "epoch": 1.0513946848231934, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2118868827819824, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7190859317779541, + "num_tokens": 247793866.0, + "step": 9574 + }, + { + "epoch": 1.0515045025258072, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 1.978874921798706, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7105447053909302, + "num_tokens": 247826488.0, + "step": 9575 + }, + { + "epoch": 1.0516143202284207, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.429206371307373, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7042278051376343, + "num_tokens": 247848265.0, + "step": 9576 + }, + { + "epoch": 1.0517241379310345, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.078441858291626, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7126414775848389, + "num_tokens": 247878027.0, + "step": 9577 + }, + { + "epoch": 1.0518339556336482, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.5270793437957764, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7313928604125977, + "num_tokens": 247899160.0, + "step": 9578 + }, + { + "epoch": 1.0519437733362618, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2994544506073, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6854048371315002, + "num_tokens": 247924094.0, + "step": 9579 + }, + { + "epoch": 1.0520535910388755, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 1.8292702436447144, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.702004075050354, + "num_tokens": 247958200.0, + "step": 9580 + }, + { + "epoch": 1.052163408741489, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2744035720825195, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.717070460319519, + "num_tokens": 247982276.0, + "step": 9581 + }, + { + "epoch": 1.0522732264441028, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.148780107498169, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7235430479049683, + "num_tokens": 248010082.0, + "step": 9582 + }, + { + "epoch": 1.0523830441467164, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.3513612747192383, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7346464395523071, + "num_tokens": 248035005.0, + "step": 9583 + }, + { + "epoch": 1.05249286184933, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.1816179752349854, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7098339796066284, + "num_tokens": 248061281.0, + "step": 9584 + }, + { + "epoch": 1.0526026795519439, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.2428178787231445, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7056065201759338, + "num_tokens": 248089291.0, + "step": 9585 + }, + { + "epoch": 1.0527124972545574, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.112119674682617, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7180477380752563, + "num_tokens": 248118085.0, + "step": 9586 + }, + { + "epoch": 1.0528223149571712, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.3805291652679443, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7145114541053772, + "num_tokens": 248141100.0, + "step": 9587 + }, + { + "epoch": 1.0529321326597847, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.304635763168335, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7230513691902161, + "num_tokens": 248165329.0, + "step": 9588 + }, + { + "epoch": 1.0530419503623984, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.0481343269348145, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7003656625747681, + "num_tokens": 248196463.0, + "step": 9589 + }, + { + "epoch": 1.053151768065012, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.4006383419036865, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7391201853752136, + "num_tokens": 248219419.0, + "step": 9590 + }, + { + "epoch": 1.0532615857676257, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.5583584308624268, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7009239792823792, + "num_tokens": 248242837.0, + "step": 9591 + }, + { + "epoch": 1.0533714034702395, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.378872871398926, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7133432626724243, + "num_tokens": 248267252.0, + "step": 9592 + }, + { + "epoch": 1.053481221172853, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 1.877256989479065, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.6995653510093689, + "num_tokens": 248303356.0, + "step": 9593 + }, + { + "epoch": 1.0535910388754668, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.1824707984924316, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7059099078178406, + "num_tokens": 248331023.0, + "step": 9594 + }, + { + "epoch": 1.0537008565780803, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 1.9595776796340942, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7225173711776733, + "num_tokens": 248360940.0, + "step": 9595 + }, + { + "epoch": 1.053810674280694, + "ewc_loss": 8.535385131835938e-05, + "grad_norm": 2.135258913040161, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7193104028701782, + "num_tokens": 248390650.0, + "step": 9596 + }, + { + "epoch": 1.0539204919833076, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1218972206115723, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.719550609588623, + "num_tokens": 248417494.0, + "step": 9597 + }, + { + "epoch": 1.0540303096859214, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.021721124649048, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7358866930007935, + "num_tokens": 248444723.0, + "step": 9598 + }, + { + "epoch": 1.0541401273885351, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4015696048736572, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7128818035125732, + "num_tokens": 248466685.0, + "step": 9599 + }, + { + "epoch": 1.0542499450911487, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5013043880462646, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7271044254302979, + "num_tokens": 248488630.0, + "step": 9600 + }, + { + "epoch": 1.0543597627937624, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.071044921875, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.685542106628418, + "num_tokens": 248518374.0, + "step": 9601 + }, + { + "epoch": 1.054469580496376, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.250894784927368, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7153611183166504, + "num_tokens": 248542665.0, + "step": 9602 + }, + { + "epoch": 1.0545793981989897, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 1.8823760747909546, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7249774932861328, + "num_tokens": 248576368.0, + "step": 9603 + }, + { + "epoch": 1.0546892159016032, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 1.9946448802947998, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6949308514595032, + "num_tokens": 248606520.0, + "step": 9604 + }, + { + "epoch": 1.054799033604217, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.05275559425354, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6881307363510132, + "num_tokens": 248634950.0, + "step": 9605 + }, + { + "epoch": 1.0549088513068308, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.086744546890259, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7203942537307739, + "num_tokens": 248663549.0, + "step": 9606 + }, + { + "epoch": 1.0550186690094443, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3241117000579834, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7076608538627625, + "num_tokens": 248688326.0, + "step": 9607 + }, + { + "epoch": 1.055128486712058, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1972391605377197, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7119559645652771, + "num_tokens": 248714511.0, + "step": 9608 + }, + { + "epoch": 1.0552383044146716, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1721553802490234, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7204892039299011, + "num_tokens": 248742304.0, + "step": 9609 + }, + { + "epoch": 1.0553481221172853, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.137030601501465, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7378228306770325, + "num_tokens": 248768175.0, + "step": 9610 + }, + { + "epoch": 1.0554579398198989, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.07863712310791, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7114858627319336, + "num_tokens": 248798423.0, + "step": 9611 + }, + { + "epoch": 1.0555677575225126, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2646644115448, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7145615220069885, + "num_tokens": 248823545.0, + "step": 9612 + }, + { + "epoch": 1.0556775752251264, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.592341423034668, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7395650744438171, + "num_tokens": 248842430.0, + "step": 9613 + }, + { + "epoch": 1.05578739292774, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.102602005004883, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7103368043899536, + "num_tokens": 248870096.0, + "step": 9614 + }, + { + "epoch": 1.0558972106303537, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.365535259246826, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7111880779266357, + "num_tokens": 248893581.0, + "step": 9615 + }, + { + "epoch": 1.0560070283329672, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1791632175445557, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7102571725845337, + "num_tokens": 248919001.0, + "step": 9616 + }, + { + "epoch": 1.056116846035581, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1288669109344482, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7095273733139038, + "num_tokens": 248946510.0, + "step": 9617 + }, + { + "epoch": 1.0562266637381945, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.0363380908966064, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7102980613708496, + "num_tokens": 248978612.0, + "step": 9618 + }, + { + "epoch": 1.0563364814408083, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.090890645980835, + "learning_rate": 1e-06, + "loss": 1.009, + "mean_token_accuracy": 0.6897826790809631, + "num_tokens": 249008011.0, + "step": 9619 + }, + { + "epoch": 1.0564462991434218, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2525815963745117, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7307297587394714, + "num_tokens": 249031806.0, + "step": 9620 + }, + { + "epoch": 1.0565561168460356, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5503087043762207, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.721106231212616, + "num_tokens": 249052702.0, + "step": 9621 + }, + { + "epoch": 1.0566659345486493, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 1.8975919485092163, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6881839036941528, + "num_tokens": 249087582.0, + "step": 9622 + }, + { + "epoch": 1.0567757522512629, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3747122287750244, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7096788287162781, + "num_tokens": 249111026.0, + "step": 9623 + }, + { + "epoch": 1.0568855699538766, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.343364715576172, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7142412662506104, + "num_tokens": 249133904.0, + "step": 9624 + }, + { + "epoch": 1.0569953876564901, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.7201197147369385, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7212049961090088, + "num_tokens": 249153840.0, + "step": 9625 + }, + { + "epoch": 1.057105205359104, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.339951753616333, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7122991681098938, + "num_tokens": 249177898.0, + "step": 9626 + }, + { + "epoch": 1.0572150230617177, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4865517616271973, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7236194014549255, + "num_tokens": 249199392.0, + "step": 9627 + }, + { + "epoch": 1.0573248407643312, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1578245162963867, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7152926325798035, + "num_tokens": 249227251.0, + "step": 9628 + }, + { + "epoch": 1.057434658466945, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4801292419433594, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7393240928649902, + "num_tokens": 249248156.0, + "step": 9629 + }, + { + "epoch": 1.0575444761695585, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5065929889678955, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6967777013778687, + "num_tokens": 249271245.0, + "step": 9630 + }, + { + "epoch": 1.0576542938721722, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3689329624176025, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.696181058883667, + "num_tokens": 249297210.0, + "step": 9631 + }, + { + "epoch": 1.0577641115747858, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5163583755493164, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7309419512748718, + "num_tokens": 249317807.0, + "step": 9632 + }, + { + "epoch": 1.0578739292773995, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.160883903503418, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7195868492126465, + "num_tokens": 249347204.0, + "step": 9633 + }, + { + "epoch": 1.057983746980013, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.049285888671875, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7149306535720825, + "num_tokens": 249377018.0, + "step": 9634 + }, + { + "epoch": 1.0580935646826268, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1480987071990967, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.726653516292572, + "num_tokens": 249404706.0, + "step": 9635 + }, + { + "epoch": 1.0582033823852406, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.405238389968872, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7241057753562927, + "num_tokens": 249428576.0, + "step": 9636 + }, + { + "epoch": 1.0583132000878541, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3083138465881348, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7331491708755493, + "num_tokens": 249455495.0, + "step": 9637 + }, + { + "epoch": 1.0584230177904679, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.584165096282959, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7294842600822449, + "num_tokens": 249474722.0, + "step": 9638 + }, + { + "epoch": 1.0585328354930814, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2482597827911377, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.720233142375946, + "num_tokens": 249501296.0, + "step": 9639 + }, + { + "epoch": 1.0586426531956952, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5557608604431152, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7339098453521729, + "num_tokens": 249521551.0, + "step": 9640 + }, + { + "epoch": 1.0587524708983087, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.306483745574951, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7259222269058228, + "num_tokens": 249546353.0, + "step": 9641 + }, + { + "epoch": 1.0588622886009225, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3463497161865234, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7084976434707642, + "num_tokens": 249572895.0, + "step": 9642 + }, + { + "epoch": 1.0589721063035362, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.400069236755371, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7397375702857971, + "num_tokens": 249596783.0, + "step": 9643 + }, + { + "epoch": 1.0590819240061498, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2352044582366943, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7170514464378357, + "num_tokens": 249623279.0, + "step": 9644 + }, + { + "epoch": 1.0591917417087635, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.30773663520813, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7148053050041199, + "num_tokens": 249651574.0, + "step": 9645 + }, + { + "epoch": 1.059301559411377, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2629029750823975, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7093484401702881, + "num_tokens": 249677631.0, + "step": 9646 + }, + { + "epoch": 1.0594113771139908, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.288853168487549, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7063229084014893, + "num_tokens": 249703995.0, + "step": 9647 + }, + { + "epoch": 1.0595211948166043, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2944138050079346, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7219477295875549, + "num_tokens": 249728629.0, + "step": 9648 + }, + { + "epoch": 1.059631012519218, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1502525806427, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6998159289360046, + "num_tokens": 249757537.0, + "step": 9649 + }, + { + "epoch": 1.0597408302218319, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.5051121711730957, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7161308526992798, + "num_tokens": 249779905.0, + "step": 9650 + }, + { + "epoch": 1.0598506479244454, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2728381156921387, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7248294353485107, + "num_tokens": 249806155.0, + "step": 9651 + }, + { + "epoch": 1.0599604656270591, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.119217872619629, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.706294059753418, + "num_tokens": 249836082.0, + "step": 9652 + }, + { + "epoch": 1.0600702833296727, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3539743423461914, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7206828594207764, + "num_tokens": 249860640.0, + "step": 9653 + }, + { + "epoch": 1.0601801010322864, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3270819187164307, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7124015092849731, + "num_tokens": 249884934.0, + "step": 9654 + }, + { + "epoch": 1.0602899187349, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2523672580718994, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7147113084793091, + "num_tokens": 249910064.0, + "step": 9655 + }, + { + "epoch": 1.0603997364375137, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1931328773498535, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7041239738464355, + "num_tokens": 249937239.0, + "step": 9656 + }, + { + "epoch": 1.0605095541401275, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.770124673843384, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7407115697860718, + "num_tokens": 249955880.0, + "step": 9657 + }, + { + "epoch": 1.060619371842741, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 1.9496055841445923, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7000023722648621, + "num_tokens": 249988295.0, + "step": 9658 + }, + { + "epoch": 1.0607291895453548, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4019792079925537, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7255002856254578, + "num_tokens": 250014371.0, + "step": 9659 + }, + { + "epoch": 1.0608390072479683, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4658942222595215, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7323403358459473, + "num_tokens": 250035108.0, + "step": 9660 + }, + { + "epoch": 1.060948824950582, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3859732151031494, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7249397039413452, + "num_tokens": 250057924.0, + "step": 9661 + }, + { + "epoch": 1.0610586426531956, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.042896032333374, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7119077444076538, + "num_tokens": 250088872.0, + "step": 9662 + }, + { + "epoch": 1.0611684603558094, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.612849235534668, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6962788701057434, + "num_tokens": 250110129.0, + "step": 9663 + }, + { + "epoch": 1.0612782780584231, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.3800601959228516, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7177740335464478, + "num_tokens": 250132915.0, + "step": 9664 + }, + { + "epoch": 1.0613880957610367, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.2274599075317383, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7110081911087036, + "num_tokens": 250158548.0, + "step": 9665 + }, + { + "epoch": 1.0614979134636504, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.292229175567627, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7104533910751343, + "num_tokens": 250183701.0, + "step": 9666 + }, + { + "epoch": 1.061607731166264, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.121851682662964, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7130609750747681, + "num_tokens": 250212131.0, + "step": 9667 + }, + { + "epoch": 1.0617175488688777, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.4274659156799316, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7375009655952454, + "num_tokens": 250235131.0, + "step": 9668 + }, + { + "epoch": 1.0618273665714912, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.388861894607544, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7182087898254395, + "num_tokens": 250258718.0, + "step": 9669 + }, + { + "epoch": 1.061937184274105, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.1886017322540283, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7436398267745972, + "num_tokens": 250283341.0, + "step": 9670 + }, + { + "epoch": 1.0620470019767188, + "ewc_loss": 8.630752563476562e-05, + "grad_norm": 2.290912389755249, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7401362061500549, + "num_tokens": 250304962.0, + "step": 9671 + }, + { + "epoch": 1.0621568196793323, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2577075958251953, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7008862495422363, + "num_tokens": 250332787.0, + "step": 9672 + }, + { + "epoch": 1.062266637381946, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 1.9051645994186401, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6916177272796631, + "num_tokens": 250368623.0, + "step": 9673 + }, + { + "epoch": 1.0623764550845596, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.366898536682129, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.723552942276001, + "num_tokens": 250392124.0, + "step": 9674 + }, + { + "epoch": 1.0624862727871733, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1953999996185303, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7199965119361877, + "num_tokens": 250417204.0, + "step": 9675 + }, + { + "epoch": 1.0625960904897869, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.251006603240967, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7065249681472778, + "num_tokens": 250444684.0, + "step": 9676 + }, + { + "epoch": 1.0627059081924006, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3187718391418457, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.7024668455123901, + "num_tokens": 250469887.0, + "step": 9677 + }, + { + "epoch": 1.0628157258950144, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.146695852279663, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7231215238571167, + "num_tokens": 250498534.0, + "step": 9678 + }, + { + "epoch": 1.062925543597628, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.183212995529175, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7405338287353516, + "num_tokens": 250525802.0, + "step": 9679 + }, + { + "epoch": 1.0630353613002417, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.128917694091797, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7258840203285217, + "num_tokens": 250553585.0, + "step": 9680 + }, + { + "epoch": 1.0631451790028552, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3986005783081055, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7334839105606079, + "num_tokens": 250576137.0, + "step": 9681 + }, + { + "epoch": 1.063254996705469, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.015756845474243, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.693940281867981, + "num_tokens": 250606474.0, + "step": 9682 + }, + { + "epoch": 1.0633648144080825, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 1.9910097122192383, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7206920385360718, + "num_tokens": 250636724.0, + "step": 9683 + }, + { + "epoch": 1.0634746321106963, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1076037883758545, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.711959958076477, + "num_tokens": 250665159.0, + "step": 9684 + }, + { + "epoch": 1.0635844498133098, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3377890586853027, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7374248504638672, + "num_tokens": 250687852.0, + "step": 9685 + }, + { + "epoch": 1.0636942675159236, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1160759925842285, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7323089838027954, + "num_tokens": 250717529.0, + "step": 9686 + }, + { + "epoch": 1.0638040852185373, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4457740783691406, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7149863243103027, + "num_tokens": 250738915.0, + "step": 9687 + }, + { + "epoch": 1.0639139029211508, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4425289630889893, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7299149036407471, + "num_tokens": 250761103.0, + "step": 9688 + }, + { + "epoch": 1.0640237206237646, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.440199136734009, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7331569194793701, + "num_tokens": 250784246.0, + "step": 9689 + }, + { + "epoch": 1.0641335383263781, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2402591705322266, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7240250110626221, + "num_tokens": 250808922.0, + "step": 9690 + }, + { + "epoch": 1.064243356028992, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.058096170425415, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7017858624458313, + "num_tokens": 250837921.0, + "step": 9691 + }, + { + "epoch": 1.0643531737316057, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2314651012420654, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7221801280975342, + "num_tokens": 250864218.0, + "step": 9692 + }, + { + "epoch": 1.0644629914342192, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1930806636810303, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7325448989868164, + "num_tokens": 250889667.0, + "step": 9693 + }, + { + "epoch": 1.064572809136833, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.742527723312378, + "learning_rate": 1e-06, + "loss": 0.7775, + "mean_token_accuracy": 0.7490636110305786, + "num_tokens": 250906367.0, + "step": 9694 + }, + { + "epoch": 1.0646826268394465, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1133975982666016, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7191929221153259, + "num_tokens": 250935459.0, + "step": 9695 + }, + { + "epoch": 1.0647924445420602, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.329428195953369, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7237293720245361, + "num_tokens": 250960205.0, + "step": 9696 + }, + { + "epoch": 1.0649022622446738, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1688311100006104, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7164748907089233, + "num_tokens": 250986395.0, + "step": 9697 + }, + { + "epoch": 1.0650120799472875, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4549167156219482, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7112768292427063, + "num_tokens": 251010206.0, + "step": 9698 + }, + { + "epoch": 1.065121897649901, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1617488861083984, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7102102041244507, + "num_tokens": 251037764.0, + "step": 9699 + }, + { + "epoch": 1.0652317153525148, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1937296390533447, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.719868540763855, + "num_tokens": 251065632.0, + "step": 9700 + }, + { + "epoch": 1.0653415330551286, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.514472723007202, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7087254524230957, + "num_tokens": 251087567.0, + "step": 9701 + }, + { + "epoch": 1.065451350757742, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.752112627029419, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7383330464363098, + "num_tokens": 251105169.0, + "step": 9702 + }, + { + "epoch": 1.0655611684603559, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2173960208892822, + "learning_rate": 1e-06, + "loss": 0.7484, + "mean_token_accuracy": 0.7600023150444031, + "num_tokens": 251129897.0, + "step": 9703 + }, + { + "epoch": 1.0656709861629694, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.5069479942321777, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7289458513259888, + "num_tokens": 251152836.0, + "step": 9704 + }, + { + "epoch": 1.0657808038655832, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3588004112243652, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7177655696868896, + "num_tokens": 251178610.0, + "step": 9705 + }, + { + "epoch": 1.0658906215681967, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.193040132522583, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7153335809707642, + "num_tokens": 251203642.0, + "step": 9706 + }, + { + "epoch": 1.0660004392708105, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2646484375, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7190204858779907, + "num_tokens": 251228970.0, + "step": 9707 + }, + { + "epoch": 1.0661102569734242, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.12126088142395, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7209730744361877, + "num_tokens": 251256481.0, + "step": 9708 + }, + { + "epoch": 1.0662200746760377, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2604329586029053, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6992471218109131, + "num_tokens": 251284419.0, + "step": 9709 + }, + { + "epoch": 1.0663298923786515, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.396045684814453, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7196906805038452, + "num_tokens": 251308932.0, + "step": 9710 + }, + { + "epoch": 1.066439710081265, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.0666513442993164, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7045760154724121, + "num_tokens": 251339253.0, + "step": 9711 + }, + { + "epoch": 1.0665495277838788, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.5807242393493652, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7259430885314941, + "num_tokens": 251360860.0, + "step": 9712 + }, + { + "epoch": 1.0666593454864923, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.234231472015381, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7271318435668945, + "num_tokens": 251385661.0, + "step": 9713 + }, + { + "epoch": 1.066769163189106, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2664053440093994, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7165701389312744, + "num_tokens": 251411174.0, + "step": 9714 + }, + { + "epoch": 1.0668789808917198, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.5068066120147705, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7222286462783813, + "num_tokens": 251433780.0, + "step": 9715 + }, + { + "epoch": 1.0669887985943334, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4545016288757324, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7158589959144592, + "num_tokens": 251457958.0, + "step": 9716 + }, + { + "epoch": 1.0670986162969471, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2328920364379883, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7192296385765076, + "num_tokens": 251483937.0, + "step": 9717 + }, + { + "epoch": 1.0672084339995607, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3548991680145264, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7259383797645569, + "num_tokens": 251506303.0, + "step": 9718 + }, + { + "epoch": 1.0673182517021744, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.356518030166626, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7079396843910217, + "num_tokens": 251531605.0, + "step": 9719 + }, + { + "epoch": 1.067428069404788, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.219766616821289, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.6924828290939331, + "num_tokens": 251559470.0, + "step": 9720 + }, + { + "epoch": 1.0675378871074017, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4984307289123535, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7435644268989563, + "num_tokens": 251580254.0, + "step": 9721 + }, + { + "epoch": 1.0676477048100155, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.179368734359741, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7149317264556885, + "num_tokens": 251606712.0, + "step": 9722 + }, + { + "epoch": 1.067757522512629, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2448551654815674, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7078021168708801, + "num_tokens": 251632873.0, + "step": 9723 + }, + { + "epoch": 1.0678673402152428, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1501803398132324, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7045201063156128, + "num_tokens": 251662345.0, + "step": 9724 + }, + { + "epoch": 1.0679771579178563, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 1.957627773284912, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7224159240722656, + "num_tokens": 251694926.0, + "step": 9725 + }, + { + "epoch": 1.06808697562047, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3952252864837646, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7209093570709229, + "num_tokens": 251717795.0, + "step": 9726 + }, + { + "epoch": 1.0681967933230836, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1098763942718506, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7250625491142273, + "num_tokens": 251744985.0, + "step": 9727 + }, + { + "epoch": 1.0683066110256974, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.268186569213867, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7105686664581299, + "num_tokens": 251769990.0, + "step": 9728 + }, + { + "epoch": 1.068416428728311, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.5011672973632812, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7048388719558716, + "num_tokens": 251793256.0, + "step": 9729 + }, + { + "epoch": 1.0685262464309246, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.060262441635132, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.6913425922393799, + "num_tokens": 251824616.0, + "step": 9730 + }, + { + "epoch": 1.0686360641335384, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.3879292011260986, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.704196572303772, + "num_tokens": 251849164.0, + "step": 9731 + }, + { + "epoch": 1.068745881836152, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.0117645263671875, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.73292076587677, + "num_tokens": 251879989.0, + "step": 9732 + }, + { + "epoch": 1.0688556995387657, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1717376708984375, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7159922122955322, + "num_tokens": 251906896.0, + "step": 9733 + }, + { + "epoch": 1.0689655172413792, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1296637058258057, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7060873508453369, + "num_tokens": 251935270.0, + "step": 9734 + }, + { + "epoch": 1.069075334943993, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.4977993965148926, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7265335321426392, + "num_tokens": 251958585.0, + "step": 9735 + }, + { + "epoch": 1.0691851526466065, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0501439571380615, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.6840598583221436, + "num_tokens": 251989265.0, + "step": 9736 + }, + { + "epoch": 1.0692949703492203, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.1073641777038574, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7375845313072205, + "num_tokens": 252015010.0, + "step": 9737 + }, + { + "epoch": 1.069404788051834, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3471791744232178, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7274891138076782, + "num_tokens": 252038183.0, + "step": 9738 + }, + { + "epoch": 1.0695146057544476, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.5245914459228516, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.705437183380127, + "num_tokens": 252061193.0, + "step": 9739 + }, + { + "epoch": 1.0696244234570613, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0543010234832764, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7164689898490906, + "num_tokens": 252091708.0, + "step": 9740 + }, + { + "epoch": 1.0697342411596749, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.597399950027466, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7370116114616394, + "num_tokens": 252110662.0, + "step": 9741 + }, + { + "epoch": 1.0698440588622886, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.100057363510132, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.719950795173645, + "num_tokens": 252138537.0, + "step": 9742 + }, + { + "epoch": 1.0699538765649024, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2313358783721924, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7163671255111694, + "num_tokens": 252164974.0, + "step": 9743 + }, + { + "epoch": 1.070063694267516, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.187504768371582, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7223584055900574, + "num_tokens": 252189644.0, + "step": 9744 + }, + { + "epoch": 1.0701735119701297, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.300114631652832, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6965383291244507, + "num_tokens": 252215816.0, + "step": 9745 + }, + { + "epoch": 1.0702833296727432, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1727566719055176, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7174229621887207, + "num_tokens": 252242546.0, + "step": 9746 + }, + { + "epoch": 1.070393147375357, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.612882137298584, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7192456722259521, + "num_tokens": 252261287.0, + "step": 9747 + }, + { + "epoch": 1.0705029650779705, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.214653730392456, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7252587080001831, + "num_tokens": 252288336.0, + "step": 9748 + }, + { + "epoch": 1.0706127827805842, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.134261131286621, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7286350131034851, + "num_tokens": 252315179.0, + "step": 9749 + }, + { + "epoch": 1.0707226004831978, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2219247817993164, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.708591878414154, + "num_tokens": 252341554.0, + "step": 9750 + }, + { + "epoch": 1.0708324181858115, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.187516927719116, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7082247138023376, + "num_tokens": 252369884.0, + "step": 9751 + }, + { + "epoch": 1.0709422358884253, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2080135345458984, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7230374217033386, + "num_tokens": 252395945.0, + "step": 9752 + }, + { + "epoch": 1.0710520535910388, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.243243932723999, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7208014726638794, + "num_tokens": 252421213.0, + "step": 9753 + }, + { + "epoch": 1.0711618712936526, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.389012575149536, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7175881266593933, + "num_tokens": 252444980.0, + "step": 9754 + }, + { + "epoch": 1.0712716889962661, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.2914340496063232, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7115379571914673, + "num_tokens": 252472709.0, + "step": 9755 + }, + { + "epoch": 1.0713815066988799, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9707845449447632, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7173411846160889, + "num_tokens": 252502923.0, + "step": 9756 + }, + { + "epoch": 1.0714913244014934, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.13655686378479, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7100504040718079, + "num_tokens": 252530288.0, + "step": 9757 + }, + { + "epoch": 1.0716011421041072, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.057363271713257, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6950724124908447, + "num_tokens": 252561438.0, + "step": 9758 + }, + { + "epoch": 1.071710959806721, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0504772663116455, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6946717500686646, + "num_tokens": 252591279.0, + "step": 9759 + }, + { + "epoch": 1.0718207775093345, + "ewc_loss": 8.678436279296875e-05, + "grad_norm": 2.1110928058624268, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7210314273834229, + "num_tokens": 252621008.0, + "step": 9760 + }, + { + "epoch": 1.0719305952119482, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.277782917022705, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7121834754943848, + "num_tokens": 252647364.0, + "step": 9761 + }, + { + "epoch": 1.0720404129145618, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3749730587005615, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7234652042388916, + "num_tokens": 252670834.0, + "step": 9762 + }, + { + "epoch": 1.0721502306171755, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.29331111907959, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7062556743621826, + "num_tokens": 252695647.0, + "step": 9763 + }, + { + "epoch": 1.072260048319789, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.04787540435791, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7198299765586853, + "num_tokens": 252725425.0, + "step": 9764 + }, + { + "epoch": 1.0723698660224028, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0352959632873535, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7008895874023438, + "num_tokens": 252755765.0, + "step": 9765 + }, + { + "epoch": 1.0724796837250166, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0749313831329346, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7201043963432312, + "num_tokens": 252784490.0, + "step": 9766 + }, + { + "epoch": 1.07258950142763, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.165363311767578, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7240861654281616, + "num_tokens": 252812843.0, + "step": 9767 + }, + { + "epoch": 1.0726993191302439, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.5900118350982666, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7203850746154785, + "num_tokens": 252834022.0, + "step": 9768 + }, + { + "epoch": 1.0728091368328574, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.299185276031494, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.719153642654419, + "num_tokens": 252860473.0, + "step": 9769 + }, + { + "epoch": 1.0729189545354711, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0579636096954346, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6905747056007385, + "num_tokens": 252892540.0, + "step": 9770 + }, + { + "epoch": 1.0730287722380847, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9455757141113281, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7208446264266968, + "num_tokens": 252925170.0, + "step": 9771 + }, + { + "epoch": 1.0731385899406984, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.47687029838562, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7237945795059204, + "num_tokens": 252946526.0, + "step": 9772 + }, + { + "epoch": 1.0732484076433122, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.4403269290924072, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7000315189361572, + "num_tokens": 252974024.0, + "step": 9773 + }, + { + "epoch": 1.0733582253459257, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.4477622509002686, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7207837104797363, + "num_tokens": 252997552.0, + "step": 9774 + }, + { + "epoch": 1.0734680430485395, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.228595733642578, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7173006534576416, + "num_tokens": 253024018.0, + "step": 9775 + }, + { + "epoch": 1.073577860751153, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.8150389194488525, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.745716392993927, + "num_tokens": 253059439.0, + "step": 9776 + }, + { + "epoch": 1.0736876784537668, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.8925100564956665, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7075783610343933, + "num_tokens": 253092060.0, + "step": 9777 + }, + { + "epoch": 1.0737974961563803, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2706847190856934, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7207677960395813, + "num_tokens": 253116732.0, + "step": 9778 + }, + { + "epoch": 1.073907313858994, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.4881303310394287, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7124080657958984, + "num_tokens": 253137444.0, + "step": 9779 + }, + { + "epoch": 1.0740171315616078, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.341890335083008, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7099982500076294, + "num_tokens": 253163728.0, + "step": 9780 + }, + { + "epoch": 1.0741269492642214, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.553084135055542, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7158333659172058, + "num_tokens": 253185007.0, + "step": 9781 + }, + { + "epoch": 1.0742367669668351, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.236774444580078, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7169091701507568, + "num_tokens": 253210039.0, + "step": 9782 + }, + { + "epoch": 1.0743465846694487, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3088293075561523, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7122175097465515, + "num_tokens": 253233730.0, + "step": 9783 + }, + { + "epoch": 1.0744564023720624, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.516007900238037, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.746039867401123, + "num_tokens": 253253245.0, + "step": 9784 + }, + { + "epoch": 1.074566220074676, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3414368629455566, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7392407655715942, + "num_tokens": 253276720.0, + "step": 9785 + }, + { + "epoch": 1.0746760377772897, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.072617530822754, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.712644636631012, + "num_tokens": 253304939.0, + "step": 9786 + }, + { + "epoch": 1.0747858554799035, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.028047800064087, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.7020477056503296, + "num_tokens": 253337224.0, + "step": 9787 + }, + { + "epoch": 1.074895673182517, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.4162817001342773, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.718049943447113, + "num_tokens": 253361449.0, + "step": 9788 + }, + { + "epoch": 1.0750054908851308, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.5487546920776367, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.73289555311203, + "num_tokens": 253382157.0, + "step": 9789 + }, + { + "epoch": 1.0751153085877443, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9368555545806885, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7079564332962036, + "num_tokens": 253416618.0, + "step": 9790 + }, + { + "epoch": 1.075225126290358, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.436527729034424, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7297057509422302, + "num_tokens": 253437656.0, + "step": 9791 + }, + { + "epoch": 1.0753349439929716, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0879507064819336, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7095769047737122, + "num_tokens": 253465676.0, + "step": 9792 + }, + { + "epoch": 1.0754447616955853, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3837332725524902, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7191824913024902, + "num_tokens": 253488247.0, + "step": 9793 + }, + { + "epoch": 1.075554579398199, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.521728515625, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7161546945571899, + "num_tokens": 253509582.0, + "step": 9794 + }, + { + "epoch": 1.0756643971008126, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3532276153564453, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7107338309288025, + "num_tokens": 253533499.0, + "step": 9795 + }, + { + "epoch": 1.0757742148034264, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.1497652530670166, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7205075025558472, + "num_tokens": 253559942.0, + "step": 9796 + }, + { + "epoch": 1.07588403250604, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2158076763153076, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7239181995391846, + "num_tokens": 253587048.0, + "step": 9797 + }, + { + "epoch": 1.0759938502086537, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.3312911987304688, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7129284739494324, + "num_tokens": 253614210.0, + "step": 9798 + }, + { + "epoch": 1.0761036679112672, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9883421659469604, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7163942456245422, + "num_tokens": 253645388.0, + "step": 9799 + }, + { + "epoch": 1.076213485613881, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.509495735168457, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7290570735931396, + "num_tokens": 253666976.0, + "step": 9800 + }, + { + "epoch": 1.0763233033164945, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.10027813911438, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6944284439086914, + "num_tokens": 253695884.0, + "step": 9801 + }, + { + "epoch": 1.0764331210191083, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.8748116493225098, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6982618570327759, + "num_tokens": 253730963.0, + "step": 9802 + }, + { + "epoch": 1.076542938721722, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.360527753829956, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.724574089050293, + "num_tokens": 253754933.0, + "step": 9803 + }, + { + "epoch": 1.0766527564243356, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.260640859603882, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7066418528556824, + "num_tokens": 253782586.0, + "step": 9804 + }, + { + "epoch": 1.0767625741269493, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9965846538543701, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7249065637588501, + "num_tokens": 253816135.0, + "step": 9805 + }, + { + "epoch": 1.0768723918295628, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2706856727600098, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7360943555831909, + "num_tokens": 253840466.0, + "step": 9806 + }, + { + "epoch": 1.0769822095321766, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.148935317993164, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7222203016281128, + "num_tokens": 253866512.0, + "step": 9807 + }, + { + "epoch": 1.0770920272347904, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2856297492980957, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7390186786651611, + "num_tokens": 253890883.0, + "step": 9808 + }, + { + "epoch": 1.077201844937404, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.4623680114746094, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7242807745933533, + "num_tokens": 253913297.0, + "step": 9809 + }, + { + "epoch": 1.0773116626400177, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.145112991333008, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7124781608581543, + "num_tokens": 253941360.0, + "step": 9810 + }, + { + "epoch": 1.0774214803426312, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.0151515007019043, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6920902132987976, + "num_tokens": 253971636.0, + "step": 9811 + }, + { + "epoch": 1.077531298045245, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.322817802429199, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.718802809715271, + "num_tokens": 253996545.0, + "step": 9812 + }, + { + "epoch": 1.0776411157478585, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 1.9593441486358643, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.699940025806427, + "num_tokens": 254030517.0, + "step": 9813 + }, + { + "epoch": 1.0777509334504722, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.5978665351867676, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7130739688873291, + "num_tokens": 254052332.0, + "step": 9814 + }, + { + "epoch": 1.0778607511530858, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.386328935623169, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7196947932243347, + "num_tokens": 254075027.0, + "step": 9815 + }, + { + "epoch": 1.0779705688556995, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.1754097938537598, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7053666114807129, + "num_tokens": 254105596.0, + "step": 9816 + }, + { + "epoch": 1.0780803865583133, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.333829402923584, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7164273262023926, + "num_tokens": 254129912.0, + "step": 9817 + }, + { + "epoch": 1.0781902042609268, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.335906505584717, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7395051121711731, + "num_tokens": 254152853.0, + "step": 9818 + }, + { + "epoch": 1.0783000219635406, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2304744720458984, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7154536247253418, + "num_tokens": 254179367.0, + "step": 9819 + }, + { + "epoch": 1.0784098396661541, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.1958775520324707, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7025264501571655, + "num_tokens": 254205615.0, + "step": 9820 + }, + { + "epoch": 1.0785196573687679, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.1579833030700684, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7298961877822876, + "num_tokens": 254232185.0, + "step": 9821 + }, + { + "epoch": 1.0786294750713814, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.2122321128845215, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7075710296630859, + "num_tokens": 254258918.0, + "step": 9822 + }, + { + "epoch": 1.0787392927739952, + "ewc_loss": 8.726119995117188e-05, + "grad_norm": 2.50141978263855, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7308410406112671, + "num_tokens": 254281886.0, + "step": 9823 + }, + { + "epoch": 1.078849110476609, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.064692497253418, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7052635550498962, + "num_tokens": 254312053.0, + "step": 9824 + }, + { + "epoch": 1.0789589281792225, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3542847633361816, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7297245264053345, + "num_tokens": 254335024.0, + "step": 9825 + }, + { + "epoch": 1.0790687458818362, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.168121814727783, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7128655910491943, + "num_tokens": 254361962.0, + "step": 9826 + }, + { + "epoch": 1.0791785635844497, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 1.8603836297988892, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7140176296234131, + "num_tokens": 254396842.0, + "step": 9827 + }, + { + "epoch": 1.0792883812870635, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2577707767486572, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7146387100219727, + "num_tokens": 254422731.0, + "step": 9828 + }, + { + "epoch": 1.079398198989677, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1625263690948486, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7339578866958618, + "num_tokens": 254448617.0, + "step": 9829 + }, + { + "epoch": 1.0795080166922908, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.573986053466797, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7067274451255798, + "num_tokens": 254468811.0, + "step": 9830 + }, + { + "epoch": 1.0796178343949046, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3600990772247314, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7130520939826965, + "num_tokens": 254492675.0, + "step": 9831 + }, + { + "epoch": 1.079727652097518, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.6511714458465576, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7216856479644775, + "num_tokens": 254512410.0, + "step": 9832 + }, + { + "epoch": 1.0798374698001318, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3841333389282227, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7167016267776489, + "num_tokens": 254535716.0, + "step": 9833 + }, + { + "epoch": 1.0799472875027454, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.5037996768951416, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.728931725025177, + "num_tokens": 254557037.0, + "step": 9834 + }, + { + "epoch": 1.0800571052053591, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2400307655334473, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6933790445327759, + "num_tokens": 254584259.0, + "step": 9835 + }, + { + "epoch": 1.0801669229079727, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2492516040802, + "learning_rate": 1e-06, + "loss": 1.032, + "mean_token_accuracy": 0.6807295680046082, + "num_tokens": 254614845.0, + "step": 9836 + }, + { + "epoch": 1.0802767406105864, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.206530809402466, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7271072864532471, + "num_tokens": 254640490.0, + "step": 9837 + }, + { + "epoch": 1.0803865583132002, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1813619136810303, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7198198437690735, + "num_tokens": 254667390.0, + "step": 9838 + }, + { + "epoch": 1.0804963760158137, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.4311516284942627, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7040233612060547, + "num_tokens": 254690296.0, + "step": 9839 + }, + { + "epoch": 1.0806061937184275, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.400054693222046, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7046176195144653, + "num_tokens": 254715809.0, + "step": 9840 + }, + { + "epoch": 1.080716011421041, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2776474952697754, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7485899925231934, + "num_tokens": 254741448.0, + "step": 9841 + }, + { + "epoch": 1.0808258291236548, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2224276065826416, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.7063815593719482, + "num_tokens": 254767857.0, + "step": 9842 + }, + { + "epoch": 1.0809356468262683, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3437294960021973, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7232183218002319, + "num_tokens": 254792749.0, + "step": 9843 + }, + { + "epoch": 1.081045464528882, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.010610342025757, + "learning_rate": 1e-06, + "loss": 1.0369, + "mean_token_accuracy": 0.6939982771873474, + "num_tokens": 254827509.0, + "step": 9844 + }, + { + "epoch": 1.0811552822314958, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.269937038421631, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7453357577323914, + "num_tokens": 254849998.0, + "step": 9845 + }, + { + "epoch": 1.0812650999341094, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.464867353439331, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7078734636306763, + "num_tokens": 254873692.0, + "step": 9846 + }, + { + "epoch": 1.081374917636723, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1411139965057373, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.728623628616333, + "num_tokens": 254901458.0, + "step": 9847 + }, + { + "epoch": 1.0814847353393366, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.4251396656036377, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7198917269706726, + "num_tokens": 254922737.0, + "step": 9848 + }, + { + "epoch": 1.0815945530419504, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1964855194091797, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7067924737930298, + "num_tokens": 254951216.0, + "step": 9849 + }, + { + "epoch": 1.081704370744564, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3696227073669434, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7145323753356934, + "num_tokens": 254975725.0, + "step": 9850 + }, + { + "epoch": 1.0818141884471777, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 1.8810443878173828, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.699414074420929, + "num_tokens": 255012158.0, + "step": 9851 + }, + { + "epoch": 1.0819240061497912, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.808969020843506, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.740347146987915, + "num_tokens": 255028102.0, + "step": 9852 + }, + { + "epoch": 1.082033823852405, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2309508323669434, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7282793521881104, + "num_tokens": 255053098.0, + "step": 9853 + }, + { + "epoch": 1.0821436415550187, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1364505290985107, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7514050006866455, + "num_tokens": 255079204.0, + "step": 9854 + }, + { + "epoch": 1.0822534592576323, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 1.9702816009521484, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.725832462310791, + "num_tokens": 255109320.0, + "step": 9855 + }, + { + "epoch": 1.082363276960246, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 1.9255049228668213, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6963778734207153, + "num_tokens": 255142634.0, + "step": 9856 + }, + { + "epoch": 1.0824730946628596, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.3052687644958496, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7437673807144165, + "num_tokens": 255166313.0, + "step": 9857 + }, + { + "epoch": 1.0825829123654733, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.2509870529174805, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7145633101463318, + "num_tokens": 255194625.0, + "step": 9858 + }, + { + "epoch": 1.082692730068087, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1441690921783447, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7315418720245361, + "num_tokens": 255220850.0, + "step": 9859 + }, + { + "epoch": 1.0828025477707006, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.5273561477661133, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7256009578704834, + "num_tokens": 255243028.0, + "step": 9860 + }, + { + "epoch": 1.0829123654733144, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.189786672592163, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.693589985370636, + "num_tokens": 255272468.0, + "step": 9861 + }, + { + "epoch": 1.083022183175928, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.4195716381073, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7183991074562073, + "num_tokens": 255296927.0, + "step": 9862 + }, + { + "epoch": 1.0831320008785417, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.165623426437378, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7401261329650879, + "num_tokens": 255323354.0, + "step": 9863 + }, + { + "epoch": 1.0832418185811552, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.016173839569092, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.6977337002754211, + "num_tokens": 255356309.0, + "step": 9864 + }, + { + "epoch": 1.083351636283769, + "ewc_loss": 8.7738037109375e-05, + "grad_norm": 2.1484992504119873, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7146697640419006, + "num_tokens": 255383162.0, + "step": 9865 + }, + { + "epoch": 1.0834614539863825, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.3103513717651367, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7050776481628418, + "num_tokens": 255408460.0, + "step": 9866 + }, + { + "epoch": 1.0835712716889963, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.2465970516204834, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.718328595161438, + "num_tokens": 255434169.0, + "step": 9867 + }, + { + "epoch": 1.08368108939161, + "ewc_loss": 8.869171142578125e-05, + "grad_norm": 2.5417277812957764, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7077286839485168, + "num_tokens": 255454970.0, + "step": 9868 + }, + { + "epoch": 1.0837909070942235, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.448404312133789, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7400147914886475, + "num_tokens": 255475839.0, + "step": 9869 + }, + { + "epoch": 1.0839007247968373, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.1036951541900635, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6976560950279236, + "num_tokens": 255505850.0, + "step": 9870 + }, + { + "epoch": 1.0840105424994508, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.1404507160186768, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6993774175643921, + "num_tokens": 255536149.0, + "step": 9871 + }, + { + "epoch": 1.0841203602020646, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.277019739151001, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7228971719741821, + "num_tokens": 255561334.0, + "step": 9872 + }, + { + "epoch": 1.0842301779046783, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.14812970161438, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7144246101379395, + "num_tokens": 255589566.0, + "step": 9873 + }, + { + "epoch": 1.0843399956072919, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.3151004314422607, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7394106388092041, + "num_tokens": 255613867.0, + "step": 9874 + }, + { + "epoch": 1.0844498133099056, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.617445468902588, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7272690534591675, + "num_tokens": 255634796.0, + "step": 9875 + }, + { + "epoch": 1.0845596310125192, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 1.9781920909881592, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7000318765640259, + "num_tokens": 255666912.0, + "step": 9876 + }, + { + "epoch": 1.084669448715133, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.235347270965576, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7080780863761902, + "num_tokens": 255693497.0, + "step": 9877 + }, + { + "epoch": 1.0847792664177465, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.2216684818267822, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7079634070396423, + "num_tokens": 255719485.0, + "step": 9878 + }, + { + "epoch": 1.0848890841203602, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.2663683891296387, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7182222604751587, + "num_tokens": 255743656.0, + "step": 9879 + }, + { + "epoch": 1.0849989018229738, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 1.9984925985336304, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7103443145751953, + "num_tokens": 255774053.0, + "step": 9880 + }, + { + "epoch": 1.0851087195255875, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.65618634223938, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7025455236434937, + "num_tokens": 255794035.0, + "step": 9881 + }, + { + "epoch": 1.0852185372282013, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.163855791091919, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7477679252624512, + "num_tokens": 255819485.0, + "step": 9882 + }, + { + "epoch": 1.0853283549308148, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.35058331489563, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6907142400741577, + "num_tokens": 255844570.0, + "step": 9883 + }, + { + "epoch": 1.0854381726334286, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.1940643787384033, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6937668323516846, + "num_tokens": 255873762.0, + "step": 9884 + }, + { + "epoch": 1.085547990336042, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.4726762771606445, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7185983657836914, + "num_tokens": 255896096.0, + "step": 9885 + }, + { + "epoch": 1.0856578080386559, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.0417630672454834, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6877595782279968, + "num_tokens": 255927815.0, + "step": 9886 + }, + { + "epoch": 1.0857676257412694, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.371134042739868, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7308937311172485, + "num_tokens": 255949988.0, + "step": 9887 + }, + { + "epoch": 1.0858774434438832, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.5571398735046387, + "learning_rate": 1e-06, + "loss": 0.7731, + "mean_token_accuracy": 0.7500469088554382, + "num_tokens": 255968301.0, + "step": 9888 + }, + { + "epoch": 1.085987261146497, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.3946821689605713, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7194615006446838, + "num_tokens": 255991431.0, + "step": 9889 + }, + { + "epoch": 1.0860970788491104, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.029552698135376, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6951411962509155, + "num_tokens": 256023317.0, + "step": 9890 + }, + { + "epoch": 1.0862068965517242, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.832031726837158, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7178064584732056, + "num_tokens": 256042239.0, + "step": 9891 + }, + { + "epoch": 1.0863167142543377, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.201740026473999, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7414937019348145, + "num_tokens": 256069235.0, + "step": 9892 + }, + { + "epoch": 1.0864265319569515, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.2906341552734375, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7231059074401855, + "num_tokens": 256094727.0, + "step": 9893 + }, + { + "epoch": 1.086536349659565, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.428676128387451, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.732856273651123, + "num_tokens": 256116410.0, + "step": 9894 + }, + { + "epoch": 1.0866461673621788, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.2974414825439453, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7326443791389465, + "num_tokens": 256140523.0, + "step": 9895 + }, + { + "epoch": 1.0867559850647925, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.0304720401763916, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7301578521728516, + "num_tokens": 256173084.0, + "step": 9896 + }, + { + "epoch": 1.086865802767406, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.423156976699829, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.707947313785553, + "num_tokens": 256197211.0, + "step": 9897 + }, + { + "epoch": 1.0869756204700198, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.3342413902282715, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7257976531982422, + "num_tokens": 256219446.0, + "step": 9898 + }, + { + "epoch": 1.0870854381726334, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.5167911052703857, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7397712469100952, + "num_tokens": 256240081.0, + "step": 9899 + }, + { + "epoch": 1.0871952558752471, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.0673935413360596, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7310550212860107, + "num_tokens": 256268252.0, + "step": 9900 + }, + { + "epoch": 1.0873050735778607, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.5515856742858887, + "learning_rate": 1e-06, + "loss": 0.8012, + "mean_token_accuracy": 0.7462335824966431, + "num_tokens": 256288236.0, + "step": 9901 + }, + { + "epoch": 1.0874148912804744, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.597015857696533, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7345007061958313, + "num_tokens": 256307306.0, + "step": 9902 + }, + { + "epoch": 1.0875247089830882, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.0793542861938477, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6928303837776184, + "num_tokens": 256336305.0, + "step": 9903 + }, + { + "epoch": 1.0876345266857017, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.319587230682373, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.737450122833252, + "num_tokens": 256360289.0, + "step": 9904 + }, + { + "epoch": 1.0877443443883155, + "ewc_loss": 8.916854858398438e-05, + "grad_norm": 2.0108866691589355, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7210183143615723, + "num_tokens": 256390078.0, + "step": 9905 + }, + { + "epoch": 1.087854162090929, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.092186450958252, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7109501361846924, + "num_tokens": 256420136.0, + "step": 9906 + }, + { + "epoch": 1.0879639797935428, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.2881977558135986, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7220728993415833, + "num_tokens": 256448256.0, + "step": 9907 + }, + { + "epoch": 1.0880737974961563, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.0869929790496826, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7049626708030701, + "num_tokens": 256477398.0, + "step": 9908 + }, + { + "epoch": 1.08818361519877, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.404219388961792, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7352207899093628, + "num_tokens": 256499772.0, + "step": 9909 + }, + { + "epoch": 1.0882934329013838, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.1380722522735596, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7149226665496826, + "num_tokens": 256528746.0, + "step": 9910 + }, + { + "epoch": 1.0884032506039973, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.4790592193603516, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6846078634262085, + "num_tokens": 256553700.0, + "step": 9911 + }, + { + "epoch": 1.088513068306611, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.1337473392486572, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7255890965461731, + "num_tokens": 256581267.0, + "step": 9912 + }, + { + "epoch": 1.0886228860092246, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.1982674598693848, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7014731764793396, + "num_tokens": 256606458.0, + "step": 9913 + }, + { + "epoch": 1.0887327037118384, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.342703104019165, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7242223024368286, + "num_tokens": 256631026.0, + "step": 9914 + }, + { + "epoch": 1.088842521414452, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.0646324157714844, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7082144618034363, + "num_tokens": 256659438.0, + "step": 9915 + }, + { + "epoch": 1.0889523391170657, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.175217866897583, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6884369254112244, + "num_tokens": 256690266.0, + "step": 9916 + }, + { + "epoch": 1.0890621568196792, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.110125780105591, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7304961085319519, + "num_tokens": 256719995.0, + "step": 9917 + }, + { + "epoch": 1.089171974522293, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.5193471908569336, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.712878406047821, + "num_tokens": 256741027.0, + "step": 9918 + }, + { + "epoch": 1.0892817922249067, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.3920345306396484, + "learning_rate": 1e-06, + "loss": 1.0121, + "mean_token_accuracy": 0.6872832775115967, + "num_tokens": 256764856.0, + "step": 9919 + }, + { + "epoch": 1.0893916099275203, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.170452833175659, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.7041258811950684, + "num_tokens": 256791611.0, + "step": 9920 + }, + { + "epoch": 1.089501427630134, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.345075845718384, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7126560211181641, + "num_tokens": 256816799.0, + "step": 9921 + }, + { + "epoch": 1.0896112453327476, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.1279029846191406, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6956222057342529, + "num_tokens": 256844587.0, + "step": 9922 + }, + { + "epoch": 1.0897210630353613, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.2407338619232178, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.717728853225708, + "num_tokens": 256869769.0, + "step": 9923 + }, + { + "epoch": 1.089830880737975, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.060509204864502, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6931227445602417, + "num_tokens": 256902427.0, + "step": 9924 + }, + { + "epoch": 1.0899406984405886, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.268495559692383, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7137022614479065, + "num_tokens": 256928302.0, + "step": 9925 + }, + { + "epoch": 1.0900505161432024, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.550612449645996, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7101285457611084, + "num_tokens": 256949129.0, + "step": 9926 + }, + { + "epoch": 1.090160333845816, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.120530366897583, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7172650098800659, + "num_tokens": 256979446.0, + "step": 9927 + }, + { + "epoch": 1.0902701515484297, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.2414710521698, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7196623682975769, + "num_tokens": 257006538.0, + "step": 9928 + }, + { + "epoch": 1.0903799692510432, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.273968458175659, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7361934185028076, + "num_tokens": 257030709.0, + "step": 9929 + }, + { + "epoch": 1.090489786953657, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.3224494457244873, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7283698320388794, + "num_tokens": 257054215.0, + "step": 9930 + }, + { + "epoch": 1.0905996046562705, + "ewc_loss": 9.012222290039062e-05, + "grad_norm": 2.1634793281555176, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.721758246421814, + "num_tokens": 257080815.0, + "step": 9931 + }, + { + "epoch": 1.0907094223588842, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.197505235671997, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7262582778930664, + "num_tokens": 257105877.0, + "step": 9932 + }, + { + "epoch": 1.090819240061498, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.465322256088257, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7215664386749268, + "num_tokens": 257128601.0, + "step": 9933 + }, + { + "epoch": 1.0909290577641115, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.4946346282958984, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7305862903594971, + "num_tokens": 257149597.0, + "step": 9934 + }, + { + "epoch": 1.0910388754667253, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.4579720497131348, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7119095325469971, + "num_tokens": 257171312.0, + "step": 9935 + }, + { + "epoch": 1.0911486931693388, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.0016489028930664, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.706116795539856, + "num_tokens": 257204631.0, + "step": 9936 + }, + { + "epoch": 1.0912585108719526, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.358137369155884, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.7039414048194885, + "num_tokens": 257228960.0, + "step": 9937 + }, + { + "epoch": 1.0913683285745663, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.4038920402526855, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.727310061454773, + "num_tokens": 257251455.0, + "step": 9938 + }, + { + "epoch": 1.0914781462771799, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.346372127532959, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7177064418792725, + "num_tokens": 257277132.0, + "step": 9939 + }, + { + "epoch": 1.0915879639797936, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.3024933338165283, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.714503288269043, + "num_tokens": 257301807.0, + "step": 9940 + }, + { + "epoch": 1.0916977816824072, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 1.9667069911956787, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7085310220718384, + "num_tokens": 257333641.0, + "step": 9941 + }, + { + "epoch": 1.091807599385021, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.0971193313598633, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7279199957847595, + "num_tokens": 257361424.0, + "step": 9942 + }, + { + "epoch": 1.0919174170876345, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.451073169708252, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7138791680335999, + "num_tokens": 257383212.0, + "step": 9943 + }, + { + "epoch": 1.0920272347902482, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.0931291580200195, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7105304598808289, + "num_tokens": 257412798.0, + "step": 9944 + }, + { + "epoch": 1.0921370524928617, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.205648183822632, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7131529450416565, + "num_tokens": 257440412.0, + "step": 9945 + }, + { + "epoch": 1.0922468701954755, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.2121188640594482, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7208080291748047, + "num_tokens": 257468699.0, + "step": 9946 + }, + { + "epoch": 1.0923566878980893, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.059573173522949, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7056195735931396, + "num_tokens": 257497606.0, + "step": 9947 + }, + { + "epoch": 1.0924665056007028, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.3722665309906006, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7307015657424927, + "num_tokens": 257520184.0, + "step": 9948 + }, + { + "epoch": 1.0925763233033166, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.166135549545288, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6965644955635071, + "num_tokens": 257550058.0, + "step": 9949 + }, + { + "epoch": 1.09268614100593, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.1740224361419678, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7120242118835449, + "num_tokens": 257577836.0, + "step": 9950 + }, + { + "epoch": 1.0927959587085438, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 1.8215267658233643, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7037126421928406, + "num_tokens": 257614681.0, + "step": 9951 + }, + { + "epoch": 1.0929057764111574, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.308403730392456, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7188865542411804, + "num_tokens": 257638113.0, + "step": 9952 + }, + { + "epoch": 1.0930155941137711, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.2850730419158936, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7317343950271606, + "num_tokens": 257661459.0, + "step": 9953 + }, + { + "epoch": 1.093125411816385, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 1.9463664293289185, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6995105743408203, + "num_tokens": 257694517.0, + "step": 9954 + }, + { + "epoch": 1.0932352295189984, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.173821210861206, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7289192080497742, + "num_tokens": 257721198.0, + "step": 9955 + }, + { + "epoch": 1.0933450472216122, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.3633902072906494, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7411867380142212, + "num_tokens": 257741074.0, + "step": 9956 + }, + { + "epoch": 1.0934548649242257, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.235661029815674, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7074844837188721, + "num_tokens": 257767076.0, + "step": 9957 + }, + { + "epoch": 1.0935646826268395, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.0299901962280273, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.6915345788002014, + "num_tokens": 257798867.0, + "step": 9958 + }, + { + "epoch": 1.093674500329453, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.4678256511688232, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7294204235076904, + "num_tokens": 257821000.0, + "step": 9959 + }, + { + "epoch": 1.0937843180320668, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.210080146789551, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7367229461669922, + "num_tokens": 257845751.0, + "step": 9960 + }, + { + "epoch": 1.0938941357346805, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.242654800415039, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7240096926689148, + "num_tokens": 257870734.0, + "step": 9961 + }, + { + "epoch": 1.094003953437294, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.46067476272583, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7105168104171753, + "num_tokens": 257892104.0, + "step": 9962 + }, + { + "epoch": 1.0941137711399078, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.2688522338867188, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7065634727478027, + "num_tokens": 257918169.0, + "step": 9963 + }, + { + "epoch": 1.0942235888425214, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.6243927478790283, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7253981232643127, + "num_tokens": 257939361.0, + "step": 9964 + }, + { + "epoch": 1.0943334065451351, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.419069290161133, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7205005884170532, + "num_tokens": 257962990.0, + "step": 9965 + }, + { + "epoch": 1.0944432242477486, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.862229824066162, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7513442635536194, + "num_tokens": 257980096.0, + "step": 9966 + }, + { + "epoch": 1.0945530419503624, + "ewc_loss": 9.059906005859375e-05, + "grad_norm": 2.126910448074341, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7392174005508423, + "num_tokens": 258007207.0, + "step": 9967 + }, + { + "epoch": 1.0946628596529762, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 1.8328934907913208, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.7008907794952393, + "num_tokens": 258043075.0, + "step": 9968 + }, + { + "epoch": 1.0947726773555897, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 1.995434045791626, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7384375333786011, + "num_tokens": 258073828.0, + "step": 9969 + }, + { + "epoch": 1.0948824950582035, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.3134827613830566, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7043758630752563, + "num_tokens": 258100626.0, + "step": 9970 + }, + { + "epoch": 1.094992312760817, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.1846396923065186, + "learning_rate": 1e-06, + "loss": 0.7901, + "mean_token_accuracy": 0.7540555596351624, + "num_tokens": 258125951.0, + "step": 9971 + }, + { + "epoch": 1.0951021304634307, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.1033849716186523, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7364040017127991, + "num_tokens": 258151984.0, + "step": 9972 + }, + { + "epoch": 1.0952119481660443, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2490196228027344, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7008876800537109, + "num_tokens": 258177525.0, + "step": 9973 + }, + { + "epoch": 1.095321765868658, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.3042829036712646, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7130919694900513, + "num_tokens": 258203254.0, + "step": 9974 + }, + { + "epoch": 1.0954315835712718, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2758796215057373, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7102826833724976, + "num_tokens": 258230192.0, + "step": 9975 + }, + { + "epoch": 1.0955414012738853, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.3984010219573975, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7010231018066406, + "num_tokens": 258253607.0, + "step": 9976 + }, + { + "epoch": 1.095651218976499, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.4773123264312744, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7188910245895386, + "num_tokens": 258276578.0, + "step": 9977 + }, + { + "epoch": 1.0957610366791126, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.313699245452881, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7141562700271606, + "num_tokens": 258302613.0, + "step": 9978 + }, + { + "epoch": 1.0958708543817264, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 1.8544062376022339, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7066942453384399, + "num_tokens": 258337946.0, + "step": 9979 + }, + { + "epoch": 1.09598067208434, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.009826898574829, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.702969491481781, + "num_tokens": 258369426.0, + "step": 9980 + }, + { + "epoch": 1.0960904897869537, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.4903564453125, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7233555316925049, + "num_tokens": 258391493.0, + "step": 9981 + }, + { + "epoch": 1.0962003074895672, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.1926815509796143, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.714943528175354, + "num_tokens": 258417199.0, + "step": 9982 + }, + { + "epoch": 1.096310125192181, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.495073080062866, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7146519422531128, + "num_tokens": 258438100.0, + "step": 9983 + }, + { + "epoch": 1.0964199428947947, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.2217986583709717, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.741798996925354, + "num_tokens": 258463474.0, + "step": 9984 + }, + { + "epoch": 1.0965297605974083, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2932372093200684, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7064377069473267, + "num_tokens": 258488155.0, + "step": 9985 + }, + { + "epoch": 1.096639578300022, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.3199212551116943, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7234113216400146, + "num_tokens": 258511403.0, + "step": 9986 + }, + { + "epoch": 1.0967493960026355, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.4759461879730225, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7036230564117432, + "num_tokens": 258533843.0, + "step": 9987 + }, + { + "epoch": 1.0968592137052493, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2158586978912354, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.721276581287384, + "num_tokens": 258559380.0, + "step": 9988 + }, + { + "epoch": 1.096969031407863, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2316765785217285, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.6807365417480469, + "num_tokens": 258589868.0, + "step": 9989 + }, + { + "epoch": 1.0970788491104766, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.401090383529663, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7375096678733826, + "num_tokens": 258611281.0, + "step": 9990 + }, + { + "epoch": 1.0971886668130904, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 2.2231433391571045, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7004810571670532, + "num_tokens": 258639167.0, + "step": 9991 + }, + { + "epoch": 1.0972984845157039, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.227194309234619, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7185636162757874, + "num_tokens": 258665202.0, + "step": 9992 + }, + { + "epoch": 1.0974083022183176, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.5209758281707764, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7314613461494446, + "num_tokens": 258686204.0, + "step": 9993 + }, + { + "epoch": 1.0975181199209312, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.1358108520507812, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.730161190032959, + "num_tokens": 258712260.0, + "step": 9994 + }, + { + "epoch": 1.097627937623545, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.6383848190307617, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7259589433670044, + "num_tokens": 258732364.0, + "step": 9995 + }, + { + "epoch": 1.0977377553261585, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.284942150115967, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7495402097702026, + "num_tokens": 258758137.0, + "step": 9996 + }, + { + "epoch": 1.0978475730287722, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.399648666381836, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7181513905525208, + "num_tokens": 258781894.0, + "step": 9997 + }, + { + "epoch": 1.097957390731386, + "ewc_loss": 9.107589721679688e-05, + "grad_norm": 1.9454234838485718, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7272019386291504, + "num_tokens": 258816433.0, + "step": 9998 + }, + { + "epoch": 1.0980672084339995, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.288405179977417, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7046118974685669, + "num_tokens": 258842293.0, + "step": 9999 + }, + { + "epoch": 1.0981770261366133, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3423917293548584, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7116997241973877, + "num_tokens": 258868498.0, + "step": 10000 + }, + { + "epoch": 1.0982868438392268, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3028452396392822, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7131540775299072, + "num_tokens": 258894056.0, + "step": 10001 + }, + { + "epoch": 1.0983966615418406, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.657031297683716, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.744092583656311, + "num_tokens": 258913204.0, + "step": 10002 + }, + { + "epoch": 1.098506479244454, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.5133402347564697, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7495204210281372, + "num_tokens": 258934560.0, + "step": 10003 + }, + { + "epoch": 1.0986162969470679, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.5473201274871826, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7387876510620117, + "num_tokens": 258957202.0, + "step": 10004 + }, + { + "epoch": 1.0987261146496816, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.5695877075195312, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7244856357574463, + "num_tokens": 258978611.0, + "step": 10005 + }, + { + "epoch": 1.0988359323522952, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3023064136505127, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7146021723747253, + "num_tokens": 259004896.0, + "step": 10006 + }, + { + "epoch": 1.098945750054909, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3871095180511475, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7133975028991699, + "num_tokens": 259029899.0, + "step": 10007 + }, + { + "epoch": 1.0990555677575224, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.0881354808807373, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7259157299995422, + "num_tokens": 259058605.0, + "step": 10008 + }, + { + "epoch": 1.0991653854601362, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3216750621795654, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7180142402648926, + "num_tokens": 259082524.0, + "step": 10009 + }, + { + "epoch": 1.0992752031627497, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.3026282787323, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7214939594268799, + "num_tokens": 259107190.0, + "step": 10010 + }, + { + "epoch": 1.0993850208653635, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 1.9567466974258423, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7182028293609619, + "num_tokens": 259140725.0, + "step": 10011 + }, + { + "epoch": 1.0994948385679773, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3027584552764893, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7010166645050049, + "num_tokens": 259166114.0, + "step": 10012 + }, + { + "epoch": 1.0996046562705908, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3507349491119385, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7206772565841675, + "num_tokens": 259188512.0, + "step": 10013 + }, + { + "epoch": 1.0997144739732045, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.1784467697143555, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7081803679466248, + "num_tokens": 259213322.0, + "step": 10014 + }, + { + "epoch": 1.099824291675818, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 1.998414397239685, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.711733341217041, + "num_tokens": 259244247.0, + "step": 10015 + }, + { + "epoch": 1.0999341093784318, + "ewc_loss": 9.1552734375e-05, + "grad_norm": 2.115605115890503, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7264624834060669, + "num_tokens": 259272335.0, + "step": 10016 + }, + { + "epoch": 1.1000439270810454, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.128509521484375, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7239780426025391, + "num_tokens": 259299824.0, + "step": 10017 + }, + { + "epoch": 1.1001537447836591, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.244490623474121, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7307168841362, + "num_tokens": 259322970.0, + "step": 10018 + }, + { + "epoch": 1.1002635624862729, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.174417495727539, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7147340178489685, + "num_tokens": 259352116.0, + "step": 10019 + }, + { + "epoch": 1.1003733801888864, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.0644407272338867, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.6966867446899414, + "num_tokens": 259383382.0, + "step": 10020 + }, + { + "epoch": 1.1004831978915002, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 1.9709385633468628, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7174527645111084, + "num_tokens": 259412079.0, + "step": 10021 + }, + { + "epoch": 1.1005930155941137, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 1.9771928787231445, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.7126656174659729, + "num_tokens": 259445070.0, + "step": 10022 + }, + { + "epoch": 1.1007028332967275, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2265894412994385, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7183836102485657, + "num_tokens": 259473069.0, + "step": 10023 + }, + { + "epoch": 1.100812650999341, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2455382347106934, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7143997550010681, + "num_tokens": 259498640.0, + "step": 10024 + }, + { + "epoch": 1.1009224687019548, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.208580255508423, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.694139838218689, + "num_tokens": 259525249.0, + "step": 10025 + }, + { + "epoch": 1.1010322864045685, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.0131828784942627, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.6948599815368652, + "num_tokens": 259558002.0, + "step": 10026 + }, + { + "epoch": 1.101142104107182, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.068683624267578, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7260928153991699, + "num_tokens": 259587005.0, + "step": 10027 + }, + { + "epoch": 1.1012519218097958, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.4899559020996094, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7020885944366455, + "num_tokens": 259608555.0, + "step": 10028 + }, + { + "epoch": 1.1013617395124093, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.276625633239746, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7153394222259521, + "num_tokens": 259632527.0, + "step": 10029 + }, + { + "epoch": 1.101471557215023, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.201899528503418, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7379575371742249, + "num_tokens": 259657114.0, + "step": 10030 + }, + { + "epoch": 1.1015813749176366, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.0987548828125, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7181800603866577, + "num_tokens": 259687056.0, + "step": 10031 + }, + { + "epoch": 1.1016911926202504, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3022313117980957, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7301481366157532, + "num_tokens": 259711041.0, + "step": 10032 + }, + { + "epoch": 1.1018010103228641, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.235820770263672, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7082608938217163, + "num_tokens": 259735799.0, + "step": 10033 + }, + { + "epoch": 1.1019108280254777, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3918488025665283, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7204433679580688, + "num_tokens": 259759245.0, + "step": 10034 + }, + { + "epoch": 1.1020206457280914, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.232794761657715, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7095086574554443, + "num_tokens": 259786430.0, + "step": 10035 + }, + { + "epoch": 1.102130463430705, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3005425930023193, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7044090628623962, + "num_tokens": 259812205.0, + "step": 10036 + }, + { + "epoch": 1.1022402811333187, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.4701199531555176, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7201193571090698, + "num_tokens": 259834090.0, + "step": 10037 + }, + { + "epoch": 1.1023500988359323, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2662880420684814, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.719062328338623, + "num_tokens": 259859684.0, + "step": 10038 + }, + { + "epoch": 1.102459916538546, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.118523359298706, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7299045324325562, + "num_tokens": 259887694.0, + "step": 10039 + }, + { + "epoch": 1.1025697342411598, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.1768362522125244, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7069041728973389, + "num_tokens": 259915361.0, + "step": 10040 + }, + { + "epoch": 1.1026795519437733, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.389538049697876, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7092713117599487, + "num_tokens": 259939604.0, + "step": 10041 + }, + { + "epoch": 1.102789369646387, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.219576597213745, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7035375237464905, + "num_tokens": 259964570.0, + "step": 10042 + }, + { + "epoch": 1.1028991873490006, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3356363773345947, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7289577722549438, + "num_tokens": 259988233.0, + "step": 10043 + }, + { + "epoch": 1.1030090050516144, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.206251621246338, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7017935514450073, + "num_tokens": 260016806.0, + "step": 10044 + }, + { + "epoch": 1.103118822754228, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2069544792175293, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7097804546356201, + "num_tokens": 260044011.0, + "step": 10045 + }, + { + "epoch": 1.1032286404568417, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3768577575683594, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7229424715042114, + "num_tokens": 260068734.0, + "step": 10046 + }, + { + "epoch": 1.1033384581594552, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.204568386077881, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7256356477737427, + "num_tokens": 260093825.0, + "step": 10047 + }, + { + "epoch": 1.103448275862069, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.0761468410491943, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6927582025527954, + "num_tokens": 260122935.0, + "step": 10048 + }, + { + "epoch": 1.1035580935646827, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2803492546081543, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7322027683258057, + "num_tokens": 260148211.0, + "step": 10049 + }, + { + "epoch": 1.1036679112672962, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.4579827785491943, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7280205488204956, + "num_tokens": 260171463.0, + "step": 10050 + }, + { + "epoch": 1.10377772896991, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.316575527191162, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7341709136962891, + "num_tokens": 260194896.0, + "step": 10051 + }, + { + "epoch": 1.1038875466725235, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.4823687076568604, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7005334496498108, + "num_tokens": 260218383.0, + "step": 10052 + }, + { + "epoch": 1.1039973643751373, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.3010406494140625, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7079861164093018, + "num_tokens": 260244991.0, + "step": 10053 + }, + { + "epoch": 1.104107182077751, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2496254444122314, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7233878374099731, + "num_tokens": 260271284.0, + "step": 10054 + }, + { + "epoch": 1.1042169997803646, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.47442626953125, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.702325701713562, + "num_tokens": 260294880.0, + "step": 10055 + }, + { + "epoch": 1.1043268174829783, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.2397875785827637, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6869341135025024, + "num_tokens": 260322375.0, + "step": 10056 + }, + { + "epoch": 1.1044366351855919, + "ewc_loss": 9.250640869140625e-05, + "grad_norm": 2.1387016773223877, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6962478160858154, + "num_tokens": 260352339.0, + "step": 10057 + }, + { + "epoch": 1.1045464528882056, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.4284746646881104, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7148684859275818, + "num_tokens": 260376081.0, + "step": 10058 + }, + { + "epoch": 1.1046562705908192, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.3454489707946777, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7111830115318298, + "num_tokens": 260400360.0, + "step": 10059 + }, + { + "epoch": 1.104766088293433, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2177064418792725, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7197344303131104, + "num_tokens": 260427736.0, + "step": 10060 + }, + { + "epoch": 1.1048759059960465, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2921884059906006, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7115082144737244, + "num_tokens": 260451174.0, + "step": 10061 + }, + { + "epoch": 1.1049857236986602, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.46565842628479, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.714948832988739, + "num_tokens": 260474631.0, + "step": 10062 + }, + { + "epoch": 1.105095541401274, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.251490831375122, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7337374687194824, + "num_tokens": 260499111.0, + "step": 10063 + }, + { + "epoch": 1.1052053591038875, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.4503602981567383, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7274171113967896, + "num_tokens": 260520362.0, + "step": 10064 + }, + { + "epoch": 1.1053151768065013, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.1717617511749268, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7290056943893433, + "num_tokens": 260547511.0, + "step": 10065 + }, + { + "epoch": 1.1054249945091148, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.410454273223877, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7435846924781799, + "num_tokens": 260569000.0, + "step": 10066 + }, + { + "epoch": 1.1055348122117286, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 1.846204400062561, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7221750020980835, + "num_tokens": 260602601.0, + "step": 10067 + }, + { + "epoch": 1.105644629914342, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.1714770793914795, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.703249990940094, + "num_tokens": 260631014.0, + "step": 10068 + }, + { + "epoch": 1.1057544476169558, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2280845642089844, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7232582569122314, + "num_tokens": 260656564.0, + "step": 10069 + }, + { + "epoch": 1.1058642653195696, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.3259387016296387, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7254385948181152, + "num_tokens": 260678529.0, + "step": 10070 + }, + { + "epoch": 1.1059740830221831, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.149993419647217, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7299314141273499, + "num_tokens": 260703650.0, + "step": 10071 + }, + { + "epoch": 1.106083900724797, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.4141924381256104, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.719519853591919, + "num_tokens": 260725587.0, + "step": 10072 + }, + { + "epoch": 1.1061937184274104, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.292524576187134, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7379543781280518, + "num_tokens": 260749706.0, + "step": 10073 + }, + { + "epoch": 1.1063035361300242, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2094357013702393, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7368465065956116, + "num_tokens": 260775255.0, + "step": 10074 + }, + { + "epoch": 1.1064133538326377, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.1475486755371094, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7210255265235901, + "num_tokens": 260803395.0, + "step": 10075 + }, + { + "epoch": 1.1065231715352515, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.3567421436309814, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7171371579170227, + "num_tokens": 260827511.0, + "step": 10076 + }, + { + "epoch": 1.1066329892378652, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.0519986152648926, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7192768454551697, + "num_tokens": 260857517.0, + "step": 10077 + }, + { + "epoch": 1.1067428069404788, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.408526659011841, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7266963124275208, + "num_tokens": 260881444.0, + "step": 10078 + }, + { + "epoch": 1.1068526246430925, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.288210153579712, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7243934273719788, + "num_tokens": 260907718.0, + "step": 10079 + }, + { + "epoch": 1.106962442345706, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2150609493255615, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7141916751861572, + "num_tokens": 260934543.0, + "step": 10080 + }, + { + "epoch": 1.1070722600483198, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.320596933364868, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7074512243270874, + "num_tokens": 260960660.0, + "step": 10081 + }, + { + "epoch": 1.1071820777509334, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.1777899265289307, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7208439707756042, + "num_tokens": 260988653.0, + "step": 10082 + }, + { + "epoch": 1.1072918954535471, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.156311511993408, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7095747590065002, + "num_tokens": 261017363.0, + "step": 10083 + }, + { + "epoch": 1.1074017131561609, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2047150135040283, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.711047351360321, + "num_tokens": 261044320.0, + "step": 10084 + }, + { + "epoch": 1.1075115308587744, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.178722381591797, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7402349710464478, + "num_tokens": 261069403.0, + "step": 10085 + }, + { + "epoch": 1.1076213485613882, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.6764230728149414, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.727144718170166, + "num_tokens": 261088395.0, + "step": 10086 + }, + { + "epoch": 1.1077311662640017, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2561914920806885, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7344620227813721, + "num_tokens": 261113142.0, + "step": 10087 + }, + { + "epoch": 1.1078409839666155, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.4454843997955322, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6931154131889343, + "num_tokens": 261135076.0, + "step": 10088 + }, + { + "epoch": 1.107950801669229, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2047643661499023, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7107570767402649, + "num_tokens": 261159556.0, + "step": 10089 + }, + { + "epoch": 1.1080606193718427, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.4060451984405518, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7247090339660645, + "num_tokens": 261185178.0, + "step": 10090 + }, + { + "epoch": 1.1081704370744565, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.196404218673706, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7426642179489136, + "num_tokens": 261210986.0, + "step": 10091 + }, + { + "epoch": 1.10828025477707, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 1.996720790863037, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7025978565216064, + "num_tokens": 261242510.0, + "step": 10092 + }, + { + "epoch": 1.1083900724796838, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.211913585662842, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7187174558639526, + "num_tokens": 261267792.0, + "step": 10093 + }, + { + "epoch": 1.1084998901822973, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.291454792022705, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7274314165115356, + "num_tokens": 261293488.0, + "step": 10094 + }, + { + "epoch": 1.108609707884911, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.579270362854004, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.723693311214447, + "num_tokens": 261313720.0, + "step": 10095 + }, + { + "epoch": 1.1087195255875246, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.6938984394073486, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.738192081451416, + "num_tokens": 261331486.0, + "step": 10096 + }, + { + "epoch": 1.1088293432901384, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.447373390197754, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7186934351921082, + "num_tokens": 261353896.0, + "step": 10097 + }, + { + "epoch": 1.108939160992752, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.622929334640503, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7111504077911377, + "num_tokens": 261376889.0, + "step": 10098 + }, + { + "epoch": 1.1090489786953657, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.677595615386963, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7221829891204834, + "num_tokens": 261395160.0, + "step": 10099 + }, + { + "epoch": 1.1091587963979794, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.2588143348693848, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7228092551231384, + "num_tokens": 261421634.0, + "step": 10100 + }, + { + "epoch": 1.109268614100593, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.12180495262146, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7176167964935303, + "num_tokens": 261450939.0, + "step": 10101 + }, + { + "epoch": 1.1093784318032067, + "ewc_loss": 9.298324584960938e-05, + "grad_norm": 2.360651731491089, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7230859994888306, + "num_tokens": 261474246.0, + "step": 10102 + }, + { + "epoch": 1.1094882495058203, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.2040727138519287, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7394874095916748, + "num_tokens": 261499381.0, + "step": 10103 + }, + { + "epoch": 1.109598067208434, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.0921080112457275, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.724143922328949, + "num_tokens": 261529173.0, + "step": 10104 + }, + { + "epoch": 1.1097078849110478, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 1.9775304794311523, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7139583826065063, + "num_tokens": 261562491.0, + "step": 10105 + }, + { + "epoch": 1.1098177026136613, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.293264389038086, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7234718799591064, + "num_tokens": 261587022.0, + "step": 10106 + }, + { + "epoch": 1.109927520316275, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.2112772464752197, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.7095040082931519, + "num_tokens": 261613628.0, + "step": 10107 + }, + { + "epoch": 1.1100373380188886, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.344820499420166, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7266070246696472, + "num_tokens": 261634952.0, + "step": 10108 + }, + { + "epoch": 1.1101471557215024, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.32907772064209, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7280017137527466, + "num_tokens": 261657788.0, + "step": 10109 + }, + { + "epoch": 1.110256973424116, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.0069682598114014, + "learning_rate": 1e-06, + "loss": 1.0808, + "mean_token_accuracy": 0.6718311905860901, + "num_tokens": 261691464.0, + "step": 10110 + }, + { + "epoch": 1.1103667911267296, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.3682868480682373, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7323243618011475, + "num_tokens": 261713940.0, + "step": 10111 + }, + { + "epoch": 1.1104766088293432, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.2465391159057617, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7312869429588318, + "num_tokens": 261739358.0, + "step": 10112 + }, + { + "epoch": 1.110586426531957, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.245035409927368, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7332882881164551, + "num_tokens": 261765350.0, + "step": 10113 + }, + { + "epoch": 1.1106962442345707, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.370302677154541, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7184842824935913, + "num_tokens": 261790069.0, + "step": 10114 + }, + { + "epoch": 1.1108060619371842, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.199446439743042, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7229030132293701, + "num_tokens": 261815604.0, + "step": 10115 + }, + { + "epoch": 1.110915879639798, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.201991081237793, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.705091118812561, + "num_tokens": 261841623.0, + "step": 10116 + }, + { + "epoch": 1.1110256973424115, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.5661773681640625, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.6996776461601257, + "num_tokens": 261862501.0, + "step": 10117 + }, + { + "epoch": 1.1111355150450253, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.1621060371398926, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7008196115493774, + "num_tokens": 261890576.0, + "step": 10118 + }, + { + "epoch": 1.111245332747639, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.2522847652435303, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7344839572906494, + "num_tokens": 261915171.0, + "step": 10119 + }, + { + "epoch": 1.1113551504502526, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.1966257095336914, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7344846129417419, + "num_tokens": 261942007.0, + "step": 10120 + }, + { + "epoch": 1.1114649681528663, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.2530324459075928, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7082945108413696, + "num_tokens": 261966752.0, + "step": 10121 + }, + { + "epoch": 1.1115747858554799, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.4725379943847656, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7152454257011414, + "num_tokens": 261990830.0, + "step": 10122 + }, + { + "epoch": 1.1116846035580936, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.223484754562378, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7055212259292603, + "num_tokens": 262015820.0, + "step": 10123 + }, + { + "epoch": 1.1117944212607072, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.691413402557373, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7379403114318848, + "num_tokens": 262034653.0, + "step": 10124 + }, + { + "epoch": 1.111904238963321, + "ewc_loss": 9.34600830078125e-05, + "grad_norm": 2.152494430541992, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.686377763748169, + "num_tokens": 262065968.0, + "step": 10125 + }, + { + "epoch": 1.1120140566659344, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 1.8504606485366821, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7040522694587708, + "num_tokens": 262102326.0, + "step": 10126 + }, + { + "epoch": 1.1121238743685482, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.19931697845459, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7049503326416016, + "num_tokens": 262129269.0, + "step": 10127 + }, + { + "epoch": 1.112233692071162, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1438045501708984, + "learning_rate": 1e-06, + "loss": 1.046, + "mean_token_accuracy": 0.6808135509490967, + "num_tokens": 262159214.0, + "step": 10128 + }, + { + "epoch": 1.1123435097737755, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.087623357772827, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.700873613357544, + "num_tokens": 262187605.0, + "step": 10129 + }, + { + "epoch": 1.1124533274763893, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2483372688293457, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7011115550994873, + "num_tokens": 262215300.0, + "step": 10130 + }, + { + "epoch": 1.1125631451790028, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.341031312942505, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7240537405014038, + "num_tokens": 262237667.0, + "step": 10131 + }, + { + "epoch": 1.1126729628816165, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.4038987159729004, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.720798671245575, + "num_tokens": 262260847.0, + "step": 10132 + }, + { + "epoch": 1.11278278058423, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1360316276550293, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7354198098182678, + "num_tokens": 262288164.0, + "step": 10133 + }, + { + "epoch": 1.1128925982868438, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.125237226486206, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7182881236076355, + "num_tokens": 262315357.0, + "step": 10134 + }, + { + "epoch": 1.1130024159894576, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.172830104827881, + "learning_rate": 1e-06, + "loss": 0.778, + "mean_token_accuracy": 0.7513519525527954, + "num_tokens": 262340817.0, + "step": 10135 + }, + { + "epoch": 1.1131122336920711, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.0831263065338135, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7219676375389099, + "num_tokens": 262369774.0, + "step": 10136 + }, + { + "epoch": 1.1132220513946849, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.191335439682007, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7067739963531494, + "num_tokens": 262395648.0, + "step": 10137 + }, + { + "epoch": 1.1133318690972984, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1481070518493652, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7179997563362122, + "num_tokens": 262421124.0, + "step": 10138 + }, + { + "epoch": 1.1134416867999122, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1955623626708984, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7381467819213867, + "num_tokens": 262446550.0, + "step": 10139 + }, + { + "epoch": 1.1135515045025257, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1837267875671387, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7346906661987305, + "num_tokens": 262471616.0, + "step": 10140 + }, + { + "epoch": 1.1136613222051395, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.8086936473846436, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7218143939971924, + "num_tokens": 262489414.0, + "step": 10141 + }, + { + "epoch": 1.1137711399077532, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 1.8099489212036133, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7032707929611206, + "num_tokens": 262526085.0, + "step": 10142 + }, + { + "epoch": 1.1138809576103668, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2633719444274902, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7167428731918335, + "num_tokens": 262550686.0, + "step": 10143 + }, + { + "epoch": 1.1139907753129805, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2543275356292725, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7026166319847107, + "num_tokens": 262575657.0, + "step": 10144 + }, + { + "epoch": 1.114100593015594, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.260190963745117, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6897463798522949, + "num_tokens": 262603127.0, + "step": 10145 + }, + { + "epoch": 1.1142104107182078, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2901723384857178, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.6991980075836182, + "num_tokens": 262628529.0, + "step": 10146 + }, + { + "epoch": 1.1143202284208213, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.166858196258545, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.685385525226593, + "num_tokens": 262656454.0, + "step": 10147 + }, + { + "epoch": 1.114430046123435, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.100724220275879, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7374469041824341, + "num_tokens": 262685153.0, + "step": 10148 + }, + { + "epoch": 1.1145398638260489, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1317241191864014, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7026447653770447, + "num_tokens": 262713813.0, + "step": 10149 + }, + { + "epoch": 1.1146496815286624, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 1.9927061796188354, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7259607315063477, + "num_tokens": 262743376.0, + "step": 10150 + }, + { + "epoch": 1.1147594992312762, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1541531085968018, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7393383383750916, + "num_tokens": 262770450.0, + "step": 10151 + }, + { + "epoch": 1.1148693169338897, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.275879383087158, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6916061639785767, + "num_tokens": 262795402.0, + "step": 10152 + }, + { + "epoch": 1.1149791346365034, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.4653139114379883, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6999704837799072, + "num_tokens": 262818722.0, + "step": 10153 + }, + { + "epoch": 1.115088952339117, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.3013079166412354, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7395950555801392, + "num_tokens": 262841738.0, + "step": 10154 + }, + { + "epoch": 1.1151987700417307, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2310471534729004, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7439083456993103, + "num_tokens": 262865964.0, + "step": 10155 + }, + { + "epoch": 1.1153085877443445, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.7571117877960205, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7030643820762634, + "num_tokens": 262885839.0, + "step": 10156 + }, + { + "epoch": 1.115418405446958, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.3942387104034424, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7237100601196289, + "num_tokens": 262908507.0, + "step": 10157 + }, + { + "epoch": 1.1155282231495718, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.467256546020508, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7214246392250061, + "num_tokens": 262931525.0, + "step": 10158 + }, + { + "epoch": 1.1156380408521853, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.4908576011657715, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7080684900283813, + "num_tokens": 262954414.0, + "step": 10159 + }, + { + "epoch": 1.115747858554799, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.455037832260132, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7066378593444824, + "num_tokens": 262976802.0, + "step": 10160 + }, + { + "epoch": 1.1158576762574126, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1043004989624023, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.71918123960495, + "num_tokens": 263006436.0, + "step": 10161 + }, + { + "epoch": 1.1159674939600264, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.173600196838379, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7020660638809204, + "num_tokens": 263033693.0, + "step": 10162 + }, + { + "epoch": 1.11607731166264, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.507021903991699, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7192316651344299, + "num_tokens": 263055731.0, + "step": 10163 + }, + { + "epoch": 1.1161871293652537, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.275677442550659, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7099582552909851, + "num_tokens": 263081564.0, + "step": 10164 + }, + { + "epoch": 1.1162969470678674, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1759161949157715, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7215359210968018, + "num_tokens": 263110244.0, + "step": 10165 + }, + { + "epoch": 1.116406764770481, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.311274528503418, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7037675976753235, + "num_tokens": 263135014.0, + "step": 10166 + }, + { + "epoch": 1.1165165824730947, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.5225942134857178, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7014265060424805, + "num_tokens": 263156727.0, + "step": 10167 + }, + { + "epoch": 1.1166264001757082, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2094430923461914, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7226401567459106, + "num_tokens": 263183522.0, + "step": 10168 + }, + { + "epoch": 1.116736217878322, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2911221981048584, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.734445333480835, + "num_tokens": 263208367.0, + "step": 10169 + }, + { + "epoch": 1.1168460355809358, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.579442262649536, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7522342205047607, + "num_tokens": 263227569.0, + "step": 10170 + }, + { + "epoch": 1.1169558532835493, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.110163688659668, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.722374439239502, + "num_tokens": 263255747.0, + "step": 10171 + }, + { + "epoch": 1.117065670986163, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.4860706329345703, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7221990823745728, + "num_tokens": 263277067.0, + "step": 10172 + }, + { + "epoch": 1.1171754886887766, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.3190088272094727, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7254340052604675, + "num_tokens": 263300135.0, + "step": 10173 + }, + { + "epoch": 1.1172853063913903, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.7506184577941895, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7339218854904175, + "num_tokens": 263318790.0, + "step": 10174 + }, + { + "epoch": 1.1173951240940039, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.158989191055298, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7142940759658813, + "num_tokens": 263348357.0, + "step": 10175 + }, + { + "epoch": 1.1175049417966176, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.263864278793335, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7350075244903564, + "num_tokens": 263372010.0, + "step": 10176 + }, + { + "epoch": 1.1176147594992312, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.2438995838165283, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7132361531257629, + "num_tokens": 263397378.0, + "step": 10177 + }, + { + "epoch": 1.117724577201845, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.1841237545013428, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.721747100353241, + "num_tokens": 263422840.0, + "step": 10178 + }, + { + "epoch": 1.1178343949044587, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.29988431930542, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.722813606262207, + "num_tokens": 263447140.0, + "step": 10179 + }, + { + "epoch": 1.1179442126070722, + "ewc_loss": 9.441375732421875e-05, + "grad_norm": 2.0689456462860107, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.703200101852417, + "num_tokens": 263477950.0, + "step": 10180 + }, + { + "epoch": 1.118054030309686, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 1.9575920104980469, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6979864239692688, + "num_tokens": 263511333.0, + "step": 10181 + }, + { + "epoch": 1.1181638480122995, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0836586952209473, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7116159200668335, + "num_tokens": 263539008.0, + "step": 10182 + }, + { + "epoch": 1.1182736657149133, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.473424196243286, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7282537221908569, + "num_tokens": 263559854.0, + "step": 10183 + }, + { + "epoch": 1.1183834834175268, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3357086181640625, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7213610410690308, + "num_tokens": 263583069.0, + "step": 10184 + }, + { + "epoch": 1.1184933011201406, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4966869354248047, + "learning_rate": 1e-06, + "loss": 0.7906, + "mean_token_accuracy": 0.754694938659668, + "num_tokens": 263603230.0, + "step": 10185 + }, + { + "epoch": 1.1186031188227543, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0602126121520996, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7102816104888916, + "num_tokens": 263632745.0, + "step": 10186 + }, + { + "epoch": 1.1187129365253679, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3667709827423096, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.708500862121582, + "num_tokens": 263656588.0, + "step": 10187 + }, + { + "epoch": 1.1188227542279816, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4577507972717285, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7287452220916748, + "num_tokens": 263677098.0, + "step": 10188 + }, + { + "epoch": 1.1189325719305951, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3480546474456787, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7042157053947449, + "num_tokens": 263700576.0, + "step": 10189 + }, + { + "epoch": 1.119042389633209, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.053114652633667, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7137663960456848, + "num_tokens": 263728753.0, + "step": 10190 + }, + { + "epoch": 1.1191522073358224, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0725257396698, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7132824659347534, + "num_tokens": 263754501.0, + "step": 10191 + }, + { + "epoch": 1.1192620250384362, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.151517152786255, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7070909738540649, + "num_tokens": 263781638.0, + "step": 10192 + }, + { + "epoch": 1.11937184274105, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3281803131103516, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7038613557815552, + "num_tokens": 263806214.0, + "step": 10193 + }, + { + "epoch": 1.1194816604436635, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.124532699584961, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7109583616256714, + "num_tokens": 263834287.0, + "step": 10194 + }, + { + "epoch": 1.1195914781462772, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.21439528465271, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.715225100517273, + "num_tokens": 263861180.0, + "step": 10195 + }, + { + "epoch": 1.1197012958488908, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4449543952941895, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.730661153793335, + "num_tokens": 263883046.0, + "step": 10196 + }, + { + "epoch": 1.1198111135515045, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1691999435424805, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7247500419616699, + "num_tokens": 263910091.0, + "step": 10197 + }, + { + "epoch": 1.119920931254118, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0999655723571777, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7276358604431152, + "num_tokens": 263937454.0, + "step": 10198 + }, + { + "epoch": 1.1200307489567318, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2040722370147705, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7207726240158081, + "num_tokens": 263963185.0, + "step": 10199 + }, + { + "epoch": 1.1201405666593456, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4870553016662598, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7287050485610962, + "num_tokens": 263985334.0, + "step": 10200 + }, + { + "epoch": 1.1202503843619591, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1769261360168457, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.7008936405181885, + "num_tokens": 264013050.0, + "step": 10201 + }, + { + "epoch": 1.1203602020645729, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.259010076522827, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7053802013397217, + "num_tokens": 264040065.0, + "step": 10202 + }, + { + "epoch": 1.1204700197671864, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.137502431869507, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7048373222351074, + "num_tokens": 264067824.0, + "step": 10203 + }, + { + "epoch": 1.1205798374698002, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1879794597625732, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7020834684371948, + "num_tokens": 264094078.0, + "step": 10204 + }, + { + "epoch": 1.1206896551724137, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0642478466033936, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7069739103317261, + "num_tokens": 264122206.0, + "step": 10205 + }, + { + "epoch": 1.1207994728750275, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3296358585357666, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7036955952644348, + "num_tokens": 264146490.0, + "step": 10206 + }, + { + "epoch": 1.1209092905776412, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.9212794303894043, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.743913471698761, + "num_tokens": 264162769.0, + "step": 10207 + }, + { + "epoch": 1.1210191082802548, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3298189640045166, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7232224941253662, + "num_tokens": 264185118.0, + "step": 10208 + }, + { + "epoch": 1.1211289259828685, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0372161865234375, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7119184136390686, + "num_tokens": 264217125.0, + "step": 10209 + }, + { + "epoch": 1.121238743685482, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.155064105987549, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7306677103042603, + "num_tokens": 264243982.0, + "step": 10210 + }, + { + "epoch": 1.1213485613880958, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3640451431274414, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7043187618255615, + "num_tokens": 264268132.0, + "step": 10211 + }, + { + "epoch": 1.1214583790907093, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.347292423248291, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7216536998748779, + "num_tokens": 264290913.0, + "step": 10212 + }, + { + "epoch": 1.121568196793323, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.03945255279541, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7334476709365845, + "num_tokens": 264317797.0, + "step": 10213 + }, + { + "epoch": 1.1216780144959368, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3423714637756348, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7129409909248352, + "num_tokens": 264343062.0, + "step": 10214 + }, + { + "epoch": 1.1217878321985504, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3621089458465576, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7131288051605225, + "num_tokens": 264365747.0, + "step": 10215 + }, + { + "epoch": 1.1218976499011641, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3280158042907715, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7072007656097412, + "num_tokens": 264389155.0, + "step": 10216 + }, + { + "epoch": 1.1220074676037777, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2049763202667236, + "learning_rate": 1e-06, + "loss": 0.7655, + "mean_token_accuracy": 0.7630266547203064, + "num_tokens": 264412724.0, + "step": 10217 + }, + { + "epoch": 1.1221172853063914, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3005006313323975, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7041438221931458, + "num_tokens": 264437737.0, + "step": 10218 + }, + { + "epoch": 1.122227103009005, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3771231174468994, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.6998367309570312, + "num_tokens": 264464141.0, + "step": 10219 + }, + { + "epoch": 1.1223369207116187, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2801225185394287, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.710186243057251, + "num_tokens": 264489951.0, + "step": 10220 + }, + { + "epoch": 1.1224467384142325, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.027907371520996, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7025021910667419, + "num_tokens": 264522420.0, + "step": 10221 + }, + { + "epoch": 1.122556556116846, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1594502925872803, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7106121182441711, + "num_tokens": 264550110.0, + "step": 10222 + }, + { + "epoch": 1.1226663738194598, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.170297861099243, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7037056684494019, + "num_tokens": 264577054.0, + "step": 10223 + }, + { + "epoch": 1.1227761915220733, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.290335178375244, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7138707637786865, + "num_tokens": 264602938.0, + "step": 10224 + }, + { + "epoch": 1.122886009224687, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.233203172683716, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7175805568695068, + "num_tokens": 264629059.0, + "step": 10225 + }, + { + "epoch": 1.1229958269273006, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1501247882843018, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7285515069961548, + "num_tokens": 264655928.0, + "step": 10226 + }, + { + "epoch": 1.1231056446299144, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2176170349121094, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7280720472335815, + "num_tokens": 264681093.0, + "step": 10227 + }, + { + "epoch": 1.123215462332528, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.252117395401001, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7289165258407593, + "num_tokens": 264705936.0, + "step": 10228 + }, + { + "epoch": 1.1233252800351416, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2087130546569824, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7132659554481506, + "num_tokens": 264731489.0, + "step": 10229 + }, + { + "epoch": 1.1234350977377554, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3461403846740723, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7381842136383057, + "num_tokens": 264753490.0, + "step": 10230 + }, + { + "epoch": 1.123544915440369, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.312232255935669, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7137319445610046, + "num_tokens": 264779633.0, + "step": 10231 + }, + { + "epoch": 1.1236547331429827, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2429661750793457, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.705634593963623, + "num_tokens": 264804496.0, + "step": 10232 + }, + { + "epoch": 1.1237645508455962, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.213675022125244, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6911652088165283, + "num_tokens": 264832883.0, + "step": 10233 + }, + { + "epoch": 1.12387436854821, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.238130807876587, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7308961749076843, + "num_tokens": 264858624.0, + "step": 10234 + }, + { + "epoch": 1.1239841862508237, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.594736337661743, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7125246524810791, + "num_tokens": 264880895.0, + "step": 10235 + }, + { + "epoch": 1.1240940039534373, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.268638849258423, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7002087235450745, + "num_tokens": 264907135.0, + "step": 10236 + }, + { + "epoch": 1.124203821656051, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2084481716156006, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7119860649108887, + "num_tokens": 264934646.0, + "step": 10237 + }, + { + "epoch": 1.1243136393586646, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.0270040035247803, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7201427221298218, + "num_tokens": 264963994.0, + "step": 10238 + }, + { + "epoch": 1.1244234570612783, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3482203483581543, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7238994836807251, + "num_tokens": 264988809.0, + "step": 10239 + }, + { + "epoch": 1.1245332747638919, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3018832206726074, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7271542549133301, + "num_tokens": 265011918.0, + "step": 10240 + }, + { + "epoch": 1.1246430924665056, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1047840118408203, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7138730883598328, + "num_tokens": 265040220.0, + "step": 10241 + }, + { + "epoch": 1.1247529101691192, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1814916133880615, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7335784435272217, + "num_tokens": 265066840.0, + "step": 10242 + }, + { + "epoch": 1.124862727871733, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.424903631210327, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7296102046966553, + "num_tokens": 265090591.0, + "step": 10243 + }, + { + "epoch": 1.1249725455743467, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2321128845214844, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7083607912063599, + "num_tokens": 265117912.0, + "step": 10244 + }, + { + "epoch": 1.1250823632769602, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2933154106140137, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7134031653404236, + "num_tokens": 265143335.0, + "step": 10245 + }, + { + "epoch": 1.125192180979574, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.342334032058716, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6939558982849121, + "num_tokens": 265168875.0, + "step": 10246 + }, + { + "epoch": 1.1253019986821875, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2295212745666504, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6967281103134155, + "num_tokens": 265195136.0, + "step": 10247 + }, + { + "epoch": 1.1254118163848013, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1038291454315186, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7114410996437073, + "num_tokens": 265222875.0, + "step": 10248 + }, + { + "epoch": 1.125521634087415, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.5923666954040527, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7221311926841736, + "num_tokens": 265242173.0, + "step": 10249 + }, + { + "epoch": 1.1256314517900285, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3346009254455566, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.737290620803833, + "num_tokens": 265265667.0, + "step": 10250 + }, + { + "epoch": 1.1257412694926423, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.229139804840088, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6949777603149414, + "num_tokens": 265295514.0, + "step": 10251 + }, + { + "epoch": 1.1258510871952558, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4422848224639893, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7393458485603333, + "num_tokens": 265317261.0, + "step": 10252 + }, + { + "epoch": 1.1259609048978696, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.440290689468384, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7026721835136414, + "num_tokens": 265342770.0, + "step": 10253 + }, + { + "epoch": 1.1260707226004831, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.5634758472442627, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7185354828834534, + "num_tokens": 265363346.0, + "step": 10254 + }, + { + "epoch": 1.1261805403030969, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1506335735321045, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.7141157388687134, + "num_tokens": 265391417.0, + "step": 10255 + }, + { + "epoch": 1.1262903580057104, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.549137830734253, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7434446811676025, + "num_tokens": 265412658.0, + "step": 10256 + }, + { + "epoch": 1.1264001757083242, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.138892650604248, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7229622006416321, + "num_tokens": 265439582.0, + "step": 10257 + }, + { + "epoch": 1.126509993410938, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2265541553497314, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7165107727050781, + "num_tokens": 265466946.0, + "step": 10258 + }, + { + "epoch": 1.1266198111135515, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.035548686981201, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7133224010467529, + "num_tokens": 265499854.0, + "step": 10259 + }, + { + "epoch": 1.1267296288161652, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.444634437561035, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7299020290374756, + "num_tokens": 265523938.0, + "step": 10260 + }, + { + "epoch": 1.1268394465187788, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 1.842026710510254, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7182756066322327, + "num_tokens": 265560468.0, + "step": 10261 + }, + { + "epoch": 1.1269492642213925, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.418898105621338, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7084833383560181, + "num_tokens": 265583667.0, + "step": 10262 + }, + { + "epoch": 1.127059081924006, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.270838975906372, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7391089797019958, + "num_tokens": 265606595.0, + "step": 10263 + }, + { + "epoch": 1.1271688996266198, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.199786901473999, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7099913358688354, + "num_tokens": 265633660.0, + "step": 10264 + }, + { + "epoch": 1.1272787173292333, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1913628578186035, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7098064422607422, + "num_tokens": 265659487.0, + "step": 10265 + }, + { + "epoch": 1.127388535031847, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1522159576416016, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.708877682685852, + "num_tokens": 265688273.0, + "step": 10266 + }, + { + "epoch": 1.1274983527344609, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2441065311431885, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.698337197303772, + "num_tokens": 265713972.0, + "step": 10267 + }, + { + "epoch": 1.1276081704370744, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.330726146697998, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6970522403717041, + "num_tokens": 265739755.0, + "step": 10268 + }, + { + "epoch": 1.1277179881396882, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4248361587524414, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6967117786407471, + "num_tokens": 265764051.0, + "step": 10269 + }, + { + "epoch": 1.1278278058423017, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.260620594024658, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7125176787376404, + "num_tokens": 265789653.0, + "step": 10270 + }, + { + "epoch": 1.1279376235449154, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1155242919921875, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7292634844779968, + "num_tokens": 265816462.0, + "step": 10271 + }, + { + "epoch": 1.1280474412475292, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.415210008621216, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7365669012069702, + "num_tokens": 265840306.0, + "step": 10272 + }, + { + "epoch": 1.1281572589501427, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.003845453262329, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7293221950531006, + "num_tokens": 265870347.0, + "step": 10273 + }, + { + "epoch": 1.1282670766527565, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.5089614391326904, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7069703936576843, + "num_tokens": 265894363.0, + "step": 10274 + }, + { + "epoch": 1.12837689435537, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.2645552158355713, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7160190343856812, + "num_tokens": 265921549.0, + "step": 10275 + }, + { + "epoch": 1.1284867120579838, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.1721928119659424, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7236818671226501, + "num_tokens": 265948626.0, + "step": 10276 + }, + { + "epoch": 1.1285965297605973, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 1.9565322399139404, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7082042694091797, + "num_tokens": 265983453.0, + "step": 10277 + }, + { + "epoch": 1.128706347463211, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.4427456855773926, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7309637069702148, + "num_tokens": 266005444.0, + "step": 10278 + }, + { + "epoch": 1.1288161651658246, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.183431386947632, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7369604110717773, + "num_tokens": 266031241.0, + "step": 10279 + }, + { + "epoch": 1.1289259828684384, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.512406587600708, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7271991968154907, + "num_tokens": 266053070.0, + "step": 10280 + }, + { + "epoch": 1.1290358005710521, + "ewc_loss": 9.489059448242188e-05, + "grad_norm": 2.3337514400482178, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7037018537521362, + "num_tokens": 266076940.0, + "step": 10281 + }, + { + "epoch": 1.1291456182736657, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 1.9634817838668823, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.704576313495636, + "num_tokens": 266111542.0, + "step": 10282 + }, + { + "epoch": 1.1292554359762794, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.077341318130493, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7477084398269653, + "num_tokens": 266137534.0, + "step": 10283 + }, + { + "epoch": 1.129365253678893, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.257387161254883, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7332801818847656, + "num_tokens": 266161658.0, + "step": 10284 + }, + { + "epoch": 1.1294750713815067, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.393453598022461, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7169308662414551, + "num_tokens": 266184514.0, + "step": 10285 + }, + { + "epoch": 1.1295848890841205, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.1943752765655518, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6930109858512878, + "num_tokens": 266214113.0, + "step": 10286 + }, + { + "epoch": 1.129694706786734, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.465054750442505, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7282571792602539, + "num_tokens": 266234590.0, + "step": 10287 + }, + { + "epoch": 1.1298045244893478, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.094308614730835, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7364436388015747, + "num_tokens": 266262165.0, + "step": 10288 + }, + { + "epoch": 1.1299143421919613, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.2230546474456787, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7275328636169434, + "num_tokens": 266287184.0, + "step": 10289 + }, + { + "epoch": 1.130024159894575, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.756138324737549, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.728324294090271, + "num_tokens": 266305758.0, + "step": 10290 + }, + { + "epoch": 1.1301339775971886, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.110710859298706, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7217915654182434, + "num_tokens": 266335294.0, + "step": 10291 + }, + { + "epoch": 1.1302437952998023, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.0939598083496094, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.7123129963874817, + "num_tokens": 266363544.0, + "step": 10292 + }, + { + "epoch": 1.1303536130024159, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.3273072242736816, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6898622512817383, + "num_tokens": 266389150.0, + "step": 10293 + }, + { + "epoch": 1.1304634307050296, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.011017322540283, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6967085599899292, + "num_tokens": 266418930.0, + "step": 10294 + }, + { + "epoch": 1.1305732484076434, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.2671597003936768, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6970749497413635, + "num_tokens": 266445519.0, + "step": 10295 + }, + { + "epoch": 1.130683066110257, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.6495964527130127, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7228490114212036, + "num_tokens": 266465326.0, + "step": 10296 + }, + { + "epoch": 1.1307928838128707, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.03369402885437, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7101143002510071, + "num_tokens": 266498034.0, + "step": 10297 + }, + { + "epoch": 1.1309027015154842, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.3749451637268066, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7093279361724854, + "num_tokens": 266523244.0, + "step": 10298 + }, + { + "epoch": 1.131012519218098, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.0068302154541016, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.6982467174530029, + "num_tokens": 266556348.0, + "step": 10299 + }, + { + "epoch": 1.1311223369207117, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.1735050678253174, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7113283276557922, + "num_tokens": 266585163.0, + "step": 10300 + }, + { + "epoch": 1.1312321546233253, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.3742880821228027, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7359889149665833, + "num_tokens": 266608597.0, + "step": 10301 + }, + { + "epoch": 1.131341972325939, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.4253647327423096, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7258468866348267, + "num_tokens": 266631533.0, + "step": 10302 + }, + { + "epoch": 1.1314517900285526, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.2056870460510254, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.6866398453712463, + "num_tokens": 266662321.0, + "step": 10303 + }, + { + "epoch": 1.1315616077311663, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.764223575592041, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7185942530632019, + "num_tokens": 266681531.0, + "step": 10304 + }, + { + "epoch": 1.1316714254337799, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.1053617000579834, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.729674220085144, + "num_tokens": 266709453.0, + "step": 10305 + }, + { + "epoch": 1.1317812431363936, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.213343620300293, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7318505644798279, + "num_tokens": 266735508.0, + "step": 10306 + }, + { + "epoch": 1.1318910608390071, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.467620611190796, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7047617435455322, + "num_tokens": 266758362.0, + "step": 10307 + }, + { + "epoch": 1.132000878541621, + "ewc_loss": 9.5367431640625e-05, + "grad_norm": 2.161625623703003, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7080692648887634, + "num_tokens": 266784633.0, + "step": 10308 + }, + { + "epoch": 1.1321106962442347, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2210910320281982, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7151282429695129, + "num_tokens": 266811149.0, + "step": 10309 + }, + { + "epoch": 1.1322205139468482, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.765409231185913, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7324388027191162, + "num_tokens": 266829030.0, + "step": 10310 + }, + { + "epoch": 1.132330331649462, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.0690672397613525, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7227572202682495, + "num_tokens": 266859845.0, + "step": 10311 + }, + { + "epoch": 1.1324401493520755, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2300264835357666, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7163023352622986, + "num_tokens": 266884529.0, + "step": 10312 + }, + { + "epoch": 1.1325499670546892, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2455902099609375, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.704702615737915, + "num_tokens": 266910291.0, + "step": 10313 + }, + { + "epoch": 1.132659784757303, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.170077085494995, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.740504264831543, + "num_tokens": 266937059.0, + "step": 10314 + }, + { + "epoch": 1.1327696024599165, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2492568492889404, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6996193528175354, + "num_tokens": 266964277.0, + "step": 10315 + }, + { + "epoch": 1.1328794201625303, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.002572774887085, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7073464393615723, + "num_tokens": 266996463.0, + "step": 10316 + }, + { + "epoch": 1.1329892378651438, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4669508934020996, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7155117988586426, + "num_tokens": 267019702.0, + "step": 10317 + }, + { + "epoch": 1.1330990555677576, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4444704055786133, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7304702997207642, + "num_tokens": 267040614.0, + "step": 10318 + }, + { + "epoch": 1.1332088732703711, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.249410629272461, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7210229635238647, + "num_tokens": 267067908.0, + "step": 10319 + }, + { + "epoch": 1.1333186909729849, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2411115169525146, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7137629985809326, + "num_tokens": 267093389.0, + "step": 10320 + }, + { + "epoch": 1.1334285086755984, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.023670196533203, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7114056944847107, + "num_tokens": 267123727.0, + "step": 10321 + }, + { + "epoch": 1.1335383263782122, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.013639450073242, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7045316696166992, + "num_tokens": 267153359.0, + "step": 10322 + }, + { + "epoch": 1.133648144080826, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.1413936614990234, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7194757461547852, + "num_tokens": 267180260.0, + "step": 10323 + }, + { + "epoch": 1.1337579617834395, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.0200862884521484, + "learning_rate": 1e-06, + "loss": 1.0355, + "mean_token_accuracy": 0.6871795654296875, + "num_tokens": 267214943.0, + "step": 10324 + }, + { + "epoch": 1.1338677794860532, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.382277727127075, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.712694525718689, + "num_tokens": 267238721.0, + "step": 10325 + }, + { + "epoch": 1.1339775971886668, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.067469835281372, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7191879749298096, + "num_tokens": 267269105.0, + "step": 10326 + }, + { + "epoch": 1.1340874148912805, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.251685380935669, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7095849514007568, + "num_tokens": 267295469.0, + "step": 10327 + }, + { + "epoch": 1.134197232593894, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.0889651775360107, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7054600715637207, + "num_tokens": 267323179.0, + "step": 10328 + }, + { + "epoch": 1.1343070502965078, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.529378652572632, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7199932932853699, + "num_tokens": 267347769.0, + "step": 10329 + }, + { + "epoch": 1.1344168679991213, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.241503953933716, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7153559923171997, + "num_tokens": 267373815.0, + "step": 10330 + }, + { + "epoch": 1.134526685701735, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.3732473850250244, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7184366583824158, + "num_tokens": 267396441.0, + "step": 10331 + }, + { + "epoch": 1.1346365034043489, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2375736236572266, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7191684246063232, + "num_tokens": 267422653.0, + "step": 10332 + }, + { + "epoch": 1.1347463211069624, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4691154956817627, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.726475179195404, + "num_tokens": 267445598.0, + "step": 10333 + }, + { + "epoch": 1.1348561388095761, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.1995856761932373, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7299187183380127, + "num_tokens": 267471406.0, + "step": 10334 + }, + { + "epoch": 1.1349659565121897, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4167160987854004, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7167788743972778, + "num_tokens": 267494400.0, + "step": 10335 + }, + { + "epoch": 1.1350757742148034, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.1261935234069824, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7294206023216248, + "num_tokens": 267524579.0, + "step": 10336 + }, + { + "epoch": 1.1351855919174172, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 1.9416354894638062, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7350505590438843, + "num_tokens": 267555767.0, + "step": 10337 + }, + { + "epoch": 1.1352954096200307, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 1.9088261127471924, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7128710150718689, + "num_tokens": 267589006.0, + "step": 10338 + }, + { + "epoch": 1.1354052273226445, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.3702504634857178, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7235180139541626, + "num_tokens": 267611517.0, + "step": 10339 + }, + { + "epoch": 1.135515045025258, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.0258138179779053, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6986203193664551, + "num_tokens": 267644383.0, + "step": 10340 + }, + { + "epoch": 1.1356248627278718, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.3330066204071045, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.703603982925415, + "num_tokens": 267667731.0, + "step": 10341 + }, + { + "epoch": 1.1357346804304853, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.388465642929077, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7107435464859009, + "num_tokens": 267692049.0, + "step": 10342 + }, + { + "epoch": 1.135844498133099, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.1759042739868164, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.726629912853241, + "num_tokens": 267719834.0, + "step": 10343 + }, + { + "epoch": 1.1359543158357126, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.644054889678955, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7341256737709045, + "num_tokens": 267741202.0, + "step": 10344 + }, + { + "epoch": 1.1360641335383264, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.081153392791748, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.6976242661476135, + "num_tokens": 267771467.0, + "step": 10345 + }, + { + "epoch": 1.1361739512409401, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4575281143188477, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7299675941467285, + "num_tokens": 267792675.0, + "step": 10346 + }, + { + "epoch": 1.1362837689435537, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.112750768661499, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7038785219192505, + "num_tokens": 267821926.0, + "step": 10347 + }, + { + "epoch": 1.1363935866461674, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.081782102584839, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7119581699371338, + "num_tokens": 267852223.0, + "step": 10348 + }, + { + "epoch": 1.136503404348781, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2474796772003174, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7134089469909668, + "num_tokens": 267876621.0, + "step": 10349 + }, + { + "epoch": 1.1366132220513947, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 1.9331704378128052, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7139897346496582, + "num_tokens": 267908585.0, + "step": 10350 + }, + { + "epoch": 1.1367230397540085, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.0578184127807617, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7248501777648926, + "num_tokens": 267940588.0, + "step": 10351 + }, + { + "epoch": 1.136832857456622, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.3759615421295166, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7197829484939575, + "num_tokens": 267965779.0, + "step": 10352 + }, + { + "epoch": 1.1369426751592357, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2666163444519043, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6948460340499878, + "num_tokens": 267992571.0, + "step": 10353 + }, + { + "epoch": 1.1370524928618493, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.23100209236145, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.712212085723877, + "num_tokens": 268019872.0, + "step": 10354 + }, + { + "epoch": 1.137162310564463, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.35006046295166, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7234413623809814, + "num_tokens": 268043544.0, + "step": 10355 + }, + { + "epoch": 1.1372721282670766, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.183767557144165, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7136985063552856, + "num_tokens": 268069948.0, + "step": 10356 + }, + { + "epoch": 1.1373819459696903, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.606797695159912, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7130454778671265, + "num_tokens": 268089960.0, + "step": 10357 + }, + { + "epoch": 1.1374917636723039, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.434825897216797, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7318199872970581, + "num_tokens": 268111988.0, + "step": 10358 + }, + { + "epoch": 1.1376015813749176, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.227437973022461, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7300472259521484, + "num_tokens": 268137247.0, + "step": 10359 + }, + { + "epoch": 1.1377113990775314, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.284203052520752, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7467645406723022, + "num_tokens": 268161393.0, + "step": 10360 + }, + { + "epoch": 1.137821216780145, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.1328630447387695, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7094732522964478, + "num_tokens": 268190923.0, + "step": 10361 + }, + { + "epoch": 1.1379310344827587, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4309451580047607, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.703085720539093, + "num_tokens": 268214444.0, + "step": 10362 + }, + { + "epoch": 1.1380408521853722, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.4397523403167725, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7188197374343872, + "num_tokens": 268236428.0, + "step": 10363 + }, + { + "epoch": 1.138150669887986, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.504255771636963, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7445945739746094, + "num_tokens": 268258604.0, + "step": 10364 + }, + { + "epoch": 1.1382604875905997, + "ewc_loss": 9.632110595703125e-05, + "grad_norm": 2.2780046463012695, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7526258230209351, + "num_tokens": 268285344.0, + "step": 10365 + }, + { + "epoch": 1.1383703052932133, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.9185491800308228, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7063318490982056, + "num_tokens": 268320968.0, + "step": 10366 + }, + { + "epoch": 1.138480122995827, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.321174144744873, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7303400039672852, + "num_tokens": 268347096.0, + "step": 10367 + }, + { + "epoch": 1.1385899406984406, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2777113914489746, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7255241870880127, + "num_tokens": 268373691.0, + "step": 10368 + }, + { + "epoch": 1.1386997584010543, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2794296741485596, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7047076225280762, + "num_tokens": 268398350.0, + "step": 10369 + }, + { + "epoch": 1.1388095761036678, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3759942054748535, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.72609543800354, + "num_tokens": 268422393.0, + "step": 10370 + }, + { + "epoch": 1.1389193938062816, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.210824966430664, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7432315349578857, + "num_tokens": 268447765.0, + "step": 10371 + }, + { + "epoch": 1.1390292115088951, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 3.1053390502929688, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7240496873855591, + "num_tokens": 268476834.0, + "step": 10372 + }, + { + "epoch": 1.139139029211509, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3244996070861816, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7237423658370972, + "num_tokens": 268501276.0, + "step": 10373 + }, + { + "epoch": 1.1392488469141226, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.9709360599517822, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7234801054000854, + "num_tokens": 268533575.0, + "step": 10374 + }, + { + "epoch": 1.1393586646167362, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.528959274291992, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7370492815971375, + "num_tokens": 268554506.0, + "step": 10375 + }, + { + "epoch": 1.13946848231935, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2755677700042725, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7217717170715332, + "num_tokens": 268579127.0, + "step": 10376 + }, + { + "epoch": 1.1395783000219635, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4240217208862305, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7107564210891724, + "num_tokens": 268602082.0, + "step": 10377 + }, + { + "epoch": 1.1396881177245772, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.251276731491089, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7011623382568359, + "num_tokens": 268629322.0, + "step": 10378 + }, + { + "epoch": 1.1397979354271908, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3022189140319824, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.739253044128418, + "num_tokens": 268654858.0, + "step": 10379 + }, + { + "epoch": 1.1399077531298045, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.6098380088806152, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7270165681838989, + "num_tokens": 268674996.0, + "step": 10380 + }, + { + "epoch": 1.1400175708324183, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.178504705429077, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7077524662017822, + "num_tokens": 268701396.0, + "step": 10381 + }, + { + "epoch": 1.1401273885350318, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3586032390594482, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7160618901252747, + "num_tokens": 268725730.0, + "step": 10382 + }, + { + "epoch": 1.1402372062376456, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.080292224884033, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7341499328613281, + "num_tokens": 268756266.0, + "step": 10383 + }, + { + "epoch": 1.140347023940259, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2844181060791016, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7345653772354126, + "num_tokens": 268779905.0, + "step": 10384 + }, + { + "epoch": 1.1404568416428729, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.178140163421631, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7275887131690979, + "num_tokens": 268805544.0, + "step": 10385 + }, + { + "epoch": 1.1405666593454864, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2787797451019287, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7414151430130005, + "num_tokens": 268828418.0, + "step": 10386 + }, + { + "epoch": 1.1406764770481002, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.687565803527832, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7177186012268066, + "num_tokens": 268846625.0, + "step": 10387 + }, + { + "epoch": 1.140786294750714, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.213648557662964, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.715540885925293, + "num_tokens": 268871966.0, + "step": 10388 + }, + { + "epoch": 1.1408961124533274, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.121978521347046, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7410346269607544, + "num_tokens": 268897336.0, + "step": 10389 + }, + { + "epoch": 1.1410059301559412, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2605068683624268, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7063071131706238, + "num_tokens": 268924918.0, + "step": 10390 + }, + { + "epoch": 1.1411157478585547, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2145400047302246, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7199633121490479, + "num_tokens": 268950848.0, + "step": 10391 + }, + { + "epoch": 1.1412255655611685, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2745559215545654, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7094486951828003, + "num_tokens": 268976844.0, + "step": 10392 + }, + { + "epoch": 1.141335383263782, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.157092809677124, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7250075340270996, + "num_tokens": 269004861.0, + "step": 10393 + }, + { + "epoch": 1.1414452009663958, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.469397783279419, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.727084755897522, + "num_tokens": 269025965.0, + "step": 10394 + }, + { + "epoch": 1.1415550186690093, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3618381023406982, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7295485734939575, + "num_tokens": 269048403.0, + "step": 10395 + }, + { + "epoch": 1.141664836371623, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.156062602996826, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6920738220214844, + "num_tokens": 269077585.0, + "step": 10396 + }, + { + "epoch": 1.1417746540742368, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.0527634620666504, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7404157519340515, + "num_tokens": 269108294.0, + "step": 10397 + }, + { + "epoch": 1.1418844717768504, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2088027000427246, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7117633819580078, + "num_tokens": 269135641.0, + "step": 10398 + }, + { + "epoch": 1.1419942894794641, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.9912492036819458, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.694770872592926, + "num_tokens": 269165374.0, + "step": 10399 + }, + { + "epoch": 1.1421041071820777, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4365522861480713, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.714520275592804, + "num_tokens": 269187117.0, + "step": 10400 + }, + { + "epoch": 1.1422139248846914, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2073941230773926, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7298235297203064, + "num_tokens": 269213528.0, + "step": 10401 + }, + { + "epoch": 1.1423237425873052, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.19179630279541, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7204368114471436, + "num_tokens": 269237503.0, + "step": 10402 + }, + { + "epoch": 1.1424335602899187, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.216484785079956, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7340089082717896, + "num_tokens": 269263747.0, + "step": 10403 + }, + { + "epoch": 1.1425433779925325, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.416250705718994, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7165055274963379, + "num_tokens": 269286290.0, + "step": 10404 + }, + { + "epoch": 1.142653195695146, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4535975456237793, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7424291372299194, + "num_tokens": 269306380.0, + "step": 10405 + }, + { + "epoch": 1.1427630133977598, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.38655948638916, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.721339762210846, + "num_tokens": 269330612.0, + "step": 10406 + }, + { + "epoch": 1.1428728311003733, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.339252233505249, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7408832907676697, + "num_tokens": 269352725.0, + "step": 10407 + }, + { + "epoch": 1.142982648802987, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.9196363687515259, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7137140035629272, + "num_tokens": 269385926.0, + "step": 10408 + }, + { + "epoch": 1.1430924665056006, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2684435844421387, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6988071203231812, + "num_tokens": 269411889.0, + "step": 10409 + }, + { + "epoch": 1.1432022842082143, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.346998691558838, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7039405703544617, + "num_tokens": 269436910.0, + "step": 10410 + }, + { + "epoch": 1.143312101910828, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.1500966548919678, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7124699354171753, + "num_tokens": 269464438.0, + "step": 10411 + }, + { + "epoch": 1.1434219196134416, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.261059284210205, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7101459503173828, + "num_tokens": 269491963.0, + "step": 10412 + }, + { + "epoch": 1.1435317373160554, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.089754343032837, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7115599513053894, + "num_tokens": 269521966.0, + "step": 10413 + }, + { + "epoch": 1.143641555018669, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.945508599281311, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7134228944778442, + "num_tokens": 269554796.0, + "step": 10414 + }, + { + "epoch": 1.1437513727212827, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2936460971832275, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7041573524475098, + "num_tokens": 269580105.0, + "step": 10415 + }, + { + "epoch": 1.1438611904238964, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.0972681045532227, + "learning_rate": 1e-06, + "loss": 1.0231, + "mean_token_accuracy": 0.6886278390884399, + "num_tokens": 269608531.0, + "step": 10416 + }, + { + "epoch": 1.14397100812651, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.1219277381896973, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7058026790618896, + "num_tokens": 269638760.0, + "step": 10417 + }, + { + "epoch": 1.1440808258291237, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.217643976211548, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.6974033713340759, + "num_tokens": 269666361.0, + "step": 10418 + }, + { + "epoch": 1.1441906435317373, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.197628974914551, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7135432958602905, + "num_tokens": 269692763.0, + "step": 10419 + }, + { + "epoch": 1.144300461234351, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 1.9525701999664307, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.6984457969665527, + "num_tokens": 269727061.0, + "step": 10420 + }, + { + "epoch": 1.1444102789369646, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.201974391937256, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7445617318153381, + "num_tokens": 269752838.0, + "step": 10421 + }, + { + "epoch": 1.1445200966395783, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.023118495941162, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7064847946166992, + "num_tokens": 269783618.0, + "step": 10422 + }, + { + "epoch": 1.1446299143421919, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.084209680557251, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7167844772338867, + "num_tokens": 269813083.0, + "step": 10423 + }, + { + "epoch": 1.1447397320448056, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.169361114501953, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7190577983856201, + "num_tokens": 269841125.0, + "step": 10424 + }, + { + "epoch": 1.1448495497474194, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2719476222991943, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7411696910858154, + "num_tokens": 269865835.0, + "step": 10425 + }, + { + "epoch": 1.144959367450033, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4076950550079346, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7127655148506165, + "num_tokens": 269888757.0, + "step": 10426 + }, + { + "epoch": 1.1450691851526467, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4627952575683594, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7111818194389343, + "num_tokens": 269912248.0, + "step": 10427 + }, + { + "epoch": 1.1451790028552602, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.339200258255005, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7378295660018921, + "num_tokens": 269936794.0, + "step": 10428 + }, + { + "epoch": 1.145288820557874, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.149013042449951, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7180744409561157, + "num_tokens": 269962722.0, + "step": 10429 + }, + { + "epoch": 1.1453986382604877, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.374290704727173, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7229422330856323, + "num_tokens": 269986040.0, + "step": 10430 + }, + { + "epoch": 1.1455084559631012, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.0238401889801025, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7149994969367981, + "num_tokens": 270018188.0, + "step": 10431 + }, + { + "epoch": 1.145618273665715, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.16739821434021, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7319862842559814, + "num_tokens": 270045447.0, + "step": 10432 + }, + { + "epoch": 1.1457280913683285, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.196916341781616, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7051907777786255, + "num_tokens": 270073881.0, + "step": 10433 + }, + { + "epoch": 1.1458379090709423, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.4195497035980225, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7239508032798767, + "num_tokens": 270097232.0, + "step": 10434 + }, + { + "epoch": 1.1459477267735558, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3399312496185303, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.71052086353302, + "num_tokens": 270122091.0, + "step": 10435 + }, + { + "epoch": 1.1460575444761696, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.2554664611816406, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.6987007856369019, + "num_tokens": 270147807.0, + "step": 10436 + }, + { + "epoch": 1.1461673621787831, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.388763904571533, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7226982116699219, + "num_tokens": 270169830.0, + "step": 10437 + }, + { + "epoch": 1.1462771798813969, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3621532917022705, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.726421594619751, + "num_tokens": 270194328.0, + "step": 10438 + }, + { + "epoch": 1.1463869975840106, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.161796808242798, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.717704176902771, + "num_tokens": 270219357.0, + "step": 10439 + }, + { + "epoch": 1.1464968152866242, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.422642946243286, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7125940918922424, + "num_tokens": 270241779.0, + "step": 10440 + }, + { + "epoch": 1.146606632989238, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.5913569927215576, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7279228568077087, + "num_tokens": 270261733.0, + "step": 10441 + }, + { + "epoch": 1.1467164506918515, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.445687770843506, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7213204503059387, + "num_tokens": 270283333.0, + "step": 10442 + }, + { + "epoch": 1.1468262683944652, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.1422739028930664, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7129602432250977, + "num_tokens": 270315335.0, + "step": 10443 + }, + { + "epoch": 1.1469360860970788, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.159449815750122, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7325606942176819, + "num_tokens": 270342964.0, + "step": 10444 + }, + { + "epoch": 1.1470459037996925, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.303018569946289, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.698390007019043, + "num_tokens": 270368653.0, + "step": 10445 + }, + { + "epoch": 1.147155721502306, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3128058910369873, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7138172388076782, + "num_tokens": 270393362.0, + "step": 10446 + }, + { + "epoch": 1.1472655392049198, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.393242597579956, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7346681952476501, + "num_tokens": 270416873.0, + "step": 10447 + }, + { + "epoch": 1.1473753569075336, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3322644233703613, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7249201536178589, + "num_tokens": 270440103.0, + "step": 10448 + }, + { + "epoch": 1.147485174610147, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.3723690509796143, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.723278284072876, + "num_tokens": 270462076.0, + "step": 10449 + }, + { + "epoch": 1.1475949923127609, + "ewc_loss": 9.72747802734375e-05, + "grad_norm": 2.184304714202881, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7039129734039307, + "num_tokens": 270488307.0, + "step": 10450 + }, + { + "epoch": 1.1477048100153744, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.0654683113098145, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7258964776992798, + "num_tokens": 270517987.0, + "step": 10451 + }, + { + "epoch": 1.1478146277179881, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3718600273132324, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7073439359664917, + "num_tokens": 270543828.0, + "step": 10452 + }, + { + "epoch": 1.147924445420602, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.247833728790283, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7047368288040161, + "num_tokens": 270570706.0, + "step": 10453 + }, + { + "epoch": 1.1480342631232154, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2372210025787354, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.736677885055542, + "num_tokens": 270596739.0, + "step": 10454 + }, + { + "epoch": 1.1481440808258292, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.082422971725464, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7368945479393005, + "num_tokens": 270623788.0, + "step": 10455 + }, + { + "epoch": 1.1482538985284427, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.202719211578369, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7346946597099304, + "num_tokens": 270649744.0, + "step": 10456 + }, + { + "epoch": 1.1483637162310565, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.4222757816314697, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7296289205551147, + "num_tokens": 270671889.0, + "step": 10457 + }, + { + "epoch": 1.14847353393367, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.510596513748169, + "learning_rate": 1e-06, + "loss": 0.7487, + "mean_token_accuracy": 0.7583535313606262, + "num_tokens": 270692169.0, + "step": 10458 + }, + { + "epoch": 1.1485833516362838, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.237292528152466, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7301063537597656, + "num_tokens": 270716963.0, + "step": 10459 + }, + { + "epoch": 1.1486931693388973, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1716816425323486, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7186940908432007, + "num_tokens": 270747232.0, + "step": 10460 + }, + { + "epoch": 1.148802987041511, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.024641275405884, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7084895372390747, + "num_tokens": 270777490.0, + "step": 10461 + }, + { + "epoch": 1.1489128047441248, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3826253414154053, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7273893356323242, + "num_tokens": 270799558.0, + "step": 10462 + }, + { + "epoch": 1.1490226224467384, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3132970333099365, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7087314128875732, + "num_tokens": 270824945.0, + "step": 10463 + }, + { + "epoch": 1.1491324401493521, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 1.9669394493103027, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7040325403213501, + "num_tokens": 270859452.0, + "step": 10464 + }, + { + "epoch": 1.1492422578519657, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.046489953994751, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7229976654052734, + "num_tokens": 270889401.0, + "step": 10465 + }, + { + "epoch": 1.1493520755545794, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.6138381958007812, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7297635674476624, + "num_tokens": 270911780.0, + "step": 10466 + }, + { + "epoch": 1.1494618932571932, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2295796871185303, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7162985801696777, + "num_tokens": 270939541.0, + "step": 10467 + }, + { + "epoch": 1.1495717109598067, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3210887908935547, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7188924551010132, + "num_tokens": 270961764.0, + "step": 10468 + }, + { + "epoch": 1.1496815286624205, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.5247983932495117, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7303051352500916, + "num_tokens": 270982189.0, + "step": 10469 + }, + { + "epoch": 1.149791346365034, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.8016252517700195, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7223866581916809, + "num_tokens": 271000362.0, + "step": 10470 + }, + { + "epoch": 1.1499011640676478, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3331761360168457, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7275044918060303, + "num_tokens": 271023383.0, + "step": 10471 + }, + { + "epoch": 1.1500109817702613, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.126103162765503, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7330260276794434, + "num_tokens": 271052828.0, + "step": 10472 + }, + { + "epoch": 1.150120799472875, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1678977012634277, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6997097730636597, + "num_tokens": 271081617.0, + "step": 10473 + }, + { + "epoch": 1.1502306171754886, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2167913913726807, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6813391447067261, + "num_tokens": 271110567.0, + "step": 10474 + }, + { + "epoch": 1.1503404348781023, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2873289585113525, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7112841010093689, + "num_tokens": 271134094.0, + "step": 10475 + }, + { + "epoch": 1.150450252580716, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.266312599182129, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7377513647079468, + "num_tokens": 271159825.0, + "step": 10476 + }, + { + "epoch": 1.1505600702833296, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3817715644836426, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7330772280693054, + "num_tokens": 271179892.0, + "step": 10477 + }, + { + "epoch": 1.1506698879859434, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1951541900634766, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7149393558502197, + "num_tokens": 271206416.0, + "step": 10478 + }, + { + "epoch": 1.150779705688557, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1147329807281494, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6917359828948975, + "num_tokens": 271235255.0, + "step": 10479 + }, + { + "epoch": 1.1508895233911707, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 1.881608247756958, + "learning_rate": 1e-06, + "loss": 1.0552, + "mean_token_accuracy": 0.6808751225471497, + "num_tokens": 271273431.0, + "step": 10480 + }, + { + "epoch": 1.1509993410937844, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.410994529724121, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7026515007019043, + "num_tokens": 271296536.0, + "step": 10481 + }, + { + "epoch": 1.151109158796398, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 1.990062952041626, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6931270956993103, + "num_tokens": 271331104.0, + "step": 10482 + }, + { + "epoch": 1.1512189764990117, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1920902729034424, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.719464123249054, + "num_tokens": 271358864.0, + "step": 10483 + }, + { + "epoch": 1.1513287942016253, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1383347511291504, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7086557149887085, + "num_tokens": 271385636.0, + "step": 10484 + }, + { + "epoch": 1.151438611904239, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.5158424377441406, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7262994050979614, + "num_tokens": 271406751.0, + "step": 10485 + }, + { + "epoch": 1.1515484296068526, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2026522159576416, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7023476362228394, + "num_tokens": 271433759.0, + "step": 10486 + }, + { + "epoch": 1.1516582473094663, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.5436437129974365, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7338554859161377, + "num_tokens": 271453057.0, + "step": 10487 + }, + { + "epoch": 1.1517680650120798, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.275163173675537, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7197107076644897, + "num_tokens": 271478297.0, + "step": 10488 + }, + { + "epoch": 1.1518778827146936, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3953278064727783, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7257027626037598, + "num_tokens": 271500563.0, + "step": 10489 + }, + { + "epoch": 1.1519877004173074, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.0152926445007324, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7025165557861328, + "num_tokens": 271533148.0, + "step": 10490 + }, + { + "epoch": 1.152097518119921, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.133790969848633, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7134401798248291, + "num_tokens": 271561841.0, + "step": 10491 + }, + { + "epoch": 1.1522073358225347, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.4830408096313477, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7321022152900696, + "num_tokens": 271583226.0, + "step": 10492 + }, + { + "epoch": 1.1523171535251482, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.376983404159546, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7287997603416443, + "num_tokens": 271605953.0, + "step": 10493 + }, + { + "epoch": 1.152426971227762, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.240180730819702, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7214933633804321, + "num_tokens": 271632154.0, + "step": 10494 + }, + { + "epoch": 1.1525367889303757, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.476820468902588, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7241702079772949, + "num_tokens": 271653841.0, + "step": 10495 + }, + { + "epoch": 1.1526466066329892, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2219033241271973, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7206698656082153, + "num_tokens": 271677977.0, + "step": 10496 + }, + { + "epoch": 1.152756424335603, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.3361713886260986, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7178163528442383, + "num_tokens": 271701221.0, + "step": 10497 + }, + { + "epoch": 1.1528662420382165, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.0898303985595703, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6975511908531189, + "num_tokens": 271732023.0, + "step": 10498 + }, + { + "epoch": 1.1529760597408303, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2103374004364014, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7240532040596008, + "num_tokens": 271757810.0, + "step": 10499 + }, + { + "epoch": 1.1530858774434438, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.9196786880493164, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7260814905166626, + "num_tokens": 271775593.0, + "step": 10500 + }, + { + "epoch": 1.1531956951460576, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.184267044067383, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7290225028991699, + "num_tokens": 271802380.0, + "step": 10501 + }, + { + "epoch": 1.153305512848671, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.186936855316162, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7147995829582214, + "num_tokens": 271829159.0, + "step": 10502 + }, + { + "epoch": 1.1534153305512849, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.352818727493286, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7126601934432983, + "num_tokens": 271853961.0, + "step": 10503 + }, + { + "epoch": 1.1535251482538986, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.1436123847961426, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7285897731781006, + "num_tokens": 271880431.0, + "step": 10504 + }, + { + "epoch": 1.1536349659565122, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.016897201538086, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7198207378387451, + "num_tokens": 271908730.0, + "step": 10505 + }, + { + "epoch": 1.153744783659126, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2848751544952393, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7339355945587158, + "num_tokens": 271931465.0, + "step": 10506 + }, + { + "epoch": 1.1538546013617395, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.314314126968384, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7230610847473145, + "num_tokens": 271957167.0, + "step": 10507 + }, + { + "epoch": 1.1539644190643532, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.2531397342681885, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7331898212432861, + "num_tokens": 271982253.0, + "step": 10508 + }, + { + "epoch": 1.1540742367669667, + "ewc_loss": 9.775161743164062e-05, + "grad_norm": 2.066281795501709, + "learning_rate": 1e-06, + "loss": 1.0667, + "mean_token_accuracy": 0.6803714632987976, + "num_tokens": 272014376.0, + "step": 10509 + }, + { + "epoch": 1.1541840544695805, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1351873874664307, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7191633582115173, + "num_tokens": 272043291.0, + "step": 10510 + }, + { + "epoch": 1.154293872172194, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1328914165496826, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.6976497173309326, + "num_tokens": 272071838.0, + "step": 10511 + }, + { + "epoch": 1.1544036898748078, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.17581844329834, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7134110927581787, + "num_tokens": 272099074.0, + "step": 10512 + }, + { + "epoch": 1.1545135075774215, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3600475788116455, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7297008633613586, + "num_tokens": 272121508.0, + "step": 10513 + }, + { + "epoch": 1.154623325280035, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2375316619873047, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7336946725845337, + "num_tokens": 272145371.0, + "step": 10514 + }, + { + "epoch": 1.1547331429826488, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2918763160705566, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7081257104873657, + "num_tokens": 272170042.0, + "step": 10515 + }, + { + "epoch": 1.1548429606852624, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.11609148979187, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7353103756904602, + "num_tokens": 272197860.0, + "step": 10516 + }, + { + "epoch": 1.1549527783878761, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.5983121395111084, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.730359673500061, + "num_tokens": 272219986.0, + "step": 10517 + }, + { + "epoch": 1.15506259609049, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.259916067123413, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7124243974685669, + "num_tokens": 272247393.0, + "step": 10518 + }, + { + "epoch": 1.1551724137931034, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.5666332244873047, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7419220209121704, + "num_tokens": 272266856.0, + "step": 10519 + }, + { + "epoch": 1.1552822314957172, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1751623153686523, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7040481567382812, + "num_tokens": 272296842.0, + "step": 10520 + }, + { + "epoch": 1.1553920491983307, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.126312494277954, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6929911971092224, + "num_tokens": 272324349.0, + "step": 10521 + }, + { + "epoch": 1.1555018669009445, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.150585174560547, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7323075532913208, + "num_tokens": 272351169.0, + "step": 10522 + }, + { + "epoch": 1.155611684603558, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3206489086151123, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.707870364189148, + "num_tokens": 272375508.0, + "step": 10523 + }, + { + "epoch": 1.1557215023061718, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.959741473197937, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7183825373649597, + "num_tokens": 272405700.0, + "step": 10524 + }, + { + "epoch": 1.1558313200087853, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0956830978393555, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6920288801193237, + "num_tokens": 272435400.0, + "step": 10525 + }, + { + "epoch": 1.155941137711399, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.276867151260376, + "learning_rate": 1e-06, + "loss": 1.0891, + "mean_token_accuracy": 0.6809430122375488, + "num_tokens": 272462424.0, + "step": 10526 + }, + { + "epoch": 1.1560509554140128, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.181755781173706, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6964279413223267, + "num_tokens": 272492794.0, + "step": 10527 + }, + { + "epoch": 1.1561607731166264, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2751975059509277, + "learning_rate": 1e-06, + "loss": 0.7369, + "mean_token_accuracy": 0.7667309045791626, + "num_tokens": 272515654.0, + "step": 10528 + }, + { + "epoch": 1.15627059081924, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2829723358154297, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7188609838485718, + "num_tokens": 272540293.0, + "step": 10529 + }, + { + "epoch": 1.1563804085218536, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1546084880828857, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7229189872741699, + "num_tokens": 272570385.0, + "step": 10530 + }, + { + "epoch": 1.1564902262244674, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.398994207382202, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.712986409664154, + "num_tokens": 272591512.0, + "step": 10531 + }, + { + "epoch": 1.1566000439270812, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.418851852416992, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7196359038352966, + "num_tokens": 272614670.0, + "step": 10532 + }, + { + "epoch": 1.1567098616296947, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.244133710861206, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7170778512954712, + "num_tokens": 272640736.0, + "step": 10533 + }, + { + "epoch": 1.1568196793323084, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2532341480255127, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7167800664901733, + "num_tokens": 272665179.0, + "step": 10534 + }, + { + "epoch": 1.156929497034922, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.024587392807007, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.728226363658905, + "num_tokens": 272693745.0, + "step": 10535 + }, + { + "epoch": 1.1570393147375357, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.5923962593078613, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.727070689201355, + "num_tokens": 272714068.0, + "step": 10536 + }, + { + "epoch": 1.1571491324401493, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.344280242919922, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7034974098205566, + "num_tokens": 272737718.0, + "step": 10537 + }, + { + "epoch": 1.157258950142763, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.5070531368255615, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7252440452575684, + "num_tokens": 272760243.0, + "step": 10538 + }, + { + "epoch": 1.1573687678453766, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.109499454498291, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7075327038764954, + "num_tokens": 272788577.0, + "step": 10539 + }, + { + "epoch": 1.1574785855479903, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1837334632873535, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6922321319580078, + "num_tokens": 272816076.0, + "step": 10540 + }, + { + "epoch": 1.157588403250604, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3925201892852783, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7197288274765015, + "num_tokens": 272838322.0, + "step": 10541 + }, + { + "epoch": 1.1576982209532176, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2382755279541016, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7394241094589233, + "num_tokens": 272864169.0, + "step": 10542 + }, + { + "epoch": 1.1578080386558314, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2504935264587402, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7265375256538391, + "num_tokens": 272888136.0, + "step": 10543 + }, + { + "epoch": 1.157917856358445, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2833304405212402, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7079785466194153, + "num_tokens": 272912632.0, + "step": 10544 + }, + { + "epoch": 1.1580276740610587, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.318047046661377, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7132554054260254, + "num_tokens": 272939621.0, + "step": 10545 + }, + { + "epoch": 1.1581374917636724, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3523480892181396, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7059526443481445, + "num_tokens": 272963177.0, + "step": 10546 + }, + { + "epoch": 1.158247309466286, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.281599283218384, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.736867368221283, + "num_tokens": 272987430.0, + "step": 10547 + }, + { + "epoch": 1.1583571271688997, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.47792387008667, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7286079525947571, + "num_tokens": 273008216.0, + "step": 10548 + }, + { + "epoch": 1.1584669448715132, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.4512405395507812, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.71236652135849, + "num_tokens": 273031222.0, + "step": 10549 + }, + { + "epoch": 1.158576762574127, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3929495811462402, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7298521399497986, + "num_tokens": 273053142.0, + "step": 10550 + }, + { + "epoch": 1.1586865802767405, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2274010181427, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.704473614692688, + "num_tokens": 273080271.0, + "step": 10551 + }, + { + "epoch": 1.1587963979793543, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3249356746673584, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7262074947357178, + "num_tokens": 273103267.0, + "step": 10552 + }, + { + "epoch": 1.1589062156819678, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2336487770080566, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7141889333724976, + "num_tokens": 273129129.0, + "step": 10553 + }, + { + "epoch": 1.1590160333845816, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0165293216705322, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7300350666046143, + "num_tokens": 273159544.0, + "step": 10554 + }, + { + "epoch": 1.1591258510871953, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1186609268188477, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7191520929336548, + "num_tokens": 273186077.0, + "step": 10555 + }, + { + "epoch": 1.1592356687898089, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1577847003936768, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7176761627197266, + "num_tokens": 273213808.0, + "step": 10556 + }, + { + "epoch": 1.1593454864924226, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2004034519195557, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7258425951004028, + "num_tokens": 273242014.0, + "step": 10557 + }, + { + "epoch": 1.1594553041950362, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0291099548339844, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7111668586730957, + "num_tokens": 273272011.0, + "step": 10558 + }, + { + "epoch": 1.15956512189765, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3543083667755127, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6953876614570618, + "num_tokens": 273296193.0, + "step": 10559 + }, + { + "epoch": 1.1596749396002635, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.153423309326172, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7188645005226135, + "num_tokens": 273324154.0, + "step": 10560 + }, + { + "epoch": 1.1597847573028772, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9895623922348022, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7018961906433105, + "num_tokens": 273356852.0, + "step": 10561 + }, + { + "epoch": 1.159894575005491, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1482012271881104, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7019339799880981, + "num_tokens": 273386090.0, + "step": 10562 + }, + { + "epoch": 1.1600043927081045, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.045691967010498, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7144994735717773, + "num_tokens": 273416319.0, + "step": 10563 + }, + { + "epoch": 1.1601142104107183, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.4146909713745117, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7087569832801819, + "num_tokens": 273439730.0, + "step": 10564 + }, + { + "epoch": 1.1602240281133318, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.238642930984497, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7009317874908447, + "num_tokens": 273465595.0, + "step": 10565 + }, + { + "epoch": 1.1603338458159456, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.092914342880249, + "learning_rate": 1e-06, + "loss": 1.0398, + "mean_token_accuracy": 0.682266354560852, + "num_tokens": 273496619.0, + "step": 10566 + }, + { + "epoch": 1.160443663518559, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.560429573059082, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7324462532997131, + "num_tokens": 273517232.0, + "step": 10567 + }, + { + "epoch": 1.1605534812211729, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.273043632507324, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7097687125205994, + "num_tokens": 273543123.0, + "step": 10568 + }, + { + "epoch": 1.1606632989237866, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.015442132949829, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.716835618019104, + "num_tokens": 273573824.0, + "step": 10569 + }, + { + "epoch": 1.1607731166264001, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3586745262145996, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.733910858631134, + "num_tokens": 273596774.0, + "step": 10570 + }, + { + "epoch": 1.160882934329014, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.018801212310791, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7185313701629639, + "num_tokens": 273625480.0, + "step": 10571 + }, + { + "epoch": 1.1609927520316274, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2366838455200195, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7200255393981934, + "num_tokens": 273650325.0, + "step": 10572 + }, + { + "epoch": 1.1611025697342412, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.330301523208618, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6982215642929077, + "num_tokens": 273673384.0, + "step": 10573 + }, + { + "epoch": 1.1612123874368547, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.350623607635498, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7198820114135742, + "num_tokens": 273696548.0, + "step": 10574 + }, + { + "epoch": 1.1613222051394685, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0586109161376953, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7116498351097107, + "num_tokens": 273725958.0, + "step": 10575 + }, + { + "epoch": 1.161432022842082, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9649068117141724, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7084190845489502, + "num_tokens": 273758568.0, + "step": 10576 + }, + { + "epoch": 1.1615418405446958, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.154106616973877, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7529184818267822, + "num_tokens": 273786322.0, + "step": 10577 + }, + { + "epoch": 1.1616516582473095, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2374956607818604, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.719661295413971, + "num_tokens": 273812749.0, + "step": 10578 + }, + { + "epoch": 1.161761475949923, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.238676071166992, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7271963357925415, + "num_tokens": 273836812.0, + "step": 10579 + }, + { + "epoch": 1.1618712936525368, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.372008800506592, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6890081763267517, + "num_tokens": 273861639.0, + "step": 10580 + }, + { + "epoch": 1.1619811113551504, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.09018611907959, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.717499852180481, + "num_tokens": 273890785.0, + "step": 10581 + }, + { + "epoch": 1.1620909290577641, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.049041509628296, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7112318277359009, + "num_tokens": 273920468.0, + "step": 10582 + }, + { + "epoch": 1.1622007467603779, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.164729118347168, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.741592288017273, + "num_tokens": 273946197.0, + "step": 10583 + }, + { + "epoch": 1.1623105644629914, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.109065055847168, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7110862731933594, + "num_tokens": 273974399.0, + "step": 10584 + }, + { + "epoch": 1.1624203821656052, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2559428215026855, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7235547304153442, + "num_tokens": 274001424.0, + "step": 10585 + }, + { + "epoch": 1.1625301998682187, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.550591468811035, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7037853598594666, + "num_tokens": 274024276.0, + "step": 10586 + }, + { + "epoch": 1.1626400175708325, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.318871259689331, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7067695260047913, + "num_tokens": 274048851.0, + "step": 10587 + }, + { + "epoch": 1.162749835273446, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.39880633354187, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7165327072143555, + "num_tokens": 274071445.0, + "step": 10588 + }, + { + "epoch": 1.1628596529760598, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.051164150238037, + "learning_rate": 1e-06, + "loss": 0.7484, + "mean_token_accuracy": 0.7586320638656616, + "num_tokens": 274098455.0, + "step": 10589 + }, + { + "epoch": 1.1629694706786733, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.205662727355957, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7433229684829712, + "num_tokens": 274124178.0, + "step": 10590 + }, + { + "epoch": 1.163079288381287, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1644797325134277, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7166064977645874, + "num_tokens": 274151069.0, + "step": 10591 + }, + { + "epoch": 1.1631891060839008, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.4709970951080322, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7086198329925537, + "num_tokens": 274173561.0, + "step": 10592 + }, + { + "epoch": 1.1632989237865143, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.047818899154663, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6850099563598633, + "num_tokens": 274205144.0, + "step": 10593 + }, + { + "epoch": 1.163408741489128, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2312114238739014, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7010186910629272, + "num_tokens": 274231929.0, + "step": 10594 + }, + { + "epoch": 1.1635185591917416, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3724138736724854, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7368436455726624, + "num_tokens": 274256158.0, + "step": 10595 + }, + { + "epoch": 1.1636283768943554, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.365189552307129, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6994552612304688, + "num_tokens": 274283026.0, + "step": 10596 + }, + { + "epoch": 1.1637381945969691, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.231938600540161, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7008107900619507, + "num_tokens": 274309775.0, + "step": 10597 + }, + { + "epoch": 1.1638480122995827, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.420484781265259, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7164472937583923, + "num_tokens": 274334540.0, + "step": 10598 + }, + { + "epoch": 1.1639578300021964, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.4600887298583984, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7167264223098755, + "num_tokens": 274357535.0, + "step": 10599 + }, + { + "epoch": 1.16406764770481, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2882041931152344, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7198039293289185, + "num_tokens": 274382477.0, + "step": 10600 + }, + { + "epoch": 1.1641774654074237, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.269951820373535, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7289917469024658, + "num_tokens": 274408585.0, + "step": 10601 + }, + { + "epoch": 1.1642872831100373, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.012921094894409, + "learning_rate": 1e-06, + "loss": 0.783, + "mean_token_accuracy": 0.748487651348114, + "num_tokens": 274437220.0, + "step": 10602 + }, + { + "epoch": 1.164397100812651, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9981858730316162, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7180298566818237, + "num_tokens": 274469802.0, + "step": 10603 + }, + { + "epoch": 1.1645069185152646, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9861376285552979, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7135621309280396, + "num_tokens": 274499495.0, + "step": 10604 + }, + { + "epoch": 1.1646167362178783, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3313281536102295, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.729688286781311, + "num_tokens": 274522048.0, + "step": 10605 + }, + { + "epoch": 1.164726553920492, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1758549213409424, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.699532151222229, + "num_tokens": 274550397.0, + "step": 10606 + }, + { + "epoch": 1.1648363716231056, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3450019359588623, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7270869016647339, + "num_tokens": 274572289.0, + "step": 10607 + }, + { + "epoch": 1.1649461893257194, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.260206699371338, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7088345289230347, + "num_tokens": 274598271.0, + "step": 10608 + }, + { + "epoch": 1.165056007028333, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9994562864303589, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7207539081573486, + "num_tokens": 274626039.0, + "step": 10609 + }, + { + "epoch": 1.1651658247309467, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.064567804336548, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.705322802066803, + "num_tokens": 274657055.0, + "step": 10610 + }, + { + "epoch": 1.1652756424335604, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1439831256866455, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.723397970199585, + "num_tokens": 274685455.0, + "step": 10611 + }, + { + "epoch": 1.165385460136174, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.400195598602295, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6952346563339233, + "num_tokens": 274709761.0, + "step": 10612 + }, + { + "epoch": 1.1654952778387877, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.3713152408599854, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.723927915096283, + "num_tokens": 274734396.0, + "step": 10613 + }, + { + "epoch": 1.1656050955414012, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0074522495269775, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.714706301689148, + "num_tokens": 274768069.0, + "step": 10614 + }, + { + "epoch": 1.165714913244015, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9811732769012451, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.724899172782898, + "num_tokens": 274800029.0, + "step": 10615 + }, + { + "epoch": 1.1658247309466285, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.6787116527557373, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7275183200836182, + "num_tokens": 274819886.0, + "step": 10616 + }, + { + "epoch": 1.1659345486492423, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2340524196624756, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7094690799713135, + "num_tokens": 274848395.0, + "step": 10617 + }, + { + "epoch": 1.1660443663518558, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2537503242492676, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7191897630691528, + "num_tokens": 274872749.0, + "step": 10618 + }, + { + "epoch": 1.1661541840544696, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2445547580718994, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7305281758308411, + "num_tokens": 274897661.0, + "step": 10619 + }, + { + "epoch": 1.1662640017570833, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.243929862976074, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7054029703140259, + "num_tokens": 274922900.0, + "step": 10620 + }, + { + "epoch": 1.1663738194596969, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.089301109313965, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7186112403869629, + "num_tokens": 274953199.0, + "step": 10621 + }, + { + "epoch": 1.1664836371623106, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.4231321811676025, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7241207361221313, + "num_tokens": 274975312.0, + "step": 10622 + }, + { + "epoch": 1.1665934548649242, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1163549423217773, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6970359086990356, + "num_tokens": 275005756.0, + "step": 10623 + }, + { + "epoch": 1.166703272567538, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.2927162647247314, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7323096990585327, + "num_tokens": 275029331.0, + "step": 10624 + }, + { + "epoch": 1.1668130902701515, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.521183967590332, + "learning_rate": 1e-06, + "loss": 1.066, + "mean_token_accuracy": 0.6810997724533081, + "num_tokens": 275054582.0, + "step": 10625 + }, + { + "epoch": 1.1669229079727652, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.36449933052063, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.738433301448822, + "num_tokens": 275079544.0, + "step": 10626 + }, + { + "epoch": 1.167032725675379, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.1126396656036377, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7073987126350403, + "num_tokens": 275109495.0, + "step": 10627 + }, + { + "epoch": 1.1671425433779925, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.324462413787842, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7167854309082031, + "num_tokens": 275132812.0, + "step": 10628 + }, + { + "epoch": 1.1672523610806063, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0773561000823975, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6949617862701416, + "num_tokens": 275161256.0, + "step": 10629 + }, + { + "epoch": 1.1673621787832198, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.247570753097534, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7276885509490967, + "num_tokens": 275186318.0, + "step": 10630 + }, + { + "epoch": 1.1674719964858336, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.509655475616455, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6951273083686829, + "num_tokens": 275207258.0, + "step": 10631 + }, + { + "epoch": 1.167581814188447, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.269918918609619, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7157078981399536, + "num_tokens": 275233345.0, + "step": 10632 + }, + { + "epoch": 1.1676916318910608, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.28387713432312, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7328322529792786, + "num_tokens": 275258457.0, + "step": 10633 + }, + { + "epoch": 1.1678014495936746, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.618170738220215, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7337608933448792, + "num_tokens": 275278010.0, + "step": 10634 + }, + { + "epoch": 1.1679112672962881, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.315265655517578, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7235422134399414, + "num_tokens": 275304279.0, + "step": 10635 + }, + { + "epoch": 1.168021084998902, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.45322322845459, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7158141732215881, + "num_tokens": 275326567.0, + "step": 10636 + }, + { + "epoch": 1.1681309027015154, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.8820531368255615, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6884053945541382, + "num_tokens": 275363466.0, + "step": 10637 + }, + { + "epoch": 1.1682407204041292, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.5249550342559814, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7357878684997559, + "num_tokens": 275383405.0, + "step": 10638 + }, + { + "epoch": 1.1683505381067427, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0794434547424316, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7287006378173828, + "num_tokens": 275410995.0, + "step": 10639 + }, + { + "epoch": 1.1684603558093565, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0076658725738525, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6963942646980286, + "num_tokens": 275442348.0, + "step": 10640 + }, + { + "epoch": 1.16857017351197, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 1.9822907447814941, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7081815004348755, + "num_tokens": 275473597.0, + "step": 10641 + }, + { + "epoch": 1.1686799912145838, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.0194079875946045, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6980591416358948, + "num_tokens": 275505412.0, + "step": 10642 + }, + { + "epoch": 1.1687898089171975, + "ewc_loss": 9.870529174804688e-05, + "grad_norm": 2.414811611175537, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7146939635276794, + "num_tokens": 275528121.0, + "step": 10643 + }, + { + "epoch": 1.168899626619811, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3312487602233887, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7401224374771118, + "num_tokens": 275549898.0, + "step": 10644 + }, + { + "epoch": 1.1690094443224248, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0558245182037354, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6931288242340088, + "num_tokens": 275579992.0, + "step": 10645 + }, + { + "epoch": 1.1691192620250384, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4750936031341553, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7100881338119507, + "num_tokens": 275602864.0, + "step": 10646 + }, + { + "epoch": 1.169229079727652, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0293378829956055, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7298522591590881, + "num_tokens": 275631977.0, + "step": 10647 + }, + { + "epoch": 1.1693388974302659, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2008867263793945, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7210246324539185, + "num_tokens": 275661055.0, + "step": 10648 + }, + { + "epoch": 1.1694487151328794, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.159048318862915, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7304353713989258, + "num_tokens": 275686776.0, + "step": 10649 + }, + { + "epoch": 1.1695585328354932, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0684592723846436, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7422812581062317, + "num_tokens": 275715084.0, + "step": 10650 + }, + { + "epoch": 1.1696683505381067, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.939643383026123, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6936672925949097, + "num_tokens": 275748446.0, + "step": 10651 + }, + { + "epoch": 1.1697781682407205, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.472865104675293, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7060370445251465, + "num_tokens": 275770040.0, + "step": 10652 + }, + { + "epoch": 1.169887985943334, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4566235542297363, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7226172685623169, + "num_tokens": 275793328.0, + "step": 10653 + }, + { + "epoch": 1.1699978036459477, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0364115238189697, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7084217071533203, + "num_tokens": 275823134.0, + "step": 10654 + }, + { + "epoch": 1.1701076213485613, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1834638118743896, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.702826201915741, + "num_tokens": 275851797.0, + "step": 10655 + }, + { + "epoch": 1.170217439051175, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4479925632476807, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7146577835083008, + "num_tokens": 275874996.0, + "step": 10656 + }, + { + "epoch": 1.1703272567537888, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.226125717163086, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.711436927318573, + "num_tokens": 275901105.0, + "step": 10657 + }, + { + "epoch": 1.1704370744564023, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.192430019378662, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7138811349868774, + "num_tokens": 275928391.0, + "step": 10658 + }, + { + "epoch": 1.170546892159016, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.221710681915283, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7211166024208069, + "num_tokens": 275955151.0, + "step": 10659 + }, + { + "epoch": 1.1706567098616296, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4206199645996094, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7178261280059814, + "num_tokens": 275977107.0, + "step": 10660 + }, + { + "epoch": 1.1707665275642434, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.069899320602417, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6865754723548889, + "num_tokens": 276006547.0, + "step": 10661 + }, + { + "epoch": 1.1708763452668571, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.9986101388931274, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7168796062469482, + "num_tokens": 276034604.0, + "step": 10662 + }, + { + "epoch": 1.1709861629694707, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.20418643951416, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.702049970626831, + "num_tokens": 276062945.0, + "step": 10663 + }, + { + "epoch": 1.1710959806720844, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.6945698261260986, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7451075315475464, + "num_tokens": 276081219.0, + "step": 10664 + }, + { + "epoch": 1.171205798374698, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3360025882720947, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.704559862613678, + "num_tokens": 276105780.0, + "step": 10665 + }, + { + "epoch": 1.1713156160773117, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.378051996231079, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7338531017303467, + "num_tokens": 276127673.0, + "step": 10666 + }, + { + "epoch": 1.1714254337799253, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.153687000274658, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7165704965591431, + "num_tokens": 276155202.0, + "step": 10667 + }, + { + "epoch": 1.171535251482539, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.253091335296631, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7087407112121582, + "num_tokens": 276182164.0, + "step": 10668 + }, + { + "epoch": 1.1716450691851525, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.9958575963974, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7312270402908325, + "num_tokens": 276210888.0, + "step": 10669 + }, + { + "epoch": 1.1717548868877663, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0359764099121094, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7251111268997192, + "num_tokens": 276238745.0, + "step": 10670 + }, + { + "epoch": 1.17186470459038, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.394282102584839, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7258564829826355, + "num_tokens": 276262017.0, + "step": 10671 + }, + { + "epoch": 1.1719745222929936, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1013543605804443, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7099720239639282, + "num_tokens": 276291389.0, + "step": 10672 + }, + { + "epoch": 1.1720843399956073, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4102442264556885, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7244888544082642, + "num_tokens": 276314507.0, + "step": 10673 + }, + { + "epoch": 1.1721941576982209, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.084019184112549, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7170530557632446, + "num_tokens": 276341334.0, + "step": 10674 + }, + { + "epoch": 1.1723039754008346, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1884067058563232, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7124311327934265, + "num_tokens": 276366493.0, + "step": 10675 + }, + { + "epoch": 1.1724137931034484, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4020936489105225, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7235140204429626, + "num_tokens": 276388674.0, + "step": 10676 + }, + { + "epoch": 1.172523610806062, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.231401205062866, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7085121870040894, + "num_tokens": 276412824.0, + "step": 10677 + }, + { + "epoch": 1.1726334285086757, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2682971954345703, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7249575853347778, + "num_tokens": 276440248.0, + "step": 10678 + }, + { + "epoch": 1.1727432462112892, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2215919494628906, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.723671555519104, + "num_tokens": 276466961.0, + "step": 10679 + }, + { + "epoch": 1.172853063913903, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4109959602355957, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7443218231201172, + "num_tokens": 276488568.0, + "step": 10680 + }, + { + "epoch": 1.1729628816165165, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.294342279434204, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7234053611755371, + "num_tokens": 276514628.0, + "step": 10681 + }, + { + "epoch": 1.1730726993191303, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0276026725769043, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7137686014175415, + "num_tokens": 276546001.0, + "step": 10682 + }, + { + "epoch": 1.1731825170217438, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3833625316619873, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7250507473945618, + "num_tokens": 276567144.0, + "step": 10683 + }, + { + "epoch": 1.1732923347243576, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.015904426574707, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7031543254852295, + "num_tokens": 276600189.0, + "step": 10684 + }, + { + "epoch": 1.1734021524269713, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.303858757019043, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7197272777557373, + "num_tokens": 276626814.0, + "step": 10685 + }, + { + "epoch": 1.1735119701295849, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0706374645233154, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6842213869094849, + "num_tokens": 276660157.0, + "step": 10686 + }, + { + "epoch": 1.1736217878321986, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.356914520263672, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7101102471351624, + "num_tokens": 276687209.0, + "step": 10687 + }, + { + "epoch": 1.1737316055348122, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.9620435237884521, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7235414981842041, + "num_tokens": 276719689.0, + "step": 10688 + }, + { + "epoch": 1.173841423237426, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.9894211292266846, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7200217247009277, + "num_tokens": 276750718.0, + "step": 10689 + }, + { + "epoch": 1.1739512409400394, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.309269428253174, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7136961817741394, + "num_tokens": 276775344.0, + "step": 10690 + }, + { + "epoch": 1.1740610586426532, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.201207399368286, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.6932756900787354, + "num_tokens": 276802480.0, + "step": 10691 + }, + { + "epoch": 1.1741708763452667, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3439910411834717, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.748584508895874, + "num_tokens": 276825093.0, + "step": 10692 + }, + { + "epoch": 1.1742806940478805, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1448662281036377, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.6893147230148315, + "num_tokens": 276859244.0, + "step": 10693 + }, + { + "epoch": 1.1743905117504942, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.231638193130493, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.719893753528595, + "num_tokens": 276883396.0, + "step": 10694 + }, + { + "epoch": 1.1745003294531078, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1305460929870605, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7314639687538147, + "num_tokens": 276910340.0, + "step": 10695 + }, + { + "epoch": 1.1746101471557215, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3372013568878174, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7264721989631653, + "num_tokens": 276935607.0, + "step": 10696 + }, + { + "epoch": 1.174719964858335, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4656479358673096, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7167748212814331, + "num_tokens": 276958407.0, + "step": 10697 + }, + { + "epoch": 1.1748297825609488, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2537665367126465, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7104881405830383, + "num_tokens": 276984274.0, + "step": 10698 + }, + { + "epoch": 1.1749396002635626, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4963510036468506, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7396047711372375, + "num_tokens": 277003924.0, + "step": 10699 + }, + { + "epoch": 1.1750494179661761, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3044464588165283, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7075549364089966, + "num_tokens": 277029326.0, + "step": 10700 + }, + { + "epoch": 1.1751592356687899, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3414225578308105, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7463361024856567, + "num_tokens": 277052982.0, + "step": 10701 + }, + { + "epoch": 1.1752690533714034, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0588817596435547, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7227574586868286, + "num_tokens": 277082856.0, + "step": 10702 + }, + { + "epoch": 1.1753788710740172, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.190222978591919, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.6939623355865479, + "num_tokens": 277110580.0, + "step": 10703 + }, + { + "epoch": 1.1754886887766307, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3700225353240967, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7197067141532898, + "num_tokens": 277133087.0, + "step": 10704 + }, + { + "epoch": 1.1755985064792445, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.252228021621704, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7186487913131714, + "num_tokens": 277157403.0, + "step": 10705 + }, + { + "epoch": 1.175708324181858, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.144096851348877, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7133444547653198, + "num_tokens": 277183739.0, + "step": 10706 + }, + { + "epoch": 1.1758181418844718, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0880491733551025, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7065210342407227, + "num_tokens": 277210821.0, + "step": 10707 + }, + { + "epoch": 1.1759279595870855, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.164092779159546, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7297497987747192, + "num_tokens": 277236725.0, + "step": 10708 + }, + { + "epoch": 1.176037777289699, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 1.9942277669906616, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7104897499084473, + "num_tokens": 277268483.0, + "step": 10709 + }, + { + "epoch": 1.1761475949923128, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.111635208129883, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7290799021720886, + "num_tokens": 277296117.0, + "step": 10710 + }, + { + "epoch": 1.1762574126949263, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.43265962600708, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7224742770195007, + "num_tokens": 277318543.0, + "step": 10711 + }, + { + "epoch": 1.17636723039754, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1224849224090576, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7089982032775879, + "num_tokens": 277347866.0, + "step": 10712 + }, + { + "epoch": 1.1764770481001539, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.17985463142395, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6919287443161011, + "num_tokens": 277376341.0, + "step": 10713 + }, + { + "epoch": 1.1765868658027674, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1685495376586914, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7296611070632935, + "num_tokens": 277403427.0, + "step": 10714 + }, + { + "epoch": 1.1766966835053811, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.103783369064331, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7244218587875366, + "num_tokens": 277430711.0, + "step": 10715 + }, + { + "epoch": 1.1768065012079947, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0427234172821045, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.6963703632354736, + "num_tokens": 277460092.0, + "step": 10716 + }, + { + "epoch": 1.1769163189106084, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.155675172805786, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.698128879070282, + "num_tokens": 277486876.0, + "step": 10717 + }, + { + "epoch": 1.177026136613222, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3960015773773193, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.709052562713623, + "num_tokens": 277510917.0, + "step": 10718 + }, + { + "epoch": 1.1771359543158357, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2314207553863525, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7056353688240051, + "num_tokens": 277537077.0, + "step": 10719 + }, + { + "epoch": 1.1772457720184493, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.4762277603149414, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7210559844970703, + "num_tokens": 277559338.0, + "step": 10720 + }, + { + "epoch": 1.177355589721063, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.5679593086242676, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7329304814338684, + "num_tokens": 277580893.0, + "step": 10721 + }, + { + "epoch": 1.1774654074236768, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.426003932952881, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7114251255989075, + "num_tokens": 277602487.0, + "step": 10722 + }, + { + "epoch": 1.1775752251262903, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.0311477184295654, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7112066745758057, + "num_tokens": 277633982.0, + "step": 10723 + }, + { + "epoch": 1.177685042828904, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3967199325561523, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7259835004806519, + "num_tokens": 277656418.0, + "step": 10724 + }, + { + "epoch": 1.1777948605315176, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.329315185546875, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7022801041603088, + "num_tokens": 277681706.0, + "step": 10725 + }, + { + "epoch": 1.1779046782341314, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3679049015045166, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7315101623535156, + "num_tokens": 277704022.0, + "step": 10726 + }, + { + "epoch": 1.1780144959367451, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.577037811279297, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7505425214767456, + "num_tokens": 277722483.0, + "step": 10727 + }, + { + "epoch": 1.1781243136393587, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.388545274734497, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7238507270812988, + "num_tokens": 277744425.0, + "step": 10728 + }, + { + "epoch": 1.1782341313419724, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.7004714012145996, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7278293371200562, + "num_tokens": 277762468.0, + "step": 10729 + }, + { + "epoch": 1.178343949044586, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.204266309738159, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7064354419708252, + "num_tokens": 277789603.0, + "step": 10730 + }, + { + "epoch": 1.1784537667471997, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2009639739990234, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.6979082822799683, + "num_tokens": 277816454.0, + "step": 10731 + }, + { + "epoch": 1.1785635844498132, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.116096258163452, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.735414981842041, + "num_tokens": 277844616.0, + "step": 10732 + }, + { + "epoch": 1.178673402152427, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.2845311164855957, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7106389999389648, + "num_tokens": 277870836.0, + "step": 10733 + }, + { + "epoch": 1.1787832198550405, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.3064911365509033, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7280179262161255, + "num_tokens": 277894579.0, + "step": 10734 + }, + { + "epoch": 1.1788930375576543, + "ewc_loss": 9.918212890625e-05, + "grad_norm": 2.1254305839538574, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.6863656640052795, + "num_tokens": 277924162.0, + "step": 10735 + }, + { + "epoch": 1.179002855260268, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.0633535385131836, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7178360819816589, + "num_tokens": 277952941.0, + "step": 10736 + }, + { + "epoch": 1.1791126729628816, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2716543674468994, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7041664123535156, + "num_tokens": 277977943.0, + "step": 10737 + }, + { + "epoch": 1.1792224906654953, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.321791887283325, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7061883211135864, + "num_tokens": 278000702.0, + "step": 10738 + }, + { + "epoch": 1.1793323083681089, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.5191476345062256, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7217836380004883, + "num_tokens": 278023195.0, + "step": 10739 + }, + { + "epoch": 1.1794421260707226, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.154261589050293, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7271608114242554, + "num_tokens": 278051203.0, + "step": 10740 + }, + { + "epoch": 1.1795519437733364, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.155033826828003, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7224371433258057, + "num_tokens": 278076421.0, + "step": 10741 + }, + { + "epoch": 1.17966176147595, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3701133728027344, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7095354795455933, + "num_tokens": 278101100.0, + "step": 10742 + }, + { + "epoch": 1.1797715791785637, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.371307611465454, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7040929794311523, + "num_tokens": 278126360.0, + "step": 10743 + }, + { + "epoch": 1.1798813968811772, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.121951103210449, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7214735746383667, + "num_tokens": 278152646.0, + "step": 10744 + }, + { + "epoch": 1.179991214583791, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.4531986713409424, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7096661925315857, + "num_tokens": 278174769.0, + "step": 10745 + }, + { + "epoch": 1.1801010322864045, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.1118102073669434, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7256952524185181, + "num_tokens": 278203501.0, + "step": 10746 + }, + { + "epoch": 1.1802108499890183, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.113180160522461, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7310315370559692, + "num_tokens": 278230519.0, + "step": 10747 + }, + { + "epoch": 1.1803206676916318, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3103113174438477, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7296303510665894, + "num_tokens": 278256524.0, + "step": 10748 + }, + { + "epoch": 1.1804304853942456, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.4819164276123047, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7340390086174011, + "num_tokens": 278278805.0, + "step": 10749 + }, + { + "epoch": 1.1805403030968593, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.1728808879852295, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7153774499893188, + "num_tokens": 278307389.0, + "step": 10750 + }, + { + "epoch": 1.1806501207994728, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2820608615875244, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7085585594177246, + "num_tokens": 278332455.0, + "step": 10751 + }, + { + "epoch": 1.1807599385020866, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.0891594886779785, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7153314352035522, + "num_tokens": 278362515.0, + "step": 10752 + }, + { + "epoch": 1.1808697562047001, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2527589797973633, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7364352345466614, + "num_tokens": 278388616.0, + "step": 10753 + }, + { + "epoch": 1.180979573907314, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3877933025360107, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.707574188709259, + "num_tokens": 278414433.0, + "step": 10754 + }, + { + "epoch": 1.1810893916099274, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.4347472190856934, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7301846742630005, + "num_tokens": 278435896.0, + "step": 10755 + }, + { + "epoch": 1.1811992093125412, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.398876428604126, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7432798743247986, + "num_tokens": 278457544.0, + "step": 10756 + }, + { + "epoch": 1.1813090270151547, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.205872058868408, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7028417587280273, + "num_tokens": 278483248.0, + "step": 10757 + }, + { + "epoch": 1.1814188447177685, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.4008116722106934, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7114126682281494, + "num_tokens": 278506615.0, + "step": 10758 + }, + { + "epoch": 1.1815286624203822, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2112033367156982, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7121543884277344, + "num_tokens": 278533229.0, + "step": 10759 + }, + { + "epoch": 1.1816384801229958, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.1533427238464355, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7253191471099854, + "num_tokens": 278559948.0, + "step": 10760 + }, + { + "epoch": 1.1817482978256095, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.31952166557312, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7153880000114441, + "num_tokens": 278583103.0, + "step": 10761 + }, + { + "epoch": 1.181858115528223, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.24422550201416, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7243087887763977, + "num_tokens": 278610819.0, + "step": 10762 + }, + { + "epoch": 1.1819679332308368, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.0693092346191406, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7421664595603943, + "num_tokens": 278639297.0, + "step": 10763 + }, + { + "epoch": 1.1820777509334506, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.168656349182129, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.7026496529579163, + "num_tokens": 278668029.0, + "step": 10764 + }, + { + "epoch": 1.1821875686360641, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 3.0151541233062744, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7187213897705078, + "num_tokens": 278684837.0, + "step": 10765 + }, + { + "epoch": 1.1822973863386779, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.164508104324341, + "learning_rate": 1e-06, + "loss": 1.0636, + "mean_token_accuracy": 0.6789295673370361, + "num_tokens": 278716488.0, + "step": 10766 + }, + { + "epoch": 1.1824072040412914, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.5626492500305176, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7141956686973572, + "num_tokens": 278739057.0, + "step": 10767 + }, + { + "epoch": 1.1825170217439052, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.6828763484954834, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7330892086029053, + "num_tokens": 278757143.0, + "step": 10768 + }, + { + "epoch": 1.1826268394465187, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.294229745864868, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7216067910194397, + "num_tokens": 278788023.0, + "step": 10769 + }, + { + "epoch": 1.1827366571491325, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2869746685028076, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7198726534843445, + "num_tokens": 278811875.0, + "step": 10770 + }, + { + "epoch": 1.182846474851746, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.377540349960327, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7082206010818481, + "num_tokens": 278837201.0, + "step": 10771 + }, + { + "epoch": 1.1829562925543597, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.6869723796844482, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7361323237419128, + "num_tokens": 278856995.0, + "step": 10772 + }, + { + "epoch": 1.1830661102569735, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.4148058891296387, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7154035568237305, + "num_tokens": 278879410.0, + "step": 10773 + }, + { + "epoch": 1.183175927959587, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3524391651153564, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7125610113143921, + "num_tokens": 278902806.0, + "step": 10774 + }, + { + "epoch": 1.1832857456622008, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3268227577209473, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7065349221229553, + "num_tokens": 278928228.0, + "step": 10775 + }, + { + "epoch": 1.1833955633648143, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2346386909484863, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6968767642974854, + "num_tokens": 278956107.0, + "step": 10776 + }, + { + "epoch": 1.183505381067428, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.198657512664795, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7391595840454102, + "num_tokens": 278980976.0, + "step": 10777 + }, + { + "epoch": 1.1836151987700418, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.32747483253479, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7179951667785645, + "num_tokens": 279006438.0, + "step": 10778 + }, + { + "epoch": 1.1837250164726554, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.7549943923950195, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7147972583770752, + "num_tokens": 279025046.0, + "step": 10779 + }, + { + "epoch": 1.1838348341752691, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.1011171340942383, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7190577983856201, + "num_tokens": 279054000.0, + "step": 10780 + }, + { + "epoch": 1.1839446518778827, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 1.9664310216903687, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6947281956672668, + "num_tokens": 279086191.0, + "step": 10781 + }, + { + "epoch": 1.1840544695804964, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.5615463256835938, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7232193946838379, + "num_tokens": 279105947.0, + "step": 10782 + }, + { + "epoch": 1.18416428728311, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.221195936203003, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7351458668708801, + "num_tokens": 279132323.0, + "step": 10783 + }, + { + "epoch": 1.1842741049857237, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2880747318267822, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7071783542633057, + "num_tokens": 279157452.0, + "step": 10784 + }, + { + "epoch": 1.1843839226883373, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.1221017837524414, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7166051268577576, + "num_tokens": 279185069.0, + "step": 10785 + }, + { + "epoch": 1.184493740390951, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.155536413192749, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7284739017486572, + "num_tokens": 279212124.0, + "step": 10786 + }, + { + "epoch": 1.1846035580935648, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2596945762634277, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.720009446144104, + "num_tokens": 279239591.0, + "step": 10787 + }, + { + "epoch": 1.1847133757961783, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.057342529296875, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.726598858833313, + "num_tokens": 279269866.0, + "step": 10788 + }, + { + "epoch": 1.184823193498792, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.341944694519043, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7103439569473267, + "num_tokens": 279293521.0, + "step": 10789 + }, + { + "epoch": 1.1849330112014056, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 1.9269969463348389, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7284886837005615, + "num_tokens": 279326470.0, + "step": 10790 + }, + { + "epoch": 1.1850428289040194, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.188821315765381, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7502346038818359, + "num_tokens": 279351307.0, + "step": 10791 + }, + { + "epoch": 1.185152646606633, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.141979932785034, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6987006664276123, + "num_tokens": 279379656.0, + "step": 10792 + }, + { + "epoch": 1.1852624643092466, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.354360580444336, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7311232089996338, + "num_tokens": 279401365.0, + "step": 10793 + }, + { + "epoch": 1.1853722820118604, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 1.9936801195144653, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7020736336708069, + "num_tokens": 279431184.0, + "step": 10794 + }, + { + "epoch": 1.185482099714474, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3611528873443604, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7111152410507202, + "num_tokens": 279455386.0, + "step": 10795 + }, + { + "epoch": 1.1855919174170877, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2589950561523438, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7166336178779602, + "num_tokens": 279480157.0, + "step": 10796 + }, + { + "epoch": 1.1857017351197012, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 1.9109622240066528, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7223094701766968, + "num_tokens": 279511184.0, + "step": 10797 + }, + { + "epoch": 1.185811552822315, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3827035427093506, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7176282405853271, + "num_tokens": 279536340.0, + "step": 10798 + }, + { + "epoch": 1.1859213705249285, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.019590377807617, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7266693115234375, + "num_tokens": 279568355.0, + "step": 10799 + }, + { + "epoch": 1.1860311882275423, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3413028717041016, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.717087984085083, + "num_tokens": 279592058.0, + "step": 10800 + }, + { + "epoch": 1.186141005930156, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.336743116378784, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7058858871459961, + "num_tokens": 279616851.0, + "step": 10801 + }, + { + "epoch": 1.1862508236327696, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.2527308464050293, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7179203033447266, + "num_tokens": 279643692.0, + "step": 10802 + }, + { + "epoch": 1.1863606413353833, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.179375648498535, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7222063541412354, + "num_tokens": 279668872.0, + "step": 10803 + }, + { + "epoch": 1.1864704590379969, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.258286952972412, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7189799547195435, + "num_tokens": 279695404.0, + "step": 10804 + }, + { + "epoch": 1.1865802767406106, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.3984129428863525, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7250882983207703, + "num_tokens": 279718073.0, + "step": 10805 + }, + { + "epoch": 1.1866900944432242, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 2.248696804046631, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7101397514343262, + "num_tokens": 279747779.0, + "step": 10806 + }, + { + "epoch": 1.186799912145838, + "ewc_loss": 9.965896606445312e-05, + "grad_norm": 1.958541989326477, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7060235738754272, + "num_tokens": 279781480.0, + "step": 10807 + }, + { + "epoch": 1.1869097298484517, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.134131908416748, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7152807116508484, + "num_tokens": 279809082.0, + "step": 10808 + }, + { + "epoch": 1.1870195475510652, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.381197214126587, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7274017333984375, + "num_tokens": 279832111.0, + "step": 10809 + }, + { + "epoch": 1.187129365253679, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1548244953155518, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7130493521690369, + "num_tokens": 279860701.0, + "step": 10810 + }, + { + "epoch": 1.1872391829562925, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.040987491607666, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7057206034660339, + "num_tokens": 279891858.0, + "step": 10811 + }, + { + "epoch": 1.1873490006589063, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.565429449081421, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7264660596847534, + "num_tokens": 279911871.0, + "step": 10812 + }, + { + "epoch": 1.1874588183615198, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1595849990844727, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.715000569820404, + "num_tokens": 279937887.0, + "step": 10813 + }, + { + "epoch": 1.1875686360641335, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1502857208251953, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7119617462158203, + "num_tokens": 279965585.0, + "step": 10814 + }, + { + "epoch": 1.1876784537667473, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.679391622543335, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7093250155448914, + "num_tokens": 279985822.0, + "step": 10815 + }, + { + "epoch": 1.1877882714693608, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.00730299949646, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7208796739578247, + "num_tokens": 280016291.0, + "step": 10816 + }, + { + "epoch": 1.1878980891719746, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1250760555267334, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7060850858688354, + "num_tokens": 280045789.0, + "step": 10817 + }, + { + "epoch": 1.1880079068745881, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.4260692596435547, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7384614944458008, + "num_tokens": 280068256.0, + "step": 10818 + }, + { + "epoch": 1.1881177245772019, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.304863452911377, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7512751817703247, + "num_tokens": 280092650.0, + "step": 10819 + }, + { + "epoch": 1.1882275422798154, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.29433274269104, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.716234564781189, + "num_tokens": 280118424.0, + "step": 10820 + }, + { + "epoch": 1.1883373599824292, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.336210012435913, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7186495065689087, + "num_tokens": 280142730.0, + "step": 10821 + }, + { + "epoch": 1.1884471776850427, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.573064088821411, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.710219144821167, + "num_tokens": 280163927.0, + "step": 10822 + }, + { + "epoch": 1.1885569953876565, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.5730767250061035, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.721481204032898, + "num_tokens": 280184953.0, + "step": 10823 + }, + { + "epoch": 1.1886668130902702, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.5625290870666504, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7109998464584351, + "num_tokens": 280206670.0, + "step": 10824 + }, + { + "epoch": 1.1887766307928838, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.162526845932007, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7233257293701172, + "num_tokens": 280232882.0, + "step": 10825 + }, + { + "epoch": 1.1888864484954975, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.333062171936035, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7279515266418457, + "num_tokens": 280257097.0, + "step": 10826 + }, + { + "epoch": 1.188996266198111, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.154311418533325, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7308299541473389, + "num_tokens": 280283838.0, + "step": 10827 + }, + { + "epoch": 1.1891060839007248, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.57391619682312, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7275680303573608, + "num_tokens": 280305715.0, + "step": 10828 + }, + { + "epoch": 1.1892159016033386, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.653698682785034, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7333199977874756, + "num_tokens": 280323948.0, + "step": 10829 + }, + { + "epoch": 1.189325719305952, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.201353073120117, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7153147459030151, + "num_tokens": 280352256.0, + "step": 10830 + }, + { + "epoch": 1.1894355370085659, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.34187912940979, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6997401714324951, + "num_tokens": 280381460.0, + "step": 10831 + }, + { + "epoch": 1.1895453547111794, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.479125499725342, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7291814088821411, + "num_tokens": 280402295.0, + "step": 10832 + }, + { + "epoch": 1.1896551724137931, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3284268379211426, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.752372145652771, + "num_tokens": 280424774.0, + "step": 10833 + }, + { + "epoch": 1.1897649901164067, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.6039860248565674, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7030496597290039, + "num_tokens": 280444824.0, + "step": 10834 + }, + { + "epoch": 1.1898748078190204, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.4097111225128174, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7131232023239136, + "num_tokens": 280469134.0, + "step": 10835 + }, + { + "epoch": 1.189984625521634, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.0286386013031006, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7115150690078735, + "num_tokens": 280499041.0, + "step": 10836 + }, + { + "epoch": 1.1900944432242477, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1206514835357666, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7095502614974976, + "num_tokens": 280528314.0, + "step": 10837 + }, + { + "epoch": 1.1902042609268615, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2473020553588867, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7122081518173218, + "num_tokens": 280554192.0, + "step": 10838 + }, + { + "epoch": 1.190314078629475, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1764602661132812, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7236573696136475, + "num_tokens": 280579201.0, + "step": 10839 + }, + { + "epoch": 1.1904238963320888, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 1.965704083442688, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6958604454994202, + "num_tokens": 280613242.0, + "step": 10840 + }, + { + "epoch": 1.1905337140347023, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2499382495880127, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7245151400566101, + "num_tokens": 280641227.0, + "step": 10841 + }, + { + "epoch": 1.190643531737316, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.0092251300811768, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7125452160835266, + "num_tokens": 280671471.0, + "step": 10842 + }, + { + "epoch": 1.1907533494399298, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.363691806793213, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7046506404876709, + "num_tokens": 280695950.0, + "step": 10843 + }, + { + "epoch": 1.1908631671425434, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.7110235691070557, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7448936104774475, + "num_tokens": 280714361.0, + "step": 10844 + }, + { + "epoch": 1.1909729848451571, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.6716527938842773, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7454731464385986, + "num_tokens": 280734389.0, + "step": 10845 + }, + { + "epoch": 1.1910828025477707, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1533613204956055, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7330609560012817, + "num_tokens": 280759798.0, + "step": 10846 + }, + { + "epoch": 1.1911926202503844, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.9200148582458496, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7287347316741943, + "num_tokens": 280782156.0, + "step": 10847 + }, + { + "epoch": 1.191302437952998, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.272113561630249, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7132089138031006, + "num_tokens": 280807703.0, + "step": 10848 + }, + { + "epoch": 1.1914122556556117, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3318610191345215, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7117623090744019, + "num_tokens": 280831959.0, + "step": 10849 + }, + { + "epoch": 1.1915220733582252, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 1.9633538722991943, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7169399261474609, + "num_tokens": 280864614.0, + "step": 10850 + }, + { + "epoch": 1.191631891060839, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.241755723953247, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7058920860290527, + "num_tokens": 280891174.0, + "step": 10851 + }, + { + "epoch": 1.1917417087634528, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.110748052597046, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7164398431777954, + "num_tokens": 280919838.0, + "step": 10852 + }, + { + "epoch": 1.1918515264660663, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.6670517921447754, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7255252599716187, + "num_tokens": 280938891.0, + "step": 10853 + }, + { + "epoch": 1.19196134416868, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2975480556488037, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7186815738677979, + "num_tokens": 280962915.0, + "step": 10854 + }, + { + "epoch": 1.1920711618712936, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.0835461616516113, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7053285241127014, + "num_tokens": 280991722.0, + "step": 10855 + }, + { + "epoch": 1.1921809795739073, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.054689407348633, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.7053616046905518, + "num_tokens": 281021599.0, + "step": 10856 + }, + { + "epoch": 1.192290797276521, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 1.9953460693359375, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7235478162765503, + "num_tokens": 281051685.0, + "step": 10857 + }, + { + "epoch": 1.1924006149791346, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.411310911178589, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7143986225128174, + "num_tokens": 281074117.0, + "step": 10858 + }, + { + "epoch": 1.1925104326817484, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.221477508544922, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7066962122917175, + "num_tokens": 281102277.0, + "step": 10859 + }, + { + "epoch": 1.192620250384362, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2487831115722656, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.710752010345459, + "num_tokens": 281128923.0, + "step": 10860 + }, + { + "epoch": 1.1927300680869757, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.0992379188537598, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7455579042434692, + "num_tokens": 281156396.0, + "step": 10861 + }, + { + "epoch": 1.1928398857895892, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3315932750701904, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7256471514701843, + "num_tokens": 281181795.0, + "step": 10862 + }, + { + "epoch": 1.192949703492203, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3788444995880127, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7118990421295166, + "num_tokens": 281205956.0, + "step": 10863 + }, + { + "epoch": 1.1930595211948165, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2021312713623047, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7095441222190857, + "num_tokens": 281231044.0, + "step": 10864 + }, + { + "epoch": 1.1931693388974303, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.222825288772583, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7340453863143921, + "num_tokens": 281256379.0, + "step": 10865 + }, + { + "epoch": 1.193279156600044, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.0556232929229736, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7157268524169922, + "num_tokens": 281286506.0, + "step": 10866 + }, + { + "epoch": 1.1933889743026576, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.2127161026000977, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7118449211120605, + "num_tokens": 281315045.0, + "step": 10867 + }, + { + "epoch": 1.1934987920052713, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3616528511047363, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7237002849578857, + "num_tokens": 281338268.0, + "step": 10868 + }, + { + "epoch": 1.1936086097078848, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.601560592651367, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.732090413570404, + "num_tokens": 281358237.0, + "step": 10869 + }, + { + "epoch": 1.1937184274104986, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.4625158309936523, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7374671697616577, + "num_tokens": 281379501.0, + "step": 10870 + }, + { + "epoch": 1.1938282451131121, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.5280091762542725, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7277034521102905, + "num_tokens": 281399907.0, + "step": 10871 + }, + { + "epoch": 1.193938062815726, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.1646370887756348, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7192673683166504, + "num_tokens": 281426171.0, + "step": 10872 + }, + { + "epoch": 1.1940478805183394, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3099303245544434, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7128850221633911, + "num_tokens": 281452592.0, + "step": 10873 + }, + { + "epoch": 1.1941576982209532, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.181241273880005, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7055182456970215, + "num_tokens": 281479924.0, + "step": 10874 + }, + { + "epoch": 1.194267515923567, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.3690319061279297, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7200301289558411, + "num_tokens": 281503398.0, + "step": 10875 + }, + { + "epoch": 1.1943773336261805, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.329554557800293, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7199636697769165, + "num_tokens": 281525676.0, + "step": 10876 + }, + { + "epoch": 1.1944871513287942, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.291097640991211, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7267266511917114, + "num_tokens": 281551631.0, + "step": 10877 + }, + { + "epoch": 1.1945969690314078, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.5666017532348633, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7183393836021423, + "num_tokens": 281572796.0, + "step": 10878 + }, + { + "epoch": 1.1947067867340215, + "ewc_loss": 0.00010013580322265625, + "grad_norm": 2.4280741214752197, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.714228093624115, + "num_tokens": 281597170.0, + "step": 10879 + }, + { + "epoch": 1.1948166044366353, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4022538661956787, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7263535261154175, + "num_tokens": 281620614.0, + "step": 10880 + }, + { + "epoch": 1.1949264221392488, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2275311946868896, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7177993059158325, + "num_tokens": 281646288.0, + "step": 10881 + }, + { + "epoch": 1.1950362398418626, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.426136016845703, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.732895016670227, + "num_tokens": 281667782.0, + "step": 10882 + }, + { + "epoch": 1.1951460575444761, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.203618049621582, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7246003150939941, + "num_tokens": 281694308.0, + "step": 10883 + }, + { + "epoch": 1.1952558752470899, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.402919292449951, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7323811054229736, + "num_tokens": 281716188.0, + "step": 10884 + }, + { + "epoch": 1.1953656929497034, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2096872329711914, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6967854499816895, + "num_tokens": 281743927.0, + "step": 10885 + }, + { + "epoch": 1.1954755106523172, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.114655017852783, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7170267701148987, + "num_tokens": 281770630.0, + "step": 10886 + }, + { + "epoch": 1.1955853283549307, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5831546783447266, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.701194167137146, + "num_tokens": 281793140.0, + "step": 10887 + }, + { + "epoch": 1.1956951460575445, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.478795051574707, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7311861515045166, + "num_tokens": 281814623.0, + "step": 10888 + }, + { + "epoch": 1.1958049637601582, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3573925495147705, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7171103954315186, + "num_tokens": 281838998.0, + "step": 10889 + }, + { + "epoch": 1.1959147814627717, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.078174114227295, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7311978340148926, + "num_tokens": 281867363.0, + "step": 10890 + }, + { + "epoch": 1.1960245991653855, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3933911323547363, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.725528359413147, + "num_tokens": 281889231.0, + "step": 10891 + }, + { + "epoch": 1.196134416867999, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1739156246185303, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7100196480751038, + "num_tokens": 281917109.0, + "step": 10892 + }, + { + "epoch": 1.1962442345706128, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3671064376831055, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7086929082870483, + "num_tokens": 281942238.0, + "step": 10893 + }, + { + "epoch": 1.1963540522732266, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.587589979171753, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7237048149108887, + "num_tokens": 281963200.0, + "step": 10894 + }, + { + "epoch": 1.19646386997584, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.187197685241699, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7060952186584473, + "num_tokens": 281991008.0, + "step": 10895 + }, + { + "epoch": 1.1965736876784538, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.516075372695923, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7347654104232788, + "num_tokens": 282013040.0, + "step": 10896 + }, + { + "epoch": 1.1966835053810674, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.482724666595459, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7276744842529297, + "num_tokens": 282033724.0, + "step": 10897 + }, + { + "epoch": 1.1967933230836811, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.045652389526367, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7018541097640991, + "num_tokens": 282064888.0, + "step": 10898 + }, + { + "epoch": 1.1969031407862947, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1917471885681152, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.720492959022522, + "num_tokens": 282091842.0, + "step": 10899 + }, + { + "epoch": 1.1970129584889084, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0613210201263428, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7300528883934021, + "num_tokens": 282121276.0, + "step": 10900 + }, + { + "epoch": 1.197122776191522, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.858175039291382, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7318587303161621, + "num_tokens": 282138436.0, + "step": 10901 + }, + { + "epoch": 1.1972325938941357, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.348299741744995, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7277597784996033, + "num_tokens": 282161974.0, + "step": 10902 + }, + { + "epoch": 1.1973424115967495, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.293161630630493, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.6962610483169556, + "num_tokens": 282187070.0, + "step": 10903 + }, + { + "epoch": 1.197452229299363, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5779201984405518, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7170503735542297, + "num_tokens": 282210391.0, + "step": 10904 + }, + { + "epoch": 1.1975620470019768, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1932284832000732, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.729179859161377, + "num_tokens": 282236268.0, + "step": 10905 + }, + { + "epoch": 1.1976718647045903, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0064873695373535, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6916759014129639, + "num_tokens": 282266725.0, + "step": 10906 + }, + { + "epoch": 1.197781682407204, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4886467456817627, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7498276233673096, + "num_tokens": 282286293.0, + "step": 10907 + }, + { + "epoch": 1.1978915001098178, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1234099864959717, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7034615874290466, + "num_tokens": 282317493.0, + "step": 10908 + }, + { + "epoch": 1.1980013178124314, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.431133270263672, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.6933515071868896, + "num_tokens": 282342822.0, + "step": 10909 + }, + { + "epoch": 1.1981111355150451, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.334348678588867, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.743183434009552, + "num_tokens": 282367191.0, + "step": 10910 + }, + { + "epoch": 1.1982209532176586, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.059229612350464, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.6959730386734009, + "num_tokens": 282397156.0, + "step": 10911 + }, + { + "epoch": 1.1983307709202724, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.335968494415283, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7297887802124023, + "num_tokens": 282422169.0, + "step": 10912 + }, + { + "epoch": 1.198440588622886, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.006697416305542, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7219635248184204, + "num_tokens": 282451878.0, + "step": 10913 + }, + { + "epoch": 1.1985504063254997, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2740771770477295, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7058408856391907, + "num_tokens": 282479141.0, + "step": 10914 + }, + { + "epoch": 1.1986602240281132, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0700080394744873, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7082341909408569, + "num_tokens": 282508341.0, + "step": 10915 + }, + { + "epoch": 1.198770041730727, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.212043523788452, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7278055548667908, + "num_tokens": 282533930.0, + "step": 10916 + }, + { + "epoch": 1.1988798594333407, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.339555263519287, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7211754322052002, + "num_tokens": 282556891.0, + "step": 10917 + }, + { + "epoch": 1.1989896771359543, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5468218326568604, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.724614679813385, + "num_tokens": 282578796.0, + "step": 10918 + }, + { + "epoch": 1.199099494838568, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3156397342681885, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.709132969379425, + "num_tokens": 282603968.0, + "step": 10919 + }, + { + "epoch": 1.1992093125411816, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.6904404163360596, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7237416505813599, + "num_tokens": 282623436.0, + "step": 10920 + }, + { + "epoch": 1.1993191302437953, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4175689220428467, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7263561487197876, + "num_tokens": 282644872.0, + "step": 10921 + }, + { + "epoch": 1.199428947946409, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0769200325012207, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6859269142150879, + "num_tokens": 282675436.0, + "step": 10922 + }, + { + "epoch": 1.1995387656490226, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4093387126922607, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7413814663887024, + "num_tokens": 282698546.0, + "step": 10923 + }, + { + "epoch": 1.1996485833516364, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3061630725860596, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7119377851486206, + "num_tokens": 282722349.0, + "step": 10924 + }, + { + "epoch": 1.19975840105425, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.472550630569458, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7020593881607056, + "num_tokens": 282745554.0, + "step": 10925 + }, + { + "epoch": 1.1998682187568637, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.398371934890747, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7228050231933594, + "num_tokens": 282770136.0, + "step": 10926 + }, + { + "epoch": 1.1999780364594772, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4323318004608154, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7132159471511841, + "num_tokens": 282793880.0, + "step": 10927 + }, + { + "epoch": 1.200087854162091, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.361426591873169, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7202587723731995, + "num_tokens": 282819050.0, + "step": 10928 + }, + { + "epoch": 1.2001976718647045, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.591566801071167, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7303277850151062, + "num_tokens": 282838645.0, + "step": 10929 + }, + { + "epoch": 1.2003074895673183, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 1.9738720655441284, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.713121771812439, + "num_tokens": 282870167.0, + "step": 10930 + }, + { + "epoch": 1.200417307269932, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1180834770202637, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7177963852882385, + "num_tokens": 282900252.0, + "step": 10931 + }, + { + "epoch": 1.2005271249725455, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2375171184539795, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.710835337638855, + "num_tokens": 282924793.0, + "step": 10932 + }, + { + "epoch": 1.2006369426751593, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1464595794677734, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7117042541503906, + "num_tokens": 282951368.0, + "step": 10933 + }, + { + "epoch": 1.2007467603777728, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.068601131439209, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6899926662445068, + "num_tokens": 282982310.0, + "step": 10934 + }, + { + "epoch": 1.2008565780803866, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4605584144592285, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.710614800453186, + "num_tokens": 283005262.0, + "step": 10935 + }, + { + "epoch": 1.2009663957830001, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.413459300994873, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7354545593261719, + "num_tokens": 283029265.0, + "step": 10936 + }, + { + "epoch": 1.2010762134856139, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.255211591720581, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7362477779388428, + "num_tokens": 283055446.0, + "step": 10937 + }, + { + "epoch": 1.2011860311882274, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3183138370513916, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7491518259048462, + "num_tokens": 283080794.0, + "step": 10938 + }, + { + "epoch": 1.2012958488908412, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4101853370666504, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7165729403495789, + "num_tokens": 283103477.0, + "step": 10939 + }, + { + "epoch": 1.201405666593455, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3437459468841553, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.7023249864578247, + "num_tokens": 283128921.0, + "step": 10940 + }, + { + "epoch": 1.2015154842960685, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 1.9593716859817505, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7139103412628174, + "num_tokens": 283159069.0, + "step": 10941 + }, + { + "epoch": 1.2016253019986822, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4388175010681152, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7196584939956665, + "num_tokens": 283182509.0, + "step": 10942 + }, + { + "epoch": 1.2017351197012958, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.16550874710083, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6925134658813477, + "num_tokens": 283211447.0, + "step": 10943 + }, + { + "epoch": 1.2018449374039095, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.157348394393921, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7223368883132935, + "num_tokens": 283239895.0, + "step": 10944 + }, + { + "epoch": 1.2019547551065233, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.841298818588257, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7120753526687622, + "num_tokens": 283260397.0, + "step": 10945 + }, + { + "epoch": 1.2020645728091368, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.296269655227661, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7334359288215637, + "num_tokens": 283285092.0, + "step": 10946 + }, + { + "epoch": 1.2021743905117506, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.412038564682007, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7305646538734436, + "num_tokens": 283308409.0, + "step": 10947 + }, + { + "epoch": 1.202284208214364, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2018728256225586, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7050650715827942, + "num_tokens": 283337028.0, + "step": 10948 + }, + { + "epoch": 1.2023940259169779, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.197399616241455, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.722209095954895, + "num_tokens": 283362535.0, + "step": 10949 + }, + { + "epoch": 1.2025038436195914, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.116401433944702, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6864648461341858, + "num_tokens": 283392480.0, + "step": 10950 + }, + { + "epoch": 1.2026136613222052, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4355270862579346, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7220150828361511, + "num_tokens": 283414290.0, + "step": 10951 + }, + { + "epoch": 1.2027234790248187, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5356688499450684, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7393811345100403, + "num_tokens": 283434515.0, + "step": 10952 + }, + { + "epoch": 1.2028332967274324, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.341928720474243, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7231848239898682, + "num_tokens": 283459691.0, + "step": 10953 + }, + { + "epoch": 1.2029431144300462, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2578859329223633, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7257739305496216, + "num_tokens": 283483621.0, + "step": 10954 + }, + { + "epoch": 1.2030529321326597, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.277411699295044, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7320549488067627, + "num_tokens": 283507753.0, + "step": 10955 + }, + { + "epoch": 1.2031627498352735, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3215715885162354, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.716118335723877, + "num_tokens": 283532007.0, + "step": 10956 + }, + { + "epoch": 1.203272567537887, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.421818494796753, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.7002103924751282, + "num_tokens": 283555781.0, + "step": 10957 + }, + { + "epoch": 1.2033823852405008, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.445514440536499, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7362022399902344, + "num_tokens": 283578730.0, + "step": 10958 + }, + { + "epoch": 1.2034922029431145, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2645838260650635, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7462325096130371, + "num_tokens": 283602621.0, + "step": 10959 + }, + { + "epoch": 1.203602020645728, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 1.8995558023452759, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7161524295806885, + "num_tokens": 283638322.0, + "step": 10960 + }, + { + "epoch": 1.2037118383483418, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.073270082473755, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.713919460773468, + "num_tokens": 283668495.0, + "step": 10961 + }, + { + "epoch": 1.2038216560509554, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.556933641433716, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.729656994342804, + "num_tokens": 283688969.0, + "step": 10962 + }, + { + "epoch": 1.2039314737535691, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2258782386779785, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.6999845504760742, + "num_tokens": 283717810.0, + "step": 10963 + }, + { + "epoch": 1.2040412914561827, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.485853672027588, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7063457369804382, + "num_tokens": 283740254.0, + "step": 10964 + }, + { + "epoch": 1.2041511091587964, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.590813159942627, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7242817878723145, + "num_tokens": 283760259.0, + "step": 10965 + }, + { + "epoch": 1.20426092686141, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.353088140487671, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7149885892868042, + "num_tokens": 283784178.0, + "step": 10966 + }, + { + "epoch": 1.2043707445640237, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3851068019866943, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7436328530311584, + "num_tokens": 283806168.0, + "step": 10967 + }, + { + "epoch": 1.2044805622666375, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3488552570343018, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7277538180351257, + "num_tokens": 283830981.0, + "step": 10968 + }, + { + "epoch": 1.204590379969251, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.223416805267334, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7134482264518738, + "num_tokens": 283856809.0, + "step": 10969 + }, + { + "epoch": 1.2047001976718648, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4330477714538574, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.720212459564209, + "num_tokens": 283880097.0, + "step": 10970 + }, + { + "epoch": 1.2048100153744783, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3144993782043457, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7315224409103394, + "num_tokens": 283904831.0, + "step": 10971 + }, + { + "epoch": 1.204919833077092, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2284023761749268, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.7091602683067322, + "num_tokens": 283931547.0, + "step": 10972 + }, + { + "epoch": 1.2050296507797058, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.615164279937744, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7256372570991516, + "num_tokens": 283950698.0, + "step": 10973 + }, + { + "epoch": 1.2051394684823193, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5615170001983643, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7253831028938293, + "num_tokens": 283974470.0, + "step": 10974 + }, + { + "epoch": 1.205249286184933, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3695573806762695, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7054802179336548, + "num_tokens": 284001087.0, + "step": 10975 + }, + { + "epoch": 1.2053591038875466, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.674067258834839, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7391027808189392, + "num_tokens": 284020322.0, + "step": 10976 + }, + { + "epoch": 1.2054689215901604, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.585972785949707, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.6998226046562195, + "num_tokens": 284040858.0, + "step": 10977 + }, + { + "epoch": 1.205578739292774, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1961684226989746, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.701607346534729, + "num_tokens": 284067636.0, + "step": 10978 + }, + { + "epoch": 1.2056885569953877, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.123971939086914, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.7032520771026611, + "num_tokens": 284096336.0, + "step": 10979 + }, + { + "epoch": 1.2057983746980012, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3987879753112793, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7086099982261658, + "num_tokens": 284120413.0, + "step": 10980 + }, + { + "epoch": 1.205908192400615, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.170529365539551, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7241465449333191, + "num_tokens": 284145303.0, + "step": 10981 + }, + { + "epoch": 1.2060180101032287, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2873685359954834, + "learning_rate": 1e-06, + "loss": 0.755, + "mean_token_accuracy": 0.7591877579689026, + "num_tokens": 284167398.0, + "step": 10982 + }, + { + "epoch": 1.2061278278058423, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3431191444396973, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.723156213760376, + "num_tokens": 284191959.0, + "step": 10983 + }, + { + "epoch": 1.206237645508456, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2649898529052734, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7170759439468384, + "num_tokens": 284217135.0, + "step": 10984 + }, + { + "epoch": 1.2063474632110696, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 1.9302541017532349, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6942053437232971, + "num_tokens": 284252636.0, + "step": 10985 + }, + { + "epoch": 1.2064572809136833, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2329137325286865, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7283577919006348, + "num_tokens": 284277462.0, + "step": 10986 + }, + { + "epoch": 1.2065670986162969, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4881343841552734, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7485681772232056, + "num_tokens": 284298181.0, + "step": 10987 + }, + { + "epoch": 1.2066769163189106, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.124053955078125, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7113360166549683, + "num_tokens": 284326972.0, + "step": 10988 + }, + { + "epoch": 1.2067867340215244, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2967634201049805, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7286049127578735, + "num_tokens": 284352298.0, + "step": 10989 + }, + { + "epoch": 1.206896551724138, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.6518685817718506, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7204854488372803, + "num_tokens": 284371893.0, + "step": 10990 + }, + { + "epoch": 1.2070063694267517, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.263962745666504, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7065598964691162, + "num_tokens": 284398281.0, + "step": 10991 + }, + { + "epoch": 1.2071161871293652, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4546093940734863, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7171111106872559, + "num_tokens": 284421089.0, + "step": 10992 + }, + { + "epoch": 1.207226004831979, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.629526138305664, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.732934832572937, + "num_tokens": 284441192.0, + "step": 10993 + }, + { + "epoch": 1.2073358225345925, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.123648166656494, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6965920925140381, + "num_tokens": 284471296.0, + "step": 10994 + }, + { + "epoch": 1.2074456402372062, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.195971727371216, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7088049054145813, + "num_tokens": 284497238.0, + "step": 10995 + }, + { + "epoch": 1.20755545793982, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1628506183624268, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.70163494348526, + "num_tokens": 284525660.0, + "step": 10996 + }, + { + "epoch": 1.2076652756424335, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.129971742630005, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7235639095306396, + "num_tokens": 284554370.0, + "step": 10997 + }, + { + "epoch": 1.2077750933450473, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.372114658355713, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7272335290908813, + "num_tokens": 284577055.0, + "step": 10998 + }, + { + "epoch": 1.2078849110476608, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.3481528759002686, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7330291867256165, + "num_tokens": 284601534.0, + "step": 10999 + }, + { + "epoch": 1.2079947287502746, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4810991287231445, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7190505266189575, + "num_tokens": 284624533.0, + "step": 11000 + }, + { + "epoch": 1.2081045464528881, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0153350830078125, + "learning_rate": 1e-06, + "loss": 1.0415, + "mean_token_accuracy": 0.67954421043396, + "num_tokens": 284655548.0, + "step": 11001 + }, + { + "epoch": 1.2082143641555019, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2878665924072266, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7169244289398193, + "num_tokens": 284678884.0, + "step": 11002 + }, + { + "epoch": 1.2083241818581154, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.4044227600097656, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7154995203018188, + "num_tokens": 284703846.0, + "step": 11003 + }, + { + "epoch": 1.2084339995607292, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.12119722366333, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.723139226436615, + "num_tokens": 284729345.0, + "step": 11004 + }, + { + "epoch": 1.208543817263343, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.0290162563323975, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7194379568099976, + "num_tokens": 284759880.0, + "step": 11005 + }, + { + "epoch": 1.2086536349659565, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.6739635467529297, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7386482357978821, + "num_tokens": 284778553.0, + "step": 11006 + }, + { + "epoch": 1.2087634526685702, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1854207515716553, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7355524301528931, + "num_tokens": 284803349.0, + "step": 11007 + }, + { + "epoch": 1.2088732703711838, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.1107177734375, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6929025650024414, + "num_tokens": 284836162.0, + "step": 11008 + }, + { + "epoch": 1.2089830880737975, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5526700019836426, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7288157939910889, + "num_tokens": 284855788.0, + "step": 11009 + }, + { + "epoch": 1.2090929057764113, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.370806932449341, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7188565731048584, + "num_tokens": 284880287.0, + "step": 11010 + }, + { + "epoch": 1.2092027234790248, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.273397445678711, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.727196216583252, + "num_tokens": 284904543.0, + "step": 11011 + }, + { + "epoch": 1.2093125411816386, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.137739658355713, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7109494805335999, + "num_tokens": 284933876.0, + "step": 11012 + }, + { + "epoch": 1.209422358884252, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2406325340270996, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7101456522941589, + "num_tokens": 284961257.0, + "step": 11013 + }, + { + "epoch": 1.2095321765868658, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.2185678482055664, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6951420903205872, + "num_tokens": 284988656.0, + "step": 11014 + }, + { + "epoch": 1.2096419942894794, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.5510993003845215, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7321116328239441, + "num_tokens": 285009556.0, + "step": 11015 + }, + { + "epoch": 1.2097518119920931, + "ewc_loss": 0.00010061264038085938, + "grad_norm": 2.27360463142395, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7389259934425354, + "num_tokens": 285034371.0, + "step": 11016 + }, + { + "epoch": 1.2098616296947067, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3666632175445557, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7264280915260315, + "num_tokens": 285056245.0, + "step": 11017 + }, + { + "epoch": 1.2099714473973204, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2560665607452393, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7192853093147278, + "num_tokens": 285081384.0, + "step": 11018 + }, + { + "epoch": 1.2100812650999342, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.5634255409240723, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7260094881057739, + "num_tokens": 285101546.0, + "step": 11019 + }, + { + "epoch": 1.2101910828025477, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1181559562683105, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7352309226989746, + "num_tokens": 285128909.0, + "step": 11020 + }, + { + "epoch": 1.2103009005051615, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.656468152999878, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7145204544067383, + "num_tokens": 285149217.0, + "step": 11021 + }, + { + "epoch": 1.210410718207775, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.276137351989746, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7419005632400513, + "num_tokens": 285173057.0, + "step": 11022 + }, + { + "epoch": 1.2105205359103888, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2789580821990967, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7291309237480164, + "num_tokens": 285198193.0, + "step": 11023 + }, + { + "epoch": 1.2106303536130025, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1962900161743164, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.710444450378418, + "num_tokens": 285226208.0, + "step": 11024 + }, + { + "epoch": 1.210740171315616, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3554203510284424, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7165395021438599, + "num_tokens": 285249441.0, + "step": 11025 + }, + { + "epoch": 1.2108499890182298, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.212048053741455, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7440064549446106, + "num_tokens": 285274785.0, + "step": 11026 + }, + { + "epoch": 1.2109598067208434, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 1.9438917636871338, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.6917933225631714, + "num_tokens": 285311003.0, + "step": 11027 + }, + { + "epoch": 1.2110696244234571, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3684816360473633, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7249252796173096, + "num_tokens": 285334538.0, + "step": 11028 + }, + { + "epoch": 1.2111794421260706, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1851541996002197, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7184208631515503, + "num_tokens": 285361596.0, + "step": 11029 + }, + { + "epoch": 1.2112892598286844, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2005786895751953, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7058336734771729, + "num_tokens": 285389838.0, + "step": 11030 + }, + { + "epoch": 1.211399077531298, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.135829448699951, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6987738013267517, + "num_tokens": 285417739.0, + "step": 11031 + }, + { + "epoch": 1.2115088952339117, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.855638265609741, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7311816215515137, + "num_tokens": 285435299.0, + "step": 11032 + }, + { + "epoch": 1.2116187129365255, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.389742851257324, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.714566171169281, + "num_tokens": 285460233.0, + "step": 11033 + }, + { + "epoch": 1.211728530639139, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4291861057281494, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7130927443504333, + "num_tokens": 285485153.0, + "step": 11034 + }, + { + "epoch": 1.2118383483417527, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2310869693756104, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7309113144874573, + "num_tokens": 285510008.0, + "step": 11035 + }, + { + "epoch": 1.2119481660443663, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2146904468536377, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7145459651947021, + "num_tokens": 285535003.0, + "step": 11036 + }, + { + "epoch": 1.21205798374698, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0588390827178955, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.710810124874115, + "num_tokens": 285565024.0, + "step": 11037 + }, + { + "epoch": 1.2121678014495938, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 1.9121423959732056, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7235784530639648, + "num_tokens": 285598590.0, + "step": 11038 + }, + { + "epoch": 1.2122776191522073, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4738807678222656, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7237573266029358, + "num_tokens": 285619570.0, + "step": 11039 + }, + { + "epoch": 1.212387436854821, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.300215482711792, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6991040706634521, + "num_tokens": 285644388.0, + "step": 11040 + }, + { + "epoch": 1.2124972545574346, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0833606719970703, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6945726871490479, + "num_tokens": 285676890.0, + "step": 11041 + }, + { + "epoch": 1.2126070722600484, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.6536812782287598, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7388740181922913, + "num_tokens": 285694848.0, + "step": 11042 + }, + { + "epoch": 1.212716889962662, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.6289877891540527, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7269362807273865, + "num_tokens": 285713171.0, + "step": 11043 + }, + { + "epoch": 1.2128267076652757, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2698609828948975, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7518184185028076, + "num_tokens": 285736481.0, + "step": 11044 + }, + { + "epoch": 1.2129365253678892, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0536792278289795, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.6916102766990662, + "num_tokens": 285766170.0, + "step": 11045 + }, + { + "epoch": 1.213046343070503, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.975290060043335, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7413281798362732, + "num_tokens": 285782962.0, + "step": 11046 + }, + { + "epoch": 1.2131561607731167, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4340779781341553, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7367947101593018, + "num_tokens": 285805223.0, + "step": 11047 + }, + { + "epoch": 1.2132659784757303, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.327906847000122, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7218915224075317, + "num_tokens": 285830478.0, + "step": 11048 + }, + { + "epoch": 1.213375796178344, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.799055576324463, + "learning_rate": 1e-06, + "loss": 0.7685, + "mean_token_accuracy": 0.7556664347648621, + "num_tokens": 285848620.0, + "step": 11049 + }, + { + "epoch": 1.2134856138809575, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1902740001678467, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7082661986351013, + "num_tokens": 285876089.0, + "step": 11050 + }, + { + "epoch": 1.2135954315835713, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3042900562286377, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7428346872329712, + "num_tokens": 285900376.0, + "step": 11051 + }, + { + "epoch": 1.2137052492861848, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.283405065536499, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7349214553833008, + "num_tokens": 285923703.0, + "step": 11052 + }, + { + "epoch": 1.2138150669887986, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3211829662323, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7004704475402832, + "num_tokens": 285950717.0, + "step": 11053 + }, + { + "epoch": 1.2139248846914121, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4017269611358643, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7040557861328125, + "num_tokens": 285973567.0, + "step": 11054 + }, + { + "epoch": 1.2140347023940259, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3893392086029053, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7087677717208862, + "num_tokens": 285997017.0, + "step": 11055 + }, + { + "epoch": 1.2141445200966396, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1216483116149902, + "learning_rate": 1e-06, + "loss": 1.0543, + "mean_token_accuracy": 0.6828006505966187, + "num_tokens": 286028655.0, + "step": 11056 + }, + { + "epoch": 1.2142543377992532, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.265514612197876, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7126481533050537, + "num_tokens": 286054292.0, + "step": 11057 + }, + { + "epoch": 1.214364155501867, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2384040355682373, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7279760837554932, + "num_tokens": 286079944.0, + "step": 11058 + }, + { + "epoch": 1.2144739732044805, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.199960470199585, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.6986081600189209, + "num_tokens": 286108619.0, + "step": 11059 + }, + { + "epoch": 1.2145837909070942, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0700645446777344, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7308913469314575, + "num_tokens": 286137641.0, + "step": 11060 + }, + { + "epoch": 1.214693608609708, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.163814067840576, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.720482587814331, + "num_tokens": 286163781.0, + "step": 11061 + }, + { + "epoch": 1.2148034263123215, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0989878177642822, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7085237503051758, + "num_tokens": 286193706.0, + "step": 11062 + }, + { + "epoch": 1.2149132440149353, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0654749870300293, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7063280940055847, + "num_tokens": 286224522.0, + "step": 11063 + }, + { + "epoch": 1.2150230617175488, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.175513744354248, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7336267232894897, + "num_tokens": 286249405.0, + "step": 11064 + }, + { + "epoch": 1.2151328794201626, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1213231086730957, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7225544452667236, + "num_tokens": 286278531.0, + "step": 11065 + }, + { + "epoch": 1.215242697122776, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 1.976219892501831, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7126365900039673, + "num_tokens": 286308331.0, + "step": 11066 + }, + { + "epoch": 1.2153525148253899, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0109589099884033, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.704807698726654, + "num_tokens": 286339208.0, + "step": 11067 + }, + { + "epoch": 1.2154623325280034, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3429319858551025, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7329933047294617, + "num_tokens": 286361971.0, + "step": 11068 + }, + { + "epoch": 1.2155721502306172, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 1.9681446552276611, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7195594310760498, + "num_tokens": 286392775.0, + "step": 11069 + }, + { + "epoch": 1.215681967933231, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.642526626586914, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7283416986465454, + "num_tokens": 286411263.0, + "step": 11070 + }, + { + "epoch": 1.2157917856358444, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1838088035583496, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7132079601287842, + "num_tokens": 286437434.0, + "step": 11071 + }, + { + "epoch": 1.2159016033384582, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.258739471435547, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7235704660415649, + "num_tokens": 286461643.0, + "step": 11072 + }, + { + "epoch": 1.2160114210410717, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.049302577972412, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7073943614959717, + "num_tokens": 286493286.0, + "step": 11073 + }, + { + "epoch": 1.2161212387436855, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.5303139686584473, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7194210886955261, + "num_tokens": 286514586.0, + "step": 11074 + }, + { + "epoch": 1.2162310564462993, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.6386775970458984, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7378715872764587, + "num_tokens": 286533142.0, + "step": 11075 + }, + { + "epoch": 1.2163408741489128, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.222350597381592, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7162674069404602, + "num_tokens": 286560108.0, + "step": 11076 + }, + { + "epoch": 1.2164506918515265, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 1.9751843214035034, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6894416213035583, + "num_tokens": 286592101.0, + "step": 11077 + }, + { + "epoch": 1.21656050955414, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.320651054382324, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7129650115966797, + "num_tokens": 286614678.0, + "step": 11078 + }, + { + "epoch": 1.2166703272567538, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.559621810913086, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7108904719352722, + "num_tokens": 286636211.0, + "step": 11079 + }, + { + "epoch": 1.2167801449593674, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2130112648010254, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7205524444580078, + "num_tokens": 286662911.0, + "step": 11080 + }, + { + "epoch": 1.2168899626619811, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.137735366821289, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7200525403022766, + "num_tokens": 286690482.0, + "step": 11081 + }, + { + "epoch": 1.2169997803645947, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.282015323638916, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7031663656234741, + "num_tokens": 286716439.0, + "step": 11082 + }, + { + "epoch": 1.2171095980672084, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.0331108570098877, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7265623807907104, + "num_tokens": 286745981.0, + "step": 11083 + }, + { + "epoch": 1.2172194157698222, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.131685256958008, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7155132293701172, + "num_tokens": 286774086.0, + "step": 11084 + }, + { + "epoch": 1.2173292334724357, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3378217220306396, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.721644401550293, + "num_tokens": 286797551.0, + "step": 11085 + }, + { + "epoch": 1.2174390511750495, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4495906829833984, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7171108722686768, + "num_tokens": 286821481.0, + "step": 11086 + }, + { + "epoch": 1.217548868877663, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.080326557159424, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7011753916740417, + "num_tokens": 286851130.0, + "step": 11087 + }, + { + "epoch": 1.2176586865802768, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.616360664367676, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7305577993392944, + "num_tokens": 286870940.0, + "step": 11088 + }, + { + "epoch": 1.2177685042828905, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.4141783714294434, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7306344509124756, + "num_tokens": 286893884.0, + "step": 11089 + }, + { + "epoch": 1.217878321985504, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.43335223197937, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7376012206077576, + "num_tokens": 286916003.0, + "step": 11090 + }, + { + "epoch": 1.2179881396881178, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.521409511566162, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7079665064811707, + "num_tokens": 286938420.0, + "step": 11091 + }, + { + "epoch": 1.2180979573907313, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.725102424621582, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7353350520133972, + "num_tokens": 286958223.0, + "step": 11092 + }, + { + "epoch": 1.218207775093345, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.2805583477020264, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7198542356491089, + "num_tokens": 286982199.0, + "step": 11093 + }, + { + "epoch": 1.2183175927959586, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.183753490447998, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7168695330619812, + "num_tokens": 287010430.0, + "step": 11094 + }, + { + "epoch": 1.2184274104985724, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.609437942504883, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7267140746116638, + "num_tokens": 287030970.0, + "step": 11095 + }, + { + "epoch": 1.218537228201186, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.1328985691070557, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7052872180938721, + "num_tokens": 287057132.0, + "step": 11096 + }, + { + "epoch": 1.2186470459037997, + "ewc_loss": 0.0001010894775390625, + "grad_norm": 2.3092057704925537, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7315424680709839, + "num_tokens": 287082012.0, + "step": 11097 + }, + { + "epoch": 1.2187568636064134, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2695913314819336, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.7017210721969604, + "num_tokens": 287109465.0, + "step": 11098 + }, + { + "epoch": 1.218866681309027, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1782500743865967, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7479239702224731, + "num_tokens": 287135634.0, + "step": 11099 + }, + { + "epoch": 1.2189764990116407, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.832254409790039, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7375009059906006, + "num_tokens": 287151167.0, + "step": 11100 + }, + { + "epoch": 1.2190863167142543, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3126728534698486, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.738979697227478, + "num_tokens": 287176289.0, + "step": 11101 + }, + { + "epoch": 1.219196134416868, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4097092151641846, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7117104530334473, + "num_tokens": 287199749.0, + "step": 11102 + }, + { + "epoch": 1.2193059521194818, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1204993724823, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.704248309135437, + "num_tokens": 287229177.0, + "step": 11103 + }, + { + "epoch": 1.2194157698220953, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1275384426116943, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.709488034248352, + "num_tokens": 287257362.0, + "step": 11104 + }, + { + "epoch": 1.219525587524709, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0965611934661865, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7230752110481262, + "num_tokens": 287286351.0, + "step": 11105 + }, + { + "epoch": 1.2196354052273226, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3679490089416504, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7214945554733276, + "num_tokens": 287312114.0, + "step": 11106 + }, + { + "epoch": 1.2197452229299364, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.151878595352173, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7256299257278442, + "num_tokens": 287340424.0, + "step": 11107 + }, + { + "epoch": 1.21985504063255, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3009703159332275, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.7060188055038452, + "num_tokens": 287365830.0, + "step": 11108 + }, + { + "epoch": 1.2199648583351637, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1619858741760254, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7330101728439331, + "num_tokens": 287392435.0, + "step": 11109 + }, + { + "epoch": 1.2200746760377772, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.652949810028076, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7300106287002563, + "num_tokens": 287412571.0, + "step": 11110 + }, + { + "epoch": 1.220184493740391, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0333688259124756, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7165891528129578, + "num_tokens": 287441777.0, + "step": 11111 + }, + { + "epoch": 1.2202943114430047, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5917203426361084, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7144830822944641, + "num_tokens": 287462481.0, + "step": 11112 + }, + { + "epoch": 1.2204041291456182, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.212139844894409, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7155043482780457, + "num_tokens": 287491128.0, + "step": 11113 + }, + { + "epoch": 1.220513946848232, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.8533544540405273, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.72397780418396, + "num_tokens": 287524819.0, + "step": 11114 + }, + { + "epoch": 1.2206237645508455, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.8552851676940918, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7156205773353577, + "num_tokens": 287562937.0, + "step": 11115 + }, + { + "epoch": 1.2207335822534593, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.405620813369751, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7402117252349854, + "num_tokens": 287584404.0, + "step": 11116 + }, + { + "epoch": 1.2208433999560728, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.188432455062866, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7443634867668152, + "num_tokens": 287609866.0, + "step": 11117 + }, + { + "epoch": 1.2209532176586866, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0529050827026367, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7062203884124756, + "num_tokens": 287639847.0, + "step": 11118 + }, + { + "epoch": 1.2210630353613001, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.9635670185089111, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7187556028366089, + "num_tokens": 287670981.0, + "step": 11119 + }, + { + "epoch": 1.2211728530639139, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0673155784606934, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7102065086364746, + "num_tokens": 287701178.0, + "step": 11120 + }, + { + "epoch": 1.2212826707665276, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.190962553024292, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7023220062255859, + "num_tokens": 287728371.0, + "step": 11121 + }, + { + "epoch": 1.2213924884691412, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.329277992248535, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7199186682701111, + "num_tokens": 287752359.0, + "step": 11122 + }, + { + "epoch": 1.221502306171755, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4482948780059814, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7313876748085022, + "num_tokens": 287774074.0, + "step": 11123 + }, + { + "epoch": 1.2216121238743685, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.022371292114258, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6988165378570557, + "num_tokens": 287805557.0, + "step": 11124 + }, + { + "epoch": 1.2217219415769822, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2834341526031494, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7035907506942749, + "num_tokens": 287831463.0, + "step": 11125 + }, + { + "epoch": 1.221831759279596, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4701266288757324, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7428969144821167, + "num_tokens": 287852361.0, + "step": 11126 + }, + { + "epoch": 1.2219415769822095, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5834786891937256, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7246440649032593, + "num_tokens": 287871387.0, + "step": 11127 + }, + { + "epoch": 1.2220513946848233, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3751919269561768, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7174752950668335, + "num_tokens": 287897123.0, + "step": 11128 + }, + { + "epoch": 1.2221612123874368, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5478768348693848, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7172638177871704, + "num_tokens": 287920282.0, + "step": 11129 + }, + { + "epoch": 1.2222710300900506, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3767623901367188, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7139912843704224, + "num_tokens": 287945621.0, + "step": 11130 + }, + { + "epoch": 1.222380847792664, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1776649951934814, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7314115762710571, + "num_tokens": 287972594.0, + "step": 11131 + }, + { + "epoch": 1.2224906654952779, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.9465129375457764, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7038373947143555, + "num_tokens": 288005889.0, + "step": 11132 + }, + { + "epoch": 1.2226004831978914, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3781232833862305, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7322871685028076, + "num_tokens": 288026620.0, + "step": 11133 + }, + { + "epoch": 1.2227103009005051, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1879918575286865, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7135261297225952, + "num_tokens": 288055937.0, + "step": 11134 + }, + { + "epoch": 1.222820118603119, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2417237758636475, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7221202254295349, + "num_tokens": 288082276.0, + "step": 11135 + }, + { + "epoch": 1.2229299363057324, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4777281284332275, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7155582904815674, + "num_tokens": 288105292.0, + "step": 11136 + }, + { + "epoch": 1.2230397540083462, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1824166774749756, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7252616286277771, + "num_tokens": 288131219.0, + "step": 11137 + }, + { + "epoch": 1.2231495717109597, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0181186199188232, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7117756009101868, + "num_tokens": 288164176.0, + "step": 11138 + }, + { + "epoch": 1.2232593894135735, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.402193784713745, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.717829704284668, + "num_tokens": 288188719.0, + "step": 11139 + }, + { + "epoch": 1.2233692071161872, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2138218879699707, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7088854908943176, + "num_tokens": 288216109.0, + "step": 11140 + }, + { + "epoch": 1.2234790248188008, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4212989807128906, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7148140668869019, + "num_tokens": 288240464.0, + "step": 11141 + }, + { + "epoch": 1.2235888425214145, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1486740112304688, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7268305420875549, + "num_tokens": 288266942.0, + "step": 11142 + }, + { + "epoch": 1.223698660224028, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2468676567077637, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7025051116943359, + "num_tokens": 288295463.0, + "step": 11143 + }, + { + "epoch": 1.2238084779266418, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2438807487487793, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7286767363548279, + "num_tokens": 288319423.0, + "step": 11144 + }, + { + "epoch": 1.2239182956292554, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.7530667781829834, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.736220121383667, + "num_tokens": 288338142.0, + "step": 11145 + }, + { + "epoch": 1.2240281133318691, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0324106216430664, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7322534322738647, + "num_tokens": 288365441.0, + "step": 11146 + }, + { + "epoch": 1.2241379310344827, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3798112869262695, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7008923292160034, + "num_tokens": 288391090.0, + "step": 11147 + }, + { + "epoch": 1.2242477487370964, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.9897761344909668, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6942875385284424, + "num_tokens": 288423660.0, + "step": 11148 + }, + { + "epoch": 1.2243575664397102, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1562247276306152, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7214986085891724, + "num_tokens": 288452342.0, + "step": 11149 + }, + { + "epoch": 1.2244673841423237, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.167320966720581, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7098562717437744, + "num_tokens": 288480932.0, + "step": 11150 + }, + { + "epoch": 1.2245772018449375, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.7612879276275635, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7467286586761475, + "num_tokens": 288498077.0, + "step": 11151 + }, + { + "epoch": 1.224687019547551, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3500962257385254, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.706271767616272, + "num_tokens": 288524421.0, + "step": 11152 + }, + { + "epoch": 1.2247968372501647, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.180593252182007, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.723045825958252, + "num_tokens": 288551768.0, + "step": 11153 + }, + { + "epoch": 1.2249066549527785, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3535895347595215, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.725783109664917, + "num_tokens": 288575893.0, + "step": 11154 + }, + { + "epoch": 1.225016472655392, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3599629402160645, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.725763201713562, + "num_tokens": 288598787.0, + "step": 11155 + }, + { + "epoch": 1.2251262903580058, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5122694969177246, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7165008783340454, + "num_tokens": 288620878.0, + "step": 11156 + }, + { + "epoch": 1.2252361080606193, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1269078254699707, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7215104103088379, + "num_tokens": 288648869.0, + "step": 11157 + }, + { + "epoch": 1.225345925763233, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5364391803741455, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7469924688339233, + "num_tokens": 288668783.0, + "step": 11158 + }, + { + "epoch": 1.2254557434658466, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.211665630340576, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7205966711044312, + "num_tokens": 288696825.0, + "step": 11159 + }, + { + "epoch": 1.2255655611684604, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.9855332374572754, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7239202260971069, + "num_tokens": 288729299.0, + "step": 11160 + }, + { + "epoch": 1.225675378871074, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3827757835388184, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7236341238021851, + "num_tokens": 288754789.0, + "step": 11161 + }, + { + "epoch": 1.2257851965736877, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3310365676879883, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7018715143203735, + "num_tokens": 288780435.0, + "step": 11162 + }, + { + "epoch": 1.2258950142763014, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3770313262939453, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7120962142944336, + "num_tokens": 288803995.0, + "step": 11163 + }, + { + "epoch": 1.226004831978915, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0902252197265625, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7036421895027161, + "num_tokens": 288832003.0, + "step": 11164 + }, + { + "epoch": 1.2261146496815287, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1159815788269043, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7068294286727905, + "num_tokens": 288862135.0, + "step": 11165 + }, + { + "epoch": 1.2262244673841423, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1327102184295654, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7024440169334412, + "num_tokens": 288890024.0, + "step": 11166 + }, + { + "epoch": 1.226334285086756, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.308392286300659, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7194544672966003, + "num_tokens": 288913336.0, + "step": 11167 + }, + { + "epoch": 1.2264441027893696, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.06416392326355, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.718457818031311, + "num_tokens": 288944772.0, + "step": 11168 + }, + { + "epoch": 1.2265539204919833, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1139652729034424, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7185549736022949, + "num_tokens": 288972703.0, + "step": 11169 + }, + { + "epoch": 1.226663738194597, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2002999782562256, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.740604817867279, + "num_tokens": 288997926.0, + "step": 11170 + }, + { + "epoch": 1.2267735558972106, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.394760847091675, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7021112442016602, + "num_tokens": 289020373.0, + "step": 11171 + }, + { + "epoch": 1.2268833735998244, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2420921325683594, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6972939968109131, + "num_tokens": 289048954.0, + "step": 11172 + }, + { + "epoch": 1.226993191302438, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3589115142822266, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7201869487762451, + "num_tokens": 289073845.0, + "step": 11173 + }, + { + "epoch": 1.2271030090050516, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.498461961746216, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7272028923034668, + "num_tokens": 289094320.0, + "step": 11174 + }, + { + "epoch": 1.2272128267076652, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2686338424682617, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7412089705467224, + "num_tokens": 289123281.0, + "step": 11175 + }, + { + "epoch": 1.227322644410279, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5139858722686768, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7075847387313843, + "num_tokens": 289146395.0, + "step": 11176 + }, + { + "epoch": 1.2274324621128927, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5422801971435547, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7217262387275696, + "num_tokens": 289166789.0, + "step": 11177 + }, + { + "epoch": 1.2275422798155062, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3239965438842773, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.7037028074264526, + "num_tokens": 289190811.0, + "step": 11178 + }, + { + "epoch": 1.22765209751812, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2980144023895264, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.721606969833374, + "num_tokens": 289216576.0, + "step": 11179 + }, + { + "epoch": 1.2277619152207335, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3285443782806396, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7209545373916626, + "num_tokens": 289241926.0, + "step": 11180 + }, + { + "epoch": 1.2278717329233473, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5108771324157715, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7442605495452881, + "num_tokens": 289261528.0, + "step": 11181 + }, + { + "epoch": 1.2279815506259608, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.520765781402588, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6923245191574097, + "num_tokens": 289284892.0, + "step": 11182 + }, + { + "epoch": 1.2280913683285746, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2078804969787598, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7087907791137695, + "num_tokens": 289311750.0, + "step": 11183 + }, + { + "epoch": 1.228201186031188, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.297391653060913, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7172092199325562, + "num_tokens": 289336750.0, + "step": 11184 + }, + { + "epoch": 1.2283110037338019, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.894907832145691, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.706436812877655, + "num_tokens": 289372893.0, + "step": 11185 + }, + { + "epoch": 1.2284208214364156, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.117382526397705, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7064964771270752, + "num_tokens": 289400788.0, + "step": 11186 + }, + { + "epoch": 1.2285306391390292, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2421252727508545, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7159151434898376, + "num_tokens": 289428204.0, + "step": 11187 + }, + { + "epoch": 1.228640456841643, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5460526943206787, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7385497093200684, + "num_tokens": 289448773.0, + "step": 11188 + }, + { + "epoch": 1.2287502745442564, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3876686096191406, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7265542149543762, + "num_tokens": 289472201.0, + "step": 11189 + }, + { + "epoch": 1.2288600922468702, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3356590270996094, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7405558824539185, + "num_tokens": 289493334.0, + "step": 11190 + }, + { + "epoch": 1.228969909949484, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1912879943847656, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6999869346618652, + "num_tokens": 289522313.0, + "step": 11191 + }, + { + "epoch": 1.2290797276520975, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 1.9419173002243042, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6885611414909363, + "num_tokens": 289555995.0, + "step": 11192 + }, + { + "epoch": 1.2291895453547113, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.0494863986968994, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7058385014533997, + "num_tokens": 289588907.0, + "step": 11193 + }, + { + "epoch": 1.2292993630573248, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3045785427093506, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7156952619552612, + "num_tokens": 289613880.0, + "step": 11194 + }, + { + "epoch": 1.2294091807599385, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4718661308288574, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7237520217895508, + "num_tokens": 289635850.0, + "step": 11195 + }, + { + "epoch": 1.229518998462552, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1705009937286377, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.719017744064331, + "num_tokens": 289661786.0, + "step": 11196 + }, + { + "epoch": 1.2296288161651658, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.152029037475586, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7290099859237671, + "num_tokens": 289690312.0, + "step": 11197 + }, + { + "epoch": 1.2297386338677794, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3223626613616943, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7072350382804871, + "num_tokens": 289716247.0, + "step": 11198 + }, + { + "epoch": 1.2298484515703931, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.497565269470215, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7053117156028748, + "num_tokens": 289737054.0, + "step": 11199 + }, + { + "epoch": 1.2299582692730069, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.245603084564209, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.69187331199646, + "num_tokens": 289765541.0, + "step": 11200 + }, + { + "epoch": 1.2300680869756204, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.223041296005249, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7194794416427612, + "num_tokens": 289793721.0, + "step": 11201 + }, + { + "epoch": 1.2301779046782342, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.176654577255249, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7253122329711914, + "num_tokens": 289818886.0, + "step": 11202 + }, + { + "epoch": 1.2302877223808477, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.5584230422973633, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7242621183395386, + "num_tokens": 289839342.0, + "step": 11203 + }, + { + "epoch": 1.2303975400834615, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.49685001373291, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7095016837120056, + "num_tokens": 289859828.0, + "step": 11204 + }, + { + "epoch": 1.2305073577860752, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.615513801574707, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7398497462272644, + "num_tokens": 289878027.0, + "step": 11205 + }, + { + "epoch": 1.2306171754886888, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.317791223526001, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7021478414535522, + "num_tokens": 289902423.0, + "step": 11206 + }, + { + "epoch": 1.2307269931913025, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3634867668151855, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7061917781829834, + "num_tokens": 289926283.0, + "step": 11207 + }, + { + "epoch": 1.230836810893916, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.481171131134033, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7254024147987366, + "num_tokens": 289947044.0, + "step": 11208 + }, + { + "epoch": 1.2309466285965298, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2104063034057617, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7066367864608765, + "num_tokens": 289973967.0, + "step": 11209 + }, + { + "epoch": 1.2310564462991433, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.457277536392212, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7179098129272461, + "num_tokens": 289997008.0, + "step": 11210 + }, + { + "epoch": 1.231166264001757, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2355852127075195, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6976157426834106, + "num_tokens": 290025815.0, + "step": 11211 + }, + { + "epoch": 1.2312760817043706, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.363187313079834, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.709248423576355, + "num_tokens": 290049001.0, + "step": 11212 + }, + { + "epoch": 1.2313858994069844, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.439581871032715, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7176016569137573, + "num_tokens": 290072534.0, + "step": 11213 + }, + { + "epoch": 1.2314957171095982, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4310076236724854, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6991336345672607, + "num_tokens": 290098182.0, + "step": 11214 + }, + { + "epoch": 1.2316055348122117, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.435404062271118, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7273038625717163, + "num_tokens": 290121546.0, + "step": 11215 + }, + { + "epoch": 1.2317153525148254, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3443174362182617, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7406372427940369, + "num_tokens": 290144929.0, + "step": 11216 + }, + { + "epoch": 1.231825170217439, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.3332412242889404, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7201907634735107, + "num_tokens": 290169041.0, + "step": 11217 + }, + { + "epoch": 1.2319349879200527, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.1388473510742188, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7301039695739746, + "num_tokens": 290196061.0, + "step": 11218 + }, + { + "epoch": 1.2320448056226665, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.2660229206085205, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7097920179367065, + "num_tokens": 290222130.0, + "step": 11219 + }, + { + "epoch": 1.23215462332528, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.46773099899292, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7284461855888367, + "num_tokens": 290245566.0, + "step": 11220 + }, + { + "epoch": 1.2322644410278938, + "ewc_loss": 0.00010156631469726562, + "grad_norm": 2.4296040534973145, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7084968090057373, + "num_tokens": 290268633.0, + "step": 11221 + }, + { + "epoch": 1.2323742587305073, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.0476768016815186, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7385851144790649, + "num_tokens": 290298052.0, + "step": 11222 + }, + { + "epoch": 1.232484076433121, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.3559281826019287, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7324970364570618, + "num_tokens": 290323306.0, + "step": 11223 + }, + { + "epoch": 1.2325938941357346, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.158346652984619, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7206209301948547, + "num_tokens": 290351222.0, + "step": 11224 + }, + { + "epoch": 1.2327037118383484, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.147538423538208, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6899694204330444, + "num_tokens": 290380297.0, + "step": 11225 + }, + { + "epoch": 1.232813529540962, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.076272964477539, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7332128286361694, + "num_tokens": 290410462.0, + "step": 11226 + }, + { + "epoch": 1.2329233472435757, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.1681301593780518, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7155404686927795, + "num_tokens": 290438048.0, + "step": 11227 + }, + { + "epoch": 1.2330331649461894, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 1.9907423257827759, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7048313617706299, + "num_tokens": 290469821.0, + "step": 11228 + }, + { + "epoch": 1.233142982648803, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.173992156982422, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7197302579879761, + "num_tokens": 290495476.0, + "step": 11229 + }, + { + "epoch": 1.2332528003514167, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.342665672302246, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7015271186828613, + "num_tokens": 290519504.0, + "step": 11230 + }, + { + "epoch": 1.2333626180540302, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.0927979946136475, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7055332660675049, + "num_tokens": 290547917.0, + "step": 11231 + }, + { + "epoch": 1.233472435756644, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.375953435897827, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7080759406089783, + "num_tokens": 290575889.0, + "step": 11232 + }, + { + "epoch": 1.2335822534592575, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.4408023357391357, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7324126958847046, + "num_tokens": 290597692.0, + "step": 11233 + }, + { + "epoch": 1.2336920711618713, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.2447433471679688, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.6997002363204956, + "num_tokens": 290624743.0, + "step": 11234 + }, + { + "epoch": 1.2338018888644848, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.2547614574432373, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7078676223754883, + "num_tokens": 290651588.0, + "step": 11235 + }, + { + "epoch": 1.2339117065670986, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.2046918869018555, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7069923877716064, + "num_tokens": 290679796.0, + "step": 11236 + }, + { + "epoch": 1.2340215242697123, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.7812721729278564, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7219898700714111, + "num_tokens": 290698190.0, + "step": 11237 + }, + { + "epoch": 1.2341313419723259, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.2834577560424805, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7273464798927307, + "num_tokens": 290722165.0, + "step": 11238 + }, + { + "epoch": 1.2342411596749396, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 1.9640129804611206, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6969216465950012, + "num_tokens": 290755251.0, + "step": 11239 + }, + { + "epoch": 1.2343509773775532, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.062805652618408, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7173687219619751, + "num_tokens": 290782938.0, + "step": 11240 + }, + { + "epoch": 1.234460795080167, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 1.929355263710022, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7305639982223511, + "num_tokens": 290814105.0, + "step": 11241 + }, + { + "epoch": 1.2345706127827807, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.0727736949920654, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7346658706665039, + "num_tokens": 290841887.0, + "step": 11242 + }, + { + "epoch": 1.2346804304853942, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.1390533447265625, + "learning_rate": 1e-06, + "loss": 1.074, + "mean_token_accuracy": 0.6738596558570862, + "num_tokens": 290872908.0, + "step": 11243 + }, + { + "epoch": 1.234790248188008, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.1121597290039062, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7129638195037842, + "num_tokens": 290902229.0, + "step": 11244 + }, + { + "epoch": 1.2349000658906215, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.123417615890503, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6952050924301147, + "num_tokens": 290930274.0, + "step": 11245 + }, + { + "epoch": 1.2350098835932353, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.248248338699341, + "learning_rate": 1e-06, + "loss": 1.0349, + "mean_token_accuracy": 0.6856465339660645, + "num_tokens": 290957237.0, + "step": 11246 + }, + { + "epoch": 1.2351197012958488, + "ewc_loss": 0.00010251998901367188, + "grad_norm": 2.231959342956543, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.725893497467041, + "num_tokens": 290982019.0, + "step": 11247 + }, + { + "epoch": 1.2352295189984626, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.426177978515625, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7358928322792053, + "num_tokens": 291004914.0, + "step": 11248 + }, + { + "epoch": 1.235339336701076, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0875258445739746, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7140363454818726, + "num_tokens": 291035370.0, + "step": 11249 + }, + { + "epoch": 1.2354491544036899, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.285318613052368, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6971103549003601, + "num_tokens": 291064938.0, + "step": 11250 + }, + { + "epoch": 1.2355589721063036, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.182889699935913, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.710250735282898, + "num_tokens": 291090672.0, + "step": 11251 + }, + { + "epoch": 1.2356687898089171, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.428114891052246, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7182342410087585, + "num_tokens": 291113233.0, + "step": 11252 + }, + { + "epoch": 1.235778607511531, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1079652309417725, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7166398763656616, + "num_tokens": 291142475.0, + "step": 11253 + }, + { + "epoch": 1.2358884252141444, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0898897647857666, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7162340879440308, + "num_tokens": 291169504.0, + "step": 11254 + }, + { + "epoch": 1.2359982429167582, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.043531656265259, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.711287796497345, + "num_tokens": 291202195.0, + "step": 11255 + }, + { + "epoch": 1.236108060619372, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2770516872406006, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6897898316383362, + "num_tokens": 291228469.0, + "step": 11256 + }, + { + "epoch": 1.2362178783219855, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.125999927520752, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6894400715827942, + "num_tokens": 291259022.0, + "step": 11257 + }, + { + "epoch": 1.2363276960245992, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2524163722991943, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7294191718101501, + "num_tokens": 291283767.0, + "step": 11258 + }, + { + "epoch": 1.2364375137272128, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.436934471130371, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.727017879486084, + "num_tokens": 291306139.0, + "step": 11259 + }, + { + "epoch": 1.2365473314298265, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 1.9942069053649902, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7158057689666748, + "num_tokens": 291336887.0, + "step": 11260 + }, + { + "epoch": 1.23665714913244, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.262237310409546, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7014609575271606, + "num_tokens": 291363600.0, + "step": 11261 + }, + { + "epoch": 1.2367669668350538, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.291215658187866, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7197604179382324, + "num_tokens": 291389067.0, + "step": 11262 + }, + { + "epoch": 1.2368767845376674, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.4560725688934326, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.694364070892334, + "num_tokens": 291413585.0, + "step": 11263 + }, + { + "epoch": 1.2369866022402811, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3752312660217285, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7348491549491882, + "num_tokens": 291434867.0, + "step": 11264 + }, + { + "epoch": 1.2370964199428949, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.28619122505188, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7208602428436279, + "num_tokens": 291461764.0, + "step": 11265 + }, + { + "epoch": 1.2372062376455084, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2224526405334473, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.707633376121521, + "num_tokens": 291488950.0, + "step": 11266 + }, + { + "epoch": 1.2373160553481222, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1541407108306885, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.6831820607185364, + "num_tokens": 291516738.0, + "step": 11267 + }, + { + "epoch": 1.2374258730507357, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0421550273895264, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.712675154209137, + "num_tokens": 291548428.0, + "step": 11268 + }, + { + "epoch": 1.2375356907533495, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.258953809738159, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7265489101409912, + "num_tokens": 291572552.0, + "step": 11269 + }, + { + "epoch": 1.2376455084559632, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.243682384490967, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7383697628974915, + "num_tokens": 291597888.0, + "step": 11270 + }, + { + "epoch": 1.2377553261585768, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1846277713775635, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7154154181480408, + "num_tokens": 291624458.0, + "step": 11271 + }, + { + "epoch": 1.2378651438611905, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.29030179977417, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7116429805755615, + "num_tokens": 291649258.0, + "step": 11272 + }, + { + "epoch": 1.237974961563804, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.216766595840454, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7360894083976746, + "num_tokens": 291673698.0, + "step": 11273 + }, + { + "epoch": 1.2380847792664178, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3294591903686523, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7250491380691528, + "num_tokens": 291697675.0, + "step": 11274 + }, + { + "epoch": 1.2381945969690313, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.224483013153076, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7104570269584656, + "num_tokens": 291724573.0, + "step": 11275 + }, + { + "epoch": 1.238304414671645, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.673560857772827, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7323589324951172, + "num_tokens": 291743913.0, + "step": 11276 + }, + { + "epoch": 1.2384142323742586, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2155990600585938, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7351173162460327, + "num_tokens": 291768326.0, + "step": 11277 + }, + { + "epoch": 1.2385240500768724, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.257032632827759, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7068254351615906, + "num_tokens": 291796062.0, + "step": 11278 + }, + { + "epoch": 1.2386338677794861, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2153334617614746, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7173711061477661, + "num_tokens": 291825094.0, + "step": 11279 + }, + { + "epoch": 1.2387436854820997, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2730214595794678, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7198660969734192, + "num_tokens": 291851180.0, + "step": 11280 + }, + { + "epoch": 1.2388535031847134, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2762768268585205, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7296119928359985, + "num_tokens": 291875864.0, + "step": 11281 + }, + { + "epoch": 1.238963320887327, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.2321364879608154, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7211757302284241, + "num_tokens": 291901757.0, + "step": 11282 + }, + { + "epoch": 1.2390731385899407, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3323793411254883, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7353737950325012, + "num_tokens": 291924418.0, + "step": 11283 + }, + { + "epoch": 1.2391829562925545, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.141298770904541, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7122053503990173, + "num_tokens": 291953676.0, + "step": 11284 + }, + { + "epoch": 1.239292773995168, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1447463035583496, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6882128715515137, + "num_tokens": 291982067.0, + "step": 11285 + }, + { + "epoch": 1.2394025916977818, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.5680177211761475, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.725565493106842, + "num_tokens": 292002505.0, + "step": 11286 + }, + { + "epoch": 1.2395124094003953, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.119673013687134, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7142771482467651, + "num_tokens": 292030540.0, + "step": 11287 + }, + { + "epoch": 1.239622227103009, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3286850452423096, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7020930647850037, + "num_tokens": 292055186.0, + "step": 11288 + }, + { + "epoch": 1.2397320448056226, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.4899003505706787, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7403601408004761, + "num_tokens": 292075507.0, + "step": 11289 + }, + { + "epoch": 1.2398418625082364, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.183082342147827, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.701765775680542, + "num_tokens": 292103532.0, + "step": 11290 + }, + { + "epoch": 1.23995168021085, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.138902425765991, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7329590916633606, + "num_tokens": 292131001.0, + "step": 11291 + }, + { + "epoch": 1.2400614979134637, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1006433963775635, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7180389165878296, + "num_tokens": 292159966.0, + "step": 11292 + }, + { + "epoch": 1.2401713156160774, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.288785457611084, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7274637222290039, + "num_tokens": 292185311.0, + "step": 11293 + }, + { + "epoch": 1.240281133318691, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.238464117050171, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7260453701019287, + "num_tokens": 292210456.0, + "step": 11294 + }, + { + "epoch": 1.2403909510213047, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.139526605606079, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7081959247589111, + "num_tokens": 292238954.0, + "step": 11295 + }, + { + "epoch": 1.2405007687239182, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0622475147247314, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7504507303237915, + "num_tokens": 292268336.0, + "step": 11296 + }, + { + "epoch": 1.240610586426532, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3917930126190186, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7339298725128174, + "num_tokens": 292290510.0, + "step": 11297 + }, + { + "epoch": 1.2407204041291455, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1799001693725586, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7333742380142212, + "num_tokens": 292319469.0, + "step": 11298 + }, + { + "epoch": 1.2408302218317593, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.32682466506958, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7105841040611267, + "num_tokens": 292345632.0, + "step": 11299 + }, + { + "epoch": 1.2409400395343728, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.23024320602417, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7090938687324524, + "num_tokens": 292372304.0, + "step": 11300 + }, + { + "epoch": 1.2410498572369866, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3396971225738525, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.751952052116394, + "num_tokens": 292395329.0, + "step": 11301 + }, + { + "epoch": 1.2411596749396003, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3505661487579346, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7210872769355774, + "num_tokens": 292420030.0, + "step": 11302 + }, + { + "epoch": 1.2412694926422139, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.248110055923462, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7126649022102356, + "num_tokens": 292444679.0, + "step": 11303 + }, + { + "epoch": 1.2413793103448276, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.186704397201538, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7212061882019043, + "num_tokens": 292471191.0, + "step": 11304 + }, + { + "epoch": 1.2414891280474412, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.542670965194702, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7211046814918518, + "num_tokens": 292493761.0, + "step": 11305 + }, + { + "epoch": 1.241598945750055, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3520572185516357, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7141542434692383, + "num_tokens": 292516844.0, + "step": 11306 + }, + { + "epoch": 1.2417087634526687, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.467939853668213, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.7014926075935364, + "num_tokens": 292540030.0, + "step": 11307 + }, + { + "epoch": 1.2418185811552822, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1176393032073975, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7201467752456665, + "num_tokens": 292568982.0, + "step": 11308 + }, + { + "epoch": 1.241928398857896, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.545872211456299, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.718834638595581, + "num_tokens": 292589148.0, + "step": 11309 + }, + { + "epoch": 1.2420382165605095, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1518030166625977, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.706376314163208, + "num_tokens": 292617921.0, + "step": 11310 + }, + { + "epoch": 1.2421480342631233, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.1897449493408203, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7388406991958618, + "num_tokens": 292644301.0, + "step": 11311 + }, + { + "epoch": 1.2422578519657368, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0606942176818848, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7222417593002319, + "num_tokens": 292674415.0, + "step": 11312 + }, + { + "epoch": 1.2423676696683505, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.3926682472229004, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7214966416358948, + "num_tokens": 292698202.0, + "step": 11313 + }, + { + "epoch": 1.242477487370964, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0980541706085205, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7157272696495056, + "num_tokens": 292727895.0, + "step": 11314 + }, + { + "epoch": 1.2425873050735778, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.4657204151153564, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.6973066329956055, + "num_tokens": 292751726.0, + "step": 11315 + }, + { + "epoch": 1.2426971227761916, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.276057243347168, + "learning_rate": 1e-06, + "loss": 1.0647, + "mean_token_accuracy": 0.6782891750335693, + "num_tokens": 292782025.0, + "step": 11316 + }, + { + "epoch": 1.2428069404788051, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.180818557739258, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7180632948875427, + "num_tokens": 292809852.0, + "step": 11317 + }, + { + "epoch": 1.242916758181419, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0229861736297607, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7060189247131348, + "num_tokens": 292840054.0, + "step": 11318 + }, + { + "epoch": 1.2430265758840324, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.6774353981018066, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7435462474822998, + "num_tokens": 292858529.0, + "step": 11319 + }, + { + "epoch": 1.2431363935866462, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.022435426712036, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.706358015537262, + "num_tokens": 292888213.0, + "step": 11320 + }, + { + "epoch": 1.24324621128926, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 1.8928459882736206, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7161244750022888, + "num_tokens": 292921633.0, + "step": 11321 + }, + { + "epoch": 1.2433560289918735, + "ewc_loss": 0.00010347366333007812, + "grad_norm": 2.0868897438049316, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7161757946014404, + "num_tokens": 292950474.0, + "step": 11322 + }, + { + "epoch": 1.2434658466944872, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3230438232421875, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.701339840888977, + "num_tokens": 292975140.0, + "step": 11323 + }, + { + "epoch": 1.2435756643971008, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.0204203128814697, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7245376110076904, + "num_tokens": 293005180.0, + "step": 11324 + }, + { + "epoch": 1.2436854820997145, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.335191011428833, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7266978621482849, + "num_tokens": 293028606.0, + "step": 11325 + }, + { + "epoch": 1.243795299802328, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.4708120822906494, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7312967777252197, + "num_tokens": 293049332.0, + "step": 11326 + }, + { + "epoch": 1.2439051175049418, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 1.9381935596466064, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7159383296966553, + "num_tokens": 293079125.0, + "step": 11327 + }, + { + "epoch": 1.2440149352075554, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1919846534729004, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.6990057229995728, + "num_tokens": 293106284.0, + "step": 11328 + }, + { + "epoch": 1.244124752910169, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 1.9130098819732666, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.724410891532898, + "num_tokens": 293137777.0, + "step": 11329 + }, + { + "epoch": 1.2442345706127829, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.679001808166504, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7408678531646729, + "num_tokens": 293157064.0, + "step": 11330 + }, + { + "epoch": 1.2443443883153964, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.4163477420806885, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6958191990852356, + "num_tokens": 293181353.0, + "step": 11331 + }, + { + "epoch": 1.2444542060180102, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.062861442565918, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.705618679523468, + "num_tokens": 293211082.0, + "step": 11332 + }, + { + "epoch": 1.2445640237206237, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.185271739959717, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7293499112129211, + "num_tokens": 293238078.0, + "step": 11333 + }, + { + "epoch": 1.2446738414232374, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.8156063556671143, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.730293869972229, + "num_tokens": 293255608.0, + "step": 11334 + }, + { + "epoch": 1.2447836591258512, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.0492427349090576, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7116626501083374, + "num_tokens": 293285103.0, + "step": 11335 + }, + { + "epoch": 1.2448934768284647, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3121519088745117, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7106168270111084, + "num_tokens": 293309605.0, + "step": 11336 + }, + { + "epoch": 1.2450032945310785, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.221251964569092, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6893608570098877, + "num_tokens": 293336245.0, + "step": 11337 + }, + { + "epoch": 1.245113112233692, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3190340995788574, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7032480835914612, + "num_tokens": 293361734.0, + "step": 11338 + }, + { + "epoch": 1.2452229299363058, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1351566314697266, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7241334915161133, + "num_tokens": 293389235.0, + "step": 11339 + }, + { + "epoch": 1.2453327476389193, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.602159023284912, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7183525562286377, + "num_tokens": 293410452.0, + "step": 11340 + }, + { + "epoch": 1.245442565341533, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.263364553451538, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7357178330421448, + "num_tokens": 293435361.0, + "step": 11341 + }, + { + "epoch": 1.2455523830441466, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.449857234954834, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7371007204055786, + "num_tokens": 293457915.0, + "step": 11342 + }, + { + "epoch": 1.2456622007467604, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.45059871673584, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7458789348602295, + "num_tokens": 293478810.0, + "step": 11343 + }, + { + "epoch": 1.2457720184493741, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3287951946258545, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7338136434555054, + "num_tokens": 293502615.0, + "step": 11344 + }, + { + "epoch": 1.2458818361519877, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2257845401763916, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7233508229255676, + "num_tokens": 293527291.0, + "step": 11345 + }, + { + "epoch": 1.2459916538546014, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1510438919067383, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7359616756439209, + "num_tokens": 293554129.0, + "step": 11346 + }, + { + "epoch": 1.246101471557215, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1911158561706543, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7247148156166077, + "num_tokens": 293579967.0, + "step": 11347 + }, + { + "epoch": 1.2462112892598287, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1693780422210693, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7328481078147888, + "num_tokens": 293606901.0, + "step": 11348 + }, + { + "epoch": 1.2463211069624422, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.250777006149292, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7030837535858154, + "num_tokens": 293634345.0, + "step": 11349 + }, + { + "epoch": 1.246430924665056, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1952881813049316, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7109622955322266, + "num_tokens": 293660439.0, + "step": 11350 + }, + { + "epoch": 1.2465407423676698, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2185444831848145, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7355822324752808, + "num_tokens": 293685799.0, + "step": 11351 + }, + { + "epoch": 1.2466505600702833, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.326451063156128, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7308009266853333, + "num_tokens": 293710985.0, + "step": 11352 + }, + { + "epoch": 1.246760377772897, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1034834384918213, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6896964311599731, + "num_tokens": 293742124.0, + "step": 11353 + }, + { + "epoch": 1.2468701954755106, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.461103677749634, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7149426937103271, + "num_tokens": 293764646.0, + "step": 11354 + }, + { + "epoch": 1.2469800131781243, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.4661359786987305, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7256656289100647, + "num_tokens": 293788042.0, + "step": 11355 + }, + { + "epoch": 1.2470898308807379, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3921077251434326, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7331204414367676, + "num_tokens": 293811110.0, + "step": 11356 + }, + { + "epoch": 1.2471996485833516, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.118485927581787, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7005094289779663, + "num_tokens": 293839656.0, + "step": 11357 + }, + { + "epoch": 1.2473094662859654, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.362814426422119, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6980904340744019, + "num_tokens": 293864921.0, + "step": 11358 + }, + { + "epoch": 1.247419283988579, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2841897010803223, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7147201299667358, + "num_tokens": 293891997.0, + "step": 11359 + }, + { + "epoch": 1.2475291016911927, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 1.9516664743423462, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.697173535823822, + "num_tokens": 293928415.0, + "step": 11360 + }, + { + "epoch": 1.2476389193938062, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 1.8735607862472534, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7103252410888672, + "num_tokens": 293964945.0, + "step": 11361 + }, + { + "epoch": 1.24774873709642, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.184830904006958, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7074268460273743, + "num_tokens": 293991970.0, + "step": 11362 + }, + { + "epoch": 1.2478585547990335, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2698352336883545, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7202260494232178, + "num_tokens": 294018661.0, + "step": 11363 + }, + { + "epoch": 1.2479683725016473, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1375796794891357, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7475509643554688, + "num_tokens": 294046716.0, + "step": 11364 + }, + { + "epoch": 1.2480781902042608, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.157543897628784, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7492592334747314, + "num_tokens": 294071987.0, + "step": 11365 + }, + { + "epoch": 1.2481880079068746, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.359127998352051, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7408782243728638, + "num_tokens": 294095503.0, + "step": 11366 + }, + { + "epoch": 1.2482978256094883, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2863659858703613, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6956441402435303, + "num_tokens": 294122424.0, + "step": 11367 + }, + { + "epoch": 1.2484076433121019, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.142015218734741, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7155501246452332, + "num_tokens": 294150376.0, + "step": 11368 + }, + { + "epoch": 1.2485174610147156, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.18896222114563, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7206017971038818, + "num_tokens": 294177393.0, + "step": 11369 + }, + { + "epoch": 1.2486272787173291, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.271132230758667, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.7108067870140076, + "num_tokens": 294203511.0, + "step": 11370 + }, + { + "epoch": 1.248737096419943, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.0575928688049316, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7159964442253113, + "num_tokens": 294232230.0, + "step": 11371 + }, + { + "epoch": 1.2488469141225567, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1161892414093018, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7413777112960815, + "num_tokens": 294261107.0, + "step": 11372 + }, + { + "epoch": 1.2489567318251702, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.6434450149536133, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7292338609695435, + "num_tokens": 294283492.0, + "step": 11373 + }, + { + "epoch": 1.249066549527784, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2109813690185547, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7229920029640198, + "num_tokens": 294308887.0, + "step": 11374 + }, + { + "epoch": 1.2491763672303975, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2838070392608643, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7098870277404785, + "num_tokens": 294336422.0, + "step": 11375 + }, + { + "epoch": 1.2492861849330112, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.220132827758789, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.700060248374939, + "num_tokens": 294363697.0, + "step": 11376 + }, + { + "epoch": 1.2493960026356248, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2489991188049316, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7321516275405884, + "num_tokens": 294387983.0, + "step": 11377 + }, + { + "epoch": 1.2495058203382385, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.192049503326416, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7093456983566284, + "num_tokens": 294414661.0, + "step": 11378 + }, + { + "epoch": 1.249615638040852, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.164412498474121, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7280109524726868, + "num_tokens": 294440692.0, + "step": 11379 + }, + { + "epoch": 1.2497254557434658, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.439197063446045, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7161552906036377, + "num_tokens": 294468741.0, + "step": 11380 + }, + { + "epoch": 1.2498352734460796, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.519120454788208, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7271418571472168, + "num_tokens": 294489488.0, + "step": 11381 + }, + { + "epoch": 1.2499450911486931, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.341909170150757, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.714789867401123, + "num_tokens": 294513430.0, + "step": 11382 + }, + { + "epoch": 1.2500549088513069, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.0482850074768066, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7185408473014832, + "num_tokens": 294542188.0, + "step": 11383 + }, + { + "epoch": 1.2501647265539204, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.365320920944214, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7156779766082764, + "num_tokens": 294566002.0, + "step": 11384 + }, + { + "epoch": 1.2502745442565342, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.194105863571167, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7142324447631836, + "num_tokens": 294591308.0, + "step": 11385 + }, + { + "epoch": 1.250384361959148, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2886006832122803, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7021619081497192, + "num_tokens": 294615500.0, + "step": 11386 + }, + { + "epoch": 1.2504941796617615, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2535343170166016, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7196677923202515, + "num_tokens": 294640524.0, + "step": 11387 + }, + { + "epoch": 1.2506039973643752, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.0187788009643555, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7105382680892944, + "num_tokens": 294670595.0, + "step": 11388 + }, + { + "epoch": 1.2507138150669888, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.27350115776062, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7034577131271362, + "num_tokens": 294696095.0, + "step": 11389 + }, + { + "epoch": 1.2508236327696025, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1713671684265137, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7270298600196838, + "num_tokens": 294722676.0, + "step": 11390 + }, + { + "epoch": 1.250933450472216, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.064389705657959, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7190357446670532, + "num_tokens": 294752360.0, + "step": 11391 + }, + { + "epoch": 1.2510432681748298, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2526750564575195, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7096998691558838, + "num_tokens": 294777603.0, + "step": 11392 + }, + { + "epoch": 1.2511530858774433, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.189727783203125, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7311407327651978, + "num_tokens": 294802529.0, + "step": 11393 + }, + { + "epoch": 1.251262903580057, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3249025344848633, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7002003192901611, + "num_tokens": 294827373.0, + "step": 11394 + }, + { + "epoch": 1.2513727212826709, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2341482639312744, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7194133996963501, + "num_tokens": 294855380.0, + "step": 11395 + }, + { + "epoch": 1.2514825389852844, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3071448802948, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7485204935073853, + "num_tokens": 294877037.0, + "step": 11396 + }, + { + "epoch": 1.2515923566878981, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.3127920627593994, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7136209011077881, + "num_tokens": 294901232.0, + "step": 11397 + }, + { + "epoch": 1.2517021743905117, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.4608218669891357, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.730938196182251, + "num_tokens": 294921628.0, + "step": 11398 + }, + { + "epoch": 1.2518119920931254, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2694175243377686, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7283082604408264, + "num_tokens": 294947376.0, + "step": 11399 + }, + { + "epoch": 1.2519218097957392, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.1492860317230225, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.732897937297821, + "num_tokens": 294976345.0, + "step": 11400 + }, + { + "epoch": 1.2520316274983527, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.2834439277648926, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.723162055015564, + "num_tokens": 295000527.0, + "step": 11401 + }, + { + "epoch": 1.2521414452009663, + "ewc_loss": 0.00010395050048828125, + "grad_norm": 2.539930582046509, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.70006263256073, + "num_tokens": 295021857.0, + "step": 11402 + }, + { + "epoch": 1.25225126290358, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.402625560760498, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.7480800747871399, + "num_tokens": 295045127.0, + "step": 11403 + }, + { + "epoch": 1.2523610806061938, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.261655807495117, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7390381097793579, + "num_tokens": 295070148.0, + "step": 11404 + }, + { + "epoch": 1.2524708983088073, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.511772394180298, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7160676717758179, + "num_tokens": 295089819.0, + "step": 11405 + }, + { + "epoch": 1.252580716011421, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.7871317863464355, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7213005423545837, + "num_tokens": 295107843.0, + "step": 11406 + }, + { + "epoch": 1.2526905337140346, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.3210175037384033, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7308573722839355, + "num_tokens": 295133757.0, + "step": 11407 + }, + { + "epoch": 1.2528003514166484, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.253308057785034, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7097326517105103, + "num_tokens": 295161640.0, + "step": 11408 + }, + { + "epoch": 1.2529101691192621, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.121457099914551, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7326726317405701, + "num_tokens": 295189968.0, + "step": 11409 + }, + { + "epoch": 1.2530199868218757, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.425513505935669, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7340804934501648, + "num_tokens": 295214500.0, + "step": 11410 + }, + { + "epoch": 1.2531298045244894, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.412015438079834, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7091965079307556, + "num_tokens": 295237895.0, + "step": 11411 + }, + { + "epoch": 1.253239622227103, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.310084342956543, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7304362654685974, + "num_tokens": 295260666.0, + "step": 11412 + }, + { + "epoch": 1.2533494399297167, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 1.9996334314346313, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7030154466629028, + "num_tokens": 295293423.0, + "step": 11413 + }, + { + "epoch": 1.2534592576323305, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.250800371170044, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7228931784629822, + "num_tokens": 295318031.0, + "step": 11414 + }, + { + "epoch": 1.253569075334944, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.1488640308380127, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.690143883228302, + "num_tokens": 295345764.0, + "step": 11415 + }, + { + "epoch": 1.2536788930375575, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.389108657836914, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6893463134765625, + "num_tokens": 295369222.0, + "step": 11416 + }, + { + "epoch": 1.2537887107401713, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.5233664512634277, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7431702017784119, + "num_tokens": 295390231.0, + "step": 11417 + }, + { + "epoch": 1.253898528442785, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.6035938262939453, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7247427701950073, + "num_tokens": 295411044.0, + "step": 11418 + }, + { + "epoch": 1.2540083461453986, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.672921657562256, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.731768786907196, + "num_tokens": 295432250.0, + "step": 11419 + }, + { + "epoch": 1.2541181638480123, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.174004077911377, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7200412750244141, + "num_tokens": 295462432.0, + "step": 11420 + }, + { + "epoch": 1.2542279815506259, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.131657600402832, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7019731998443604, + "num_tokens": 295490547.0, + "step": 11421 + }, + { + "epoch": 1.2543377992532396, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.3276994228363037, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7183947563171387, + "num_tokens": 295513531.0, + "step": 11422 + }, + { + "epoch": 1.2544476169558534, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.405937433242798, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7387688159942627, + "num_tokens": 295535425.0, + "step": 11423 + }, + { + "epoch": 1.254557434658467, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.302002191543579, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.722390353679657, + "num_tokens": 295561283.0, + "step": 11424 + }, + { + "epoch": 1.2546672523610807, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.1383261680603027, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7217297554016113, + "num_tokens": 295589515.0, + "step": 11425 + }, + { + "epoch": 1.2547770700636942, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.3353614807128906, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7100340127944946, + "num_tokens": 295615953.0, + "step": 11426 + }, + { + "epoch": 1.254886887766308, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.1914498805999756, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7176902890205383, + "num_tokens": 295644770.0, + "step": 11427 + }, + { + "epoch": 1.2549967054689217, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.4612786769866943, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7266689538955688, + "num_tokens": 295665904.0, + "step": 11428 + }, + { + "epoch": 1.2551065231715353, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 1.920671820640564, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7108376026153564, + "num_tokens": 295698262.0, + "step": 11429 + }, + { + "epoch": 1.2552163408741488, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.2651777267456055, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6896785497665405, + "num_tokens": 295723447.0, + "step": 11430 + }, + { + "epoch": 1.2553261585767626, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.433419942855835, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7331794500350952, + "num_tokens": 295744698.0, + "step": 11431 + }, + { + "epoch": 1.2554359762793763, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.5968384742736816, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7294586896896362, + "num_tokens": 295764393.0, + "step": 11432 + }, + { + "epoch": 1.2555457939819898, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.3087964057922363, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7126253843307495, + "num_tokens": 295790134.0, + "step": 11433 + }, + { + "epoch": 1.2556556116846036, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.494443893432617, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.722293496131897, + "num_tokens": 295811072.0, + "step": 11434 + }, + { + "epoch": 1.2557654293872171, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.132936477661133, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7451256513595581, + "num_tokens": 295837017.0, + "step": 11435 + }, + { + "epoch": 1.255875247089831, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.374940872192383, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7170017957687378, + "num_tokens": 295859040.0, + "step": 11436 + }, + { + "epoch": 1.2559850647924446, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.803027868270874, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7461539506912231, + "num_tokens": 295875491.0, + "step": 11437 + }, + { + "epoch": 1.2560948824950582, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.3130860328674316, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7331977486610413, + "num_tokens": 295897898.0, + "step": 11438 + }, + { + "epoch": 1.256204700197672, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.0295262336730957, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.6964707374572754, + "num_tokens": 295928789.0, + "step": 11439 + }, + { + "epoch": 1.2563145179002855, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.2296102046966553, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7225348949432373, + "num_tokens": 295953425.0, + "step": 11440 + }, + { + "epoch": 1.2564243356028992, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.0871825218200684, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.6987799406051636, + "num_tokens": 295982609.0, + "step": 11441 + }, + { + "epoch": 1.2565341533055128, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.2979395389556885, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6928171515464783, + "num_tokens": 296008120.0, + "step": 11442 + }, + { + "epoch": 1.2566439710081265, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.731330394744873, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7356480956077576, + "num_tokens": 296025613.0, + "step": 11443 + }, + { + "epoch": 1.25675378871074, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.114503860473633, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.714372456073761, + "num_tokens": 296053874.0, + "step": 11444 + }, + { + "epoch": 1.2568636064133538, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.659747362136841, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7322404384613037, + "num_tokens": 296074215.0, + "step": 11445 + }, + { + "epoch": 1.2569734241159676, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.1487956047058105, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7057672142982483, + "num_tokens": 296102474.0, + "step": 11446 + }, + { + "epoch": 1.257083241818581, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.1451447010040283, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.697237491607666, + "num_tokens": 296131694.0, + "step": 11447 + }, + { + "epoch": 1.2571930595211949, + "ewc_loss": 0.00010442733764648438, + "grad_norm": 2.436445951461792, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7379903197288513, + "num_tokens": 296153126.0, + "step": 11448 + }, + { + "epoch": 1.2573028772238084, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.362802743911743, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7014960646629333, + "num_tokens": 296179671.0, + "step": 11449 + }, + { + "epoch": 1.2574126949264222, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.4606990814208984, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7017571926116943, + "num_tokens": 296204581.0, + "step": 11450 + }, + { + "epoch": 1.257522512629036, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.6689062118530273, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7171639204025269, + "num_tokens": 296224412.0, + "step": 11451 + }, + { + "epoch": 1.2576323303316495, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.510775327682495, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7173556089401245, + "num_tokens": 296246127.0, + "step": 11452 + }, + { + "epoch": 1.257742148034263, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.296332836151123, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.710018515586853, + "num_tokens": 296269949.0, + "step": 11453 + }, + { + "epoch": 1.2578519657368767, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.5648255348205566, + "learning_rate": 1e-06, + "loss": 1.0385, + "mean_token_accuracy": 0.6939403414726257, + "num_tokens": 296291955.0, + "step": 11454 + }, + { + "epoch": 1.2579617834394905, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.2546017169952393, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7241367697715759, + "num_tokens": 296319512.0, + "step": 11455 + }, + { + "epoch": 1.258071601142104, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.354618549346924, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7122097015380859, + "num_tokens": 296344515.0, + "step": 11456 + }, + { + "epoch": 1.2581814188447178, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.2586073875427246, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7063015103340149, + "num_tokens": 296371510.0, + "step": 11457 + }, + { + "epoch": 1.2582912365473313, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.3146026134490967, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7257370948791504, + "num_tokens": 296395037.0, + "step": 11458 + }, + { + "epoch": 1.258401054249945, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.2788538932800293, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7208594083786011, + "num_tokens": 296420556.0, + "step": 11459 + }, + { + "epoch": 1.2585108719525588, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.1400201320648193, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7018505334854126, + "num_tokens": 296449227.0, + "step": 11460 + }, + { + "epoch": 1.2586206896551724, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.1690807342529297, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7213160395622253, + "num_tokens": 296474535.0, + "step": 11461 + }, + { + "epoch": 1.2587305073577861, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.3565731048583984, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7049962878227234, + "num_tokens": 296498255.0, + "step": 11462 + }, + { + "epoch": 1.2588403250603997, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.087435007095337, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7280607223510742, + "num_tokens": 296527472.0, + "step": 11463 + }, + { + "epoch": 1.2589501427630134, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.403595209121704, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7120761275291443, + "num_tokens": 296551741.0, + "step": 11464 + }, + { + "epoch": 1.2590599604656272, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.11704158782959, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7451658248901367, + "num_tokens": 296577777.0, + "step": 11465 + }, + { + "epoch": 1.2591697781682407, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.0168378353118896, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7165522575378418, + "num_tokens": 296611448.0, + "step": 11466 + }, + { + "epoch": 1.2592795958708543, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.4471027851104736, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7157471179962158, + "num_tokens": 296638612.0, + "step": 11467 + }, + { + "epoch": 1.259389413573468, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.241039276123047, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7188600897789001, + "num_tokens": 296663290.0, + "step": 11468 + }, + { + "epoch": 1.2594992312760818, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.236828088760376, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7325177192687988, + "num_tokens": 296688612.0, + "step": 11469 + }, + { + "epoch": 1.2596090489786953, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.337336778640747, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6915526986122131, + "num_tokens": 296718311.0, + "step": 11470 + }, + { + "epoch": 1.259718866681309, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.337641716003418, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.712019681930542, + "num_tokens": 296741384.0, + "step": 11471 + }, + { + "epoch": 1.2598286843839226, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.333540439605713, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7346233129501343, + "num_tokens": 296764291.0, + "step": 11472 + }, + { + "epoch": 1.2599385020865363, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.250610113143921, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7154864072799683, + "num_tokens": 296790654.0, + "step": 11473 + }, + { + "epoch": 1.26004831978915, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.012805700302124, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7290829420089722, + "num_tokens": 296821205.0, + "step": 11474 + }, + { + "epoch": 1.2601581374917636, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.030879497528076, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7313125133514404, + "num_tokens": 296849946.0, + "step": 11475 + }, + { + "epoch": 1.2602679551943774, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.1926651000976562, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7139162421226501, + "num_tokens": 296879116.0, + "step": 11476 + }, + { + "epoch": 1.260377772896991, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.2458200454711914, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6966066956520081, + "num_tokens": 296905714.0, + "step": 11477 + }, + { + "epoch": 1.2604875905996047, + "ewc_loss": 0.00010538101196289062, + "grad_norm": 2.1945788860321045, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7133339643478394, + "num_tokens": 296932215.0, + "step": 11478 + }, + { + "epoch": 1.2605974083022184, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.091130495071411, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7624753713607788, + "num_tokens": 296959360.0, + "step": 11479 + }, + { + "epoch": 1.260707226004832, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.385089874267578, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7039686441421509, + "num_tokens": 296981937.0, + "step": 11480 + }, + { + "epoch": 1.2608170437074455, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.95167875289917, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7112282514572144, + "num_tokens": 297013783.0, + "step": 11481 + }, + { + "epoch": 1.2609268614100593, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.002898693084717, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7005738019943237, + "num_tokens": 297045639.0, + "step": 11482 + }, + { + "epoch": 1.261036679112673, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1122612953186035, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7041741013526917, + "num_tokens": 297074810.0, + "step": 11483 + }, + { + "epoch": 1.2611464968152866, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1894426345825195, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7384572625160217, + "num_tokens": 297100410.0, + "step": 11484 + }, + { + "epoch": 1.2612563145179003, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.9731500148773193, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7249605059623718, + "num_tokens": 297132999.0, + "step": 11485 + }, + { + "epoch": 1.2613661322205139, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3672375679016113, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7213088274002075, + "num_tokens": 297157221.0, + "step": 11486 + }, + { + "epoch": 1.2614759499231276, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4276769161224365, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7088791131973267, + "num_tokens": 297183062.0, + "step": 11487 + }, + { + "epoch": 1.2615857676257414, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2941527366638184, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7319395542144775, + "num_tokens": 297206154.0, + "step": 11488 + }, + { + "epoch": 1.261695585328355, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.353259325027466, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7325437664985657, + "num_tokens": 297228651.0, + "step": 11489 + }, + { + "epoch": 1.2618054030309687, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0933330059051514, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7151072025299072, + "num_tokens": 297259028.0, + "step": 11490 + }, + { + "epoch": 1.2619152207335822, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1673905849456787, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7211513519287109, + "num_tokens": 297285857.0, + "step": 11491 + }, + { + "epoch": 1.262025038436196, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.324995756149292, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6944984197616577, + "num_tokens": 297312496.0, + "step": 11492 + }, + { + "epoch": 1.2621348561388097, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.223416328430176, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7007801532745361, + "num_tokens": 297338471.0, + "step": 11493 + }, + { + "epoch": 1.2622446738414232, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4082000255584717, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7253410816192627, + "num_tokens": 297363122.0, + "step": 11494 + }, + { + "epoch": 1.2623544915440368, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.103956460952759, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7098054885864258, + "num_tokens": 297393466.0, + "step": 11495 + }, + { + "epoch": 1.2624643092466505, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2757177352905273, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7198820114135742, + "num_tokens": 297422391.0, + "step": 11496 + }, + { + "epoch": 1.2625741269492643, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.104889392852783, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6958326101303101, + "num_tokens": 297451099.0, + "step": 11497 + }, + { + "epoch": 1.2626839446518778, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.26259183883667, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7063020467758179, + "num_tokens": 297476115.0, + "step": 11498 + }, + { + "epoch": 1.2627937623544916, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4659345149993896, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7026157975196838, + "num_tokens": 297498409.0, + "step": 11499 + }, + { + "epoch": 1.2629035800571051, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.378960371017456, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7311860918998718, + "num_tokens": 297521611.0, + "step": 11500 + }, + { + "epoch": 1.2630133977597189, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1123807430267334, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7327574491500854, + "num_tokens": 297548205.0, + "step": 11501 + }, + { + "epoch": 1.2631232154623326, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3249683380126953, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.697866678237915, + "num_tokens": 297573850.0, + "step": 11502 + }, + { + "epoch": 1.2632330331649462, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.7234506607055664, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7248543500900269, + "num_tokens": 297593424.0, + "step": 11503 + }, + { + "epoch": 1.26334285086756, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.320770502090454, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7025640606880188, + "num_tokens": 297618308.0, + "step": 11504 + }, + { + "epoch": 1.2634526685701735, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.363229513168335, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7194726467132568, + "num_tokens": 297642742.0, + "step": 11505 + }, + { + "epoch": 1.2635624862727872, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.326766014099121, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7271007895469666, + "num_tokens": 297667547.0, + "step": 11506 + }, + { + "epoch": 1.2636723039754008, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2607297897338867, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7474499940872192, + "num_tokens": 297691845.0, + "step": 11507 + }, + { + "epoch": 1.2637821216780145, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3323826789855957, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7263926863670349, + "num_tokens": 297716930.0, + "step": 11508 + }, + { + "epoch": 1.263891939380628, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.045231342315674, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7064869403839111, + "num_tokens": 297747549.0, + "step": 11509 + }, + { + "epoch": 1.2640017570832418, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.9467732906341553, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7340760231018066, + "num_tokens": 297766898.0, + "step": 11510 + }, + { + "epoch": 1.2641115747858556, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3406331539154053, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7034081220626831, + "num_tokens": 297791316.0, + "step": 11511 + }, + { + "epoch": 1.264221392488469, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.5848939418792725, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.716280460357666, + "num_tokens": 297811952.0, + "step": 11512 + }, + { + "epoch": 1.2643312101910829, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4494261741638184, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.723180890083313, + "num_tokens": 297833174.0, + "step": 11513 + }, + { + "epoch": 1.2644410278936964, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.571040630340576, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7146059274673462, + "num_tokens": 297853684.0, + "step": 11514 + }, + { + "epoch": 1.2645508455963101, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.209368944168091, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.700377345085144, + "num_tokens": 297880080.0, + "step": 11515 + }, + { + "epoch": 1.264660663298924, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.102731227874756, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.70090651512146, + "num_tokens": 297909560.0, + "step": 11516 + }, + { + "epoch": 1.2647704810015374, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3385820388793945, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7165141105651855, + "num_tokens": 297932492.0, + "step": 11517 + }, + { + "epoch": 1.264880298704151, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.092529058456421, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.700691282749176, + "num_tokens": 297965319.0, + "step": 11518 + }, + { + "epoch": 1.2649901164067647, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.251678228378296, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6997540593147278, + "num_tokens": 297993903.0, + "step": 11519 + }, + { + "epoch": 1.2650999341093785, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2242655754089355, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7054322957992554, + "num_tokens": 298022800.0, + "step": 11520 + }, + { + "epoch": 1.265209751811992, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1081438064575195, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7322183847427368, + "num_tokens": 298052026.0, + "step": 11521 + }, + { + "epoch": 1.2653195695146058, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3535208702087402, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7292335033416748, + "num_tokens": 298076619.0, + "step": 11522 + }, + { + "epoch": 1.2654293872172193, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.400686502456665, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7232917547225952, + "num_tokens": 298100965.0, + "step": 11523 + }, + { + "epoch": 1.265539204919833, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3835678100585938, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.689924418926239, + "num_tokens": 298124809.0, + "step": 11524 + }, + { + "epoch": 1.2656490226224468, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0847957134246826, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7046023607254028, + "num_tokens": 298155550.0, + "step": 11525 + }, + { + "epoch": 1.2657588403250604, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1540982723236084, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.71063232421875, + "num_tokens": 298181998.0, + "step": 11526 + }, + { + "epoch": 1.2658686580276741, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.024869203567505, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7269912958145142, + "num_tokens": 298211803.0, + "step": 11527 + }, + { + "epoch": 1.2659784757302877, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.546053647994995, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7072464227676392, + "num_tokens": 298234110.0, + "step": 11528 + }, + { + "epoch": 1.2660882934329014, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.034560441970825, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.6955379247665405, + "num_tokens": 298265890.0, + "step": 11529 + }, + { + "epoch": 1.2661981111355152, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2693068981170654, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7249819040298462, + "num_tokens": 298290478.0, + "step": 11530 + }, + { + "epoch": 1.2663079288381287, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1310513019561768, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7298423051834106, + "num_tokens": 298318961.0, + "step": 11531 + }, + { + "epoch": 1.2664177465407422, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.376481771469116, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7351384162902832, + "num_tokens": 298341171.0, + "step": 11532 + }, + { + "epoch": 1.266527564243356, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3225064277648926, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.727241039276123, + "num_tokens": 298365083.0, + "step": 11533 + }, + { + "epoch": 1.2666373819459698, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0249173641204834, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7110031843185425, + "num_tokens": 298396270.0, + "step": 11534 + }, + { + "epoch": 1.2667471996485833, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.416531801223755, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7135922908782959, + "num_tokens": 298420344.0, + "step": 11535 + }, + { + "epoch": 1.266857017351197, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.9795857667922974, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7161500453948975, + "num_tokens": 298451386.0, + "step": 11536 + }, + { + "epoch": 1.2669668350538106, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.484384775161743, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7089008092880249, + "num_tokens": 298473767.0, + "step": 11537 + }, + { + "epoch": 1.2670766527564243, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.493682861328125, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7292451858520508, + "num_tokens": 298495330.0, + "step": 11538 + }, + { + "epoch": 1.267186470459038, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1762661933898926, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7050119042396545, + "num_tokens": 298524502.0, + "step": 11539 + }, + { + "epoch": 1.2672962881616516, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1776576042175293, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7146386504173279, + "num_tokens": 298552364.0, + "step": 11540 + }, + { + "epoch": 1.2674061058642654, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2236239910125732, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7169440984725952, + "num_tokens": 298578078.0, + "step": 11541 + }, + { + "epoch": 1.267515923566879, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1384029388427734, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7182174921035767, + "num_tokens": 298605107.0, + "step": 11542 + }, + { + "epoch": 1.2676257412694927, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.47308611869812, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7314268350601196, + "num_tokens": 298625913.0, + "step": 11543 + }, + { + "epoch": 1.2677355589721064, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.067929744720459, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6921882629394531, + "num_tokens": 298656895.0, + "step": 11544 + }, + { + "epoch": 1.26784537667472, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4652698040008545, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7172839641571045, + "num_tokens": 298680044.0, + "step": 11545 + }, + { + "epoch": 1.2679551943773335, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.235133171081543, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7087386846542358, + "num_tokens": 298707118.0, + "step": 11546 + }, + { + "epoch": 1.2680650120799473, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.170569896697998, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7040162086486816, + "num_tokens": 298731089.0, + "step": 11547 + }, + { + "epoch": 1.268174829782561, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1445059776306152, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7393237948417664, + "num_tokens": 298757668.0, + "step": 11548 + }, + { + "epoch": 1.2682846474851746, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.105334520339966, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.72966468334198, + "num_tokens": 298785147.0, + "step": 11549 + }, + { + "epoch": 1.2683944651877883, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.093074321746826, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.709735095500946, + "num_tokens": 298812858.0, + "step": 11550 + }, + { + "epoch": 1.2685042828904018, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2513644695281982, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7282170057296753, + "num_tokens": 298835848.0, + "step": 11551 + }, + { + "epoch": 1.2686141005930156, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.493927240371704, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7140443325042725, + "num_tokens": 298856536.0, + "step": 11552 + }, + { + "epoch": 1.2687239182956294, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1598691940307617, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7290364503860474, + "num_tokens": 298883334.0, + "step": 11553 + }, + { + "epoch": 1.268833735998243, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.6492490768432617, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7442591190338135, + "num_tokens": 298902690.0, + "step": 11554 + }, + { + "epoch": 1.2689435537008567, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.9639612436294556, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7098771333694458, + "num_tokens": 298935723.0, + "step": 11555 + }, + { + "epoch": 1.2690533714034702, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2588987350463867, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7166811227798462, + "num_tokens": 298961129.0, + "step": 11556 + }, + { + "epoch": 1.269163189106084, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.085015296936035, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7273005843162537, + "num_tokens": 298989359.0, + "step": 11557 + }, + { + "epoch": 1.2692730068086975, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1164095401763916, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7284396290779114, + "num_tokens": 299018271.0, + "step": 11558 + }, + { + "epoch": 1.2693828245113112, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.444584369659424, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6960698962211609, + "num_tokens": 299041732.0, + "step": 11559 + }, + { + "epoch": 1.2694926422139248, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2674808502197266, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7270578145980835, + "num_tokens": 299066729.0, + "step": 11560 + }, + { + "epoch": 1.2696024599165385, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.441136121749878, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.708032488822937, + "num_tokens": 299088342.0, + "step": 11561 + }, + { + "epoch": 1.2697122776191523, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.355973243713379, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.6981193423271179, + "num_tokens": 299113587.0, + "step": 11562 + }, + { + "epoch": 1.2698220953217658, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2920265197753906, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.710837721824646, + "num_tokens": 299140214.0, + "step": 11563 + }, + { + "epoch": 1.2699319130243796, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.9438602924346924, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7068751454353333, + "num_tokens": 299174699.0, + "step": 11564 + }, + { + "epoch": 1.2700417307269931, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.95583975315094, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7286039590835571, + "num_tokens": 299206309.0, + "step": 11565 + }, + { + "epoch": 1.2701515484296069, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.5786800384521484, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7235029935836792, + "num_tokens": 299225998.0, + "step": 11566 + }, + { + "epoch": 1.2702613661322206, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0599844455718994, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7155452370643616, + "num_tokens": 299258633.0, + "step": 11567 + }, + { + "epoch": 1.2703711838348342, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4940595626831055, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7090340852737427, + "num_tokens": 299280740.0, + "step": 11568 + }, + { + "epoch": 1.270481001537448, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4307310581207275, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7174023389816284, + "num_tokens": 299304369.0, + "step": 11569 + }, + { + "epoch": 1.2705908192400615, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2253336906433105, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7006677985191345, + "num_tokens": 299332152.0, + "step": 11570 + }, + { + "epoch": 1.2707006369426752, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1764931678771973, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.6980880498886108, + "num_tokens": 299361670.0, + "step": 11571 + }, + { + "epoch": 1.2708104546452887, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.2572200298309326, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.6983561515808105, + "num_tokens": 299389803.0, + "step": 11572 + }, + { + "epoch": 1.2709202723479025, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0269711017608643, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.716586709022522, + "num_tokens": 299420332.0, + "step": 11573 + }, + { + "epoch": 1.271030090050516, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3478169441223145, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6983948945999146, + "num_tokens": 299444266.0, + "step": 11574 + }, + { + "epoch": 1.2711399077531298, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 1.9675921201705933, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7117462158203125, + "num_tokens": 299476084.0, + "step": 11575 + }, + { + "epoch": 1.2712497254557436, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3051183223724365, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7185558080673218, + "num_tokens": 299501733.0, + "step": 11576 + }, + { + "epoch": 1.271359543158357, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3919677734375, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7108360528945923, + "num_tokens": 299524569.0, + "step": 11577 + }, + { + "epoch": 1.2714693608609708, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.285475015640259, + "learning_rate": 1e-06, + "loss": 0.796, + "mean_token_accuracy": 0.7455956935882568, + "num_tokens": 299546818.0, + "step": 11578 + }, + { + "epoch": 1.2715791785635844, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.5401670932769775, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7188625335693359, + "num_tokens": 299568630.0, + "step": 11579 + }, + { + "epoch": 1.2716889962661981, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.240387201309204, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7052586078643799, + "num_tokens": 299595091.0, + "step": 11580 + }, + { + "epoch": 1.271798813968812, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.078749895095825, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7206668257713318, + "num_tokens": 299624124.0, + "step": 11581 + }, + { + "epoch": 1.2719086316714254, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.286757707595825, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.69999760389328, + "num_tokens": 299649454.0, + "step": 11582 + }, + { + "epoch": 1.272018449374039, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3724308013916016, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.706138551235199, + "num_tokens": 299675046.0, + "step": 11583 + }, + { + "epoch": 1.2721282670766527, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.190290927886963, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7001059651374817, + "num_tokens": 299704160.0, + "step": 11584 + }, + { + "epoch": 1.2722380847792665, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.286201238632202, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6881295442581177, + "num_tokens": 299730361.0, + "step": 11585 + }, + { + "epoch": 1.27234790248188, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.1859633922576904, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7180464863777161, + "num_tokens": 299757251.0, + "step": 11586 + }, + { + "epoch": 1.2724577201844938, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4915311336517334, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7320983409881592, + "num_tokens": 299777812.0, + "step": 11587 + }, + { + "epoch": 1.2725675378871073, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.3945116996765137, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7456461191177368, + "num_tokens": 299799763.0, + "step": 11588 + }, + { + "epoch": 1.272677355589721, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.241605520248413, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7106667757034302, + "num_tokens": 299823944.0, + "step": 11589 + }, + { + "epoch": 1.2727871732923348, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.4436211585998535, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7382342219352722, + "num_tokens": 299846384.0, + "step": 11590 + }, + { + "epoch": 1.2728969909949484, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.058443546295166, + "learning_rate": 1e-06, + "loss": 1.0527, + "mean_token_accuracy": 0.6787025928497314, + "num_tokens": 299877965.0, + "step": 11591 + }, + { + "epoch": 1.273006808697562, + "ewc_loss": 0.00010585784912109375, + "grad_norm": 2.0857491493225098, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7295259237289429, + "num_tokens": 299907103.0, + "step": 11592 + }, + { + "epoch": 1.2731166264001756, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0808722972869873, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.7048226594924927, + "num_tokens": 299939433.0, + "step": 11593 + }, + { + "epoch": 1.2732264441027894, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.21297550201416, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7431298494338989, + "num_tokens": 299965674.0, + "step": 11594 + }, + { + "epoch": 1.2733362618054032, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.4906225204467773, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7081549167633057, + "num_tokens": 299988125.0, + "step": 11595 + }, + { + "epoch": 1.2734460795080167, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.585466146469116, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7397589683532715, + "num_tokens": 300007609.0, + "step": 11596 + }, + { + "epoch": 1.2735558972106302, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.97416090965271, + "learning_rate": 1e-06, + "loss": 1.0836, + "mean_token_accuracy": 0.6741334795951843, + "num_tokens": 300042810.0, + "step": 11597 + }, + { + "epoch": 1.273665714913244, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.42573881149292, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7098554372787476, + "num_tokens": 300065000.0, + "step": 11598 + }, + { + "epoch": 1.2737755326158577, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.13897705078125, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7268402576446533, + "num_tokens": 300093292.0, + "step": 11599 + }, + { + "epoch": 1.2738853503184713, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2637243270874023, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7321707606315613, + "num_tokens": 300116894.0, + "step": 11600 + }, + { + "epoch": 1.273995168021085, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.572962522506714, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.706378698348999, + "num_tokens": 300138074.0, + "step": 11601 + }, + { + "epoch": 1.2741049857236986, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0324888229370117, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7274380922317505, + "num_tokens": 300167547.0, + "step": 11602 + }, + { + "epoch": 1.2742148034263123, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.068108320236206, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7076317071914673, + "num_tokens": 300197864.0, + "step": 11603 + }, + { + "epoch": 1.274324621128926, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.34232759475708, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7058782577514648, + "num_tokens": 300223913.0, + "step": 11604 + }, + { + "epoch": 1.2744344388315396, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.101091146469116, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7227340936660767, + "num_tokens": 300253707.0, + "step": 11605 + }, + { + "epoch": 1.2745442565341534, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2895822525024414, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7320834994316101, + "num_tokens": 300277603.0, + "step": 11606 + }, + { + "epoch": 1.274654074236767, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0464534759521484, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7140218019485474, + "num_tokens": 300308542.0, + "step": 11607 + }, + { + "epoch": 1.2747638919393807, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.7519752979278564, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.719957709312439, + "num_tokens": 300326984.0, + "step": 11608 + }, + { + "epoch": 1.2748737096419944, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.165199041366577, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7196514010429382, + "num_tokens": 300353479.0, + "step": 11609 + }, + { + "epoch": 1.274983527344608, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3569719791412354, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6972247362136841, + "num_tokens": 300377842.0, + "step": 11610 + }, + { + "epoch": 1.2750933450472215, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.5961012840270996, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7189130187034607, + "num_tokens": 300399800.0, + "step": 11611 + }, + { + "epoch": 1.2752031627498353, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.5998148918151855, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.700185239315033, + "num_tokens": 300420136.0, + "step": 11612 + }, + { + "epoch": 1.275312980452449, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.354325294494629, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7368862628936768, + "num_tokens": 300443520.0, + "step": 11613 + }, + { + "epoch": 1.2754227981550625, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3041372299194336, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7328488230705261, + "num_tokens": 300468267.0, + "step": 11614 + }, + { + "epoch": 1.2755326158576763, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.15598464012146, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7101837396621704, + "num_tokens": 300498103.0, + "step": 11615 + }, + { + "epoch": 1.2756424335602898, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2672786712646484, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7324244976043701, + "num_tokens": 300521745.0, + "step": 11616 + }, + { + "epoch": 1.2757522512629036, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.093782901763916, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7257406711578369, + "num_tokens": 300550409.0, + "step": 11617 + }, + { + "epoch": 1.2758620689655173, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3747551441192627, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7294486165046692, + "num_tokens": 300573697.0, + "step": 11618 + }, + { + "epoch": 1.2759718866681309, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.4770562648773193, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7220079898834229, + "num_tokens": 300595692.0, + "step": 11619 + }, + { + "epoch": 1.2760817043707446, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1868293285369873, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7206228971481323, + "num_tokens": 300621855.0, + "step": 11620 + }, + { + "epoch": 1.2761915220733582, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3559229373931885, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.744160532951355, + "num_tokens": 300648529.0, + "step": 11621 + }, + { + "epoch": 1.276301339775972, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.111269235610962, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7060643434524536, + "num_tokens": 300680264.0, + "step": 11622 + }, + { + "epoch": 1.2764111574785855, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.501979112625122, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7146007418632507, + "num_tokens": 300703359.0, + "step": 11623 + }, + { + "epoch": 1.2765209751811992, + "ewc_loss": 0.0001068115234375, + "grad_norm": 3.0139238834381104, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7076483964920044, + "num_tokens": 300722624.0, + "step": 11624 + }, + { + "epoch": 1.2766307928838128, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.176020383834839, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7233186960220337, + "num_tokens": 300749852.0, + "step": 11625 + }, + { + "epoch": 1.2767406105864265, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.394963026046753, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.740847110748291, + "num_tokens": 300772727.0, + "step": 11626 + }, + { + "epoch": 1.2768504282890403, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.329955577850342, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7140752077102661, + "num_tokens": 300796476.0, + "step": 11627 + }, + { + "epoch": 1.2769602459916538, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.47137713432312, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7148104310035706, + "num_tokens": 300821612.0, + "step": 11628 + }, + { + "epoch": 1.2770700636942676, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2127678394317627, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7398880124092102, + "num_tokens": 300847114.0, + "step": 11629 + }, + { + "epoch": 1.277179881396881, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.57944655418396, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7170788645744324, + "num_tokens": 300867258.0, + "step": 11630 + }, + { + "epoch": 1.2772896990994949, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1052255630493164, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6882445812225342, + "num_tokens": 300895077.0, + "step": 11631 + }, + { + "epoch": 1.2773995168021086, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1055638790130615, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7049145698547363, + "num_tokens": 300926950.0, + "step": 11632 + }, + { + "epoch": 1.2775093345047221, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0647268295288086, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7221275568008423, + "num_tokens": 300954676.0, + "step": 11633 + }, + { + "epoch": 1.2776191522073357, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3386077880859375, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7223493456840515, + "num_tokens": 300979637.0, + "step": 11634 + }, + { + "epoch": 1.2777289699099494, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.089310646057129, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7065295577049255, + "num_tokens": 301008969.0, + "step": 11635 + }, + { + "epoch": 1.2778387876125632, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1559247970581055, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.7057997584342957, + "num_tokens": 301039515.0, + "step": 11636 + }, + { + "epoch": 1.2779486053151767, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.172358512878418, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7254899740219116, + "num_tokens": 301066854.0, + "step": 11637 + }, + { + "epoch": 1.2780584230177905, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.045461893081665, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7038114666938782, + "num_tokens": 301097544.0, + "step": 11638 + }, + { + "epoch": 1.278168240720404, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.9601908922195435, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7190802693367004, + "num_tokens": 301128404.0, + "step": 11639 + }, + { + "epoch": 1.2782780584230178, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2927350997924805, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7146339416503906, + "num_tokens": 301154182.0, + "step": 11640 + }, + { + "epoch": 1.2783878761256315, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.542937994003296, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7287344932556152, + "num_tokens": 301174498.0, + "step": 11641 + }, + { + "epoch": 1.278497693828245, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.33774995803833, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7215822339057922, + "num_tokens": 301198593.0, + "step": 11642 + }, + { + "epoch": 1.2786075115308588, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.4107861518859863, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7342016696929932, + "num_tokens": 301219877.0, + "step": 11643 + }, + { + "epoch": 1.2787173292334724, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.7051806449890137, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7426900863647461, + "num_tokens": 301237174.0, + "step": 11644 + }, + { + "epoch": 1.2788271469360861, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2273640632629395, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7253977060317993, + "num_tokens": 301264961.0, + "step": 11645 + }, + { + "epoch": 1.2789369646386999, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.244749069213867, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7226011753082275, + "num_tokens": 301288854.0, + "step": 11646 + }, + { + "epoch": 1.2790467823413134, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.120530366897583, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7222104072570801, + "num_tokens": 301317681.0, + "step": 11647 + }, + { + "epoch": 1.279156600043927, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.190804958343506, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7323793172836304, + "num_tokens": 301342976.0, + "step": 11648 + }, + { + "epoch": 1.2792664177465407, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0906946659088135, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7142904996871948, + "num_tokens": 301371968.0, + "step": 11649 + }, + { + "epoch": 1.2793762354491545, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1280293464660645, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7115362882614136, + "num_tokens": 301399863.0, + "step": 11650 + }, + { + "epoch": 1.279486053151768, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.5429155826568604, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7155606746673584, + "num_tokens": 301421566.0, + "step": 11651 + }, + { + "epoch": 1.2795958708543818, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2326478958129883, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7301947474479675, + "num_tokens": 301448732.0, + "step": 11652 + }, + { + "epoch": 1.2797056885569953, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3322970867156982, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6943411827087402, + "num_tokens": 301472964.0, + "step": 11653 + }, + { + "epoch": 1.279815506259609, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3068737983703613, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7199733257293701, + "num_tokens": 301499186.0, + "step": 11654 + }, + { + "epoch": 1.2799253239622228, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0835022926330566, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7188410758972168, + "num_tokens": 301526014.0, + "step": 11655 + }, + { + "epoch": 1.2800351416648363, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.259805202484131, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7075768709182739, + "num_tokens": 301551649.0, + "step": 11656 + }, + { + "epoch": 1.28014495936745, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.206913471221924, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6939547061920166, + "num_tokens": 301580729.0, + "step": 11657 + }, + { + "epoch": 1.2802547770700636, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0110127925872803, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6982396841049194, + "num_tokens": 301610803.0, + "step": 11658 + }, + { + "epoch": 1.2803645947726774, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.317195415496826, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7250680923461914, + "num_tokens": 301634563.0, + "step": 11659 + }, + { + "epoch": 1.2804744124752911, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.4437527656555176, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7320470809936523, + "num_tokens": 301655487.0, + "step": 11660 + }, + { + "epoch": 1.2805842301779047, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.8829079866409302, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.70240318775177, + "num_tokens": 301692870.0, + "step": 11661 + }, + { + "epoch": 1.2806940478805182, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.324105978012085, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7023775577545166, + "num_tokens": 301718774.0, + "step": 11662 + }, + { + "epoch": 1.280803865583132, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.0345497131347656, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.73935866355896, + "num_tokens": 301745583.0, + "step": 11663 + }, + { + "epoch": 1.2809136832857457, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.6651272773742676, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7196406126022339, + "num_tokens": 301766400.0, + "step": 11664 + }, + { + "epoch": 1.2810235009883593, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.37463641166687, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7345994114875793, + "num_tokens": 301788337.0, + "step": 11665 + }, + { + "epoch": 1.281133318690973, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3383917808532715, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7101142406463623, + "num_tokens": 301812213.0, + "step": 11666 + }, + { + "epoch": 1.2812431363935866, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.4131155014038086, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7393084168434143, + "num_tokens": 301834284.0, + "step": 11667 + }, + { + "epoch": 1.2813529540962003, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.9951618909835815, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7095428705215454, + "num_tokens": 301866714.0, + "step": 11668 + }, + { + "epoch": 1.281462771798814, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1435444355010986, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7177163362503052, + "num_tokens": 301895602.0, + "step": 11669 + }, + { + "epoch": 1.2815725895014276, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.2816450595855713, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7105554342269897, + "num_tokens": 301922347.0, + "step": 11670 + }, + { + "epoch": 1.2816824072040414, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.19952654838562, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7246150970458984, + "num_tokens": 301948200.0, + "step": 11671 + }, + { + "epoch": 1.281792224906655, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.382580041885376, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7167806625366211, + "num_tokens": 301972825.0, + "step": 11672 + }, + { + "epoch": 1.2819020426092687, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.630380392074585, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.729304313659668, + "num_tokens": 301992609.0, + "step": 11673 + }, + { + "epoch": 1.2820118603118824, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1221060752868652, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7059231400489807, + "num_tokens": 302019194.0, + "step": 11674 + }, + { + "epoch": 1.282121678014496, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3049237728118896, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7061589956283569, + "num_tokens": 302045526.0, + "step": 11675 + }, + { + "epoch": 1.2822314957171095, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3398354053497314, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7075564861297607, + "num_tokens": 302071242.0, + "step": 11676 + }, + { + "epoch": 1.2823413134197232, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.234713554382324, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7053483128547668, + "num_tokens": 302099235.0, + "step": 11677 + }, + { + "epoch": 1.282451131122337, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3112361431121826, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7158156633377075, + "num_tokens": 302125416.0, + "step": 11678 + }, + { + "epoch": 1.2825609488249505, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.083528518676758, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7146642208099365, + "num_tokens": 302155002.0, + "step": 11679 + }, + { + "epoch": 1.2826707665275643, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.38910174369812, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7164201140403748, + "num_tokens": 302177355.0, + "step": 11680 + }, + { + "epoch": 1.2827805842301778, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3501384258270264, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7285879254341125, + "num_tokens": 302199821.0, + "step": 11681 + }, + { + "epoch": 1.2828904019327916, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1725094318389893, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.711245059967041, + "num_tokens": 302228048.0, + "step": 11682 + }, + { + "epoch": 1.2830002196354053, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.340822458267212, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7332099676132202, + "num_tokens": 302250998.0, + "step": 11683 + }, + { + "epoch": 1.2831100373380189, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3119444847106934, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7349897027015686, + "num_tokens": 302275876.0, + "step": 11684 + }, + { + "epoch": 1.2832198550406326, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1279051303863525, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7240732908248901, + "num_tokens": 302303707.0, + "step": 11685 + }, + { + "epoch": 1.2833296727432462, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.148989677429199, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.741966962814331, + "num_tokens": 302329954.0, + "step": 11686 + }, + { + "epoch": 1.28343949044586, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.5450997352600098, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.726038932800293, + "num_tokens": 302352435.0, + "step": 11687 + }, + { + "epoch": 1.2835493081484735, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.9782525300979614, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7246812582015991, + "num_tokens": 302383080.0, + "step": 11688 + }, + { + "epoch": 1.2836591258510872, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.279240608215332, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.6989402770996094, + "num_tokens": 302407828.0, + "step": 11689 + }, + { + "epoch": 1.2837689435537007, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3156492710113525, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7172765731811523, + "num_tokens": 302434324.0, + "step": 11690 + }, + { + "epoch": 1.2838787612563145, + "ewc_loss": 0.0001068115234375, + "grad_norm": 1.980207920074463, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7185083627700806, + "num_tokens": 302466784.0, + "step": 11691 + }, + { + "epoch": 1.2839885789589283, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3699779510498047, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7391247749328613, + "num_tokens": 302490422.0, + "step": 11692 + }, + { + "epoch": 1.2840983966615418, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.1659586429595947, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.69319748878479, + "num_tokens": 302521864.0, + "step": 11693 + }, + { + "epoch": 1.2842082143641556, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.138162612915039, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7168891429901123, + "num_tokens": 302550059.0, + "step": 11694 + }, + { + "epoch": 1.284318032066769, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.3337106704711914, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7238812446594238, + "num_tokens": 302572257.0, + "step": 11695 + }, + { + "epoch": 1.2844278497693828, + "ewc_loss": 0.0001068115234375, + "grad_norm": 2.235234022140503, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7185819149017334, + "num_tokens": 302597184.0, + "step": 11696 + }, + { + "epoch": 1.2845376674719966, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1658565998077393, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.723059892654419, + "num_tokens": 302625370.0, + "step": 11697 + }, + { + "epoch": 1.2846474851746101, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.5533387660980225, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7137905359268188, + "num_tokens": 302647698.0, + "step": 11698 + }, + { + "epoch": 1.2847573028772237, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3428101539611816, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7222381830215454, + "num_tokens": 302675217.0, + "step": 11699 + }, + { + "epoch": 1.2848671205798374, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.253401517868042, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7125272750854492, + "num_tokens": 302701118.0, + "step": 11700 + }, + { + "epoch": 1.2849769382824512, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.656327486038208, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7063107490539551, + "num_tokens": 302721586.0, + "step": 11701 + }, + { + "epoch": 1.2850867559850647, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.166213274002075, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7193648815155029, + "num_tokens": 302749143.0, + "step": 11702 + }, + { + "epoch": 1.2851965736876785, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.388732433319092, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7211679220199585, + "num_tokens": 302771892.0, + "step": 11703 + }, + { + "epoch": 1.285306391390292, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2641358375549316, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7185957431793213, + "num_tokens": 302795399.0, + "step": 11704 + }, + { + "epoch": 1.2854162090929058, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2405483722686768, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7287275791168213, + "num_tokens": 302820330.0, + "step": 11705 + }, + { + "epoch": 1.2855260267955195, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.300553321838379, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7016459703445435, + "num_tokens": 302844470.0, + "step": 11706 + }, + { + "epoch": 1.285635844498133, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.090885877609253, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6910144090652466, + "num_tokens": 302875201.0, + "step": 11707 + }, + { + "epoch": 1.2857456622007468, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2520294189453125, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.72947758436203, + "num_tokens": 302900116.0, + "step": 11708 + }, + { + "epoch": 1.2858554799033604, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.254392147064209, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7014011144638062, + "num_tokens": 302926066.0, + "step": 11709 + }, + { + "epoch": 1.2859652976059741, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.338181972503662, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.718206524848938, + "num_tokens": 302951977.0, + "step": 11710 + }, + { + "epoch": 1.2860751153085879, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3629305362701416, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7148738503456116, + "num_tokens": 302976941.0, + "step": 11711 + }, + { + "epoch": 1.2861849330112014, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4965391159057617, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7207417488098145, + "num_tokens": 302999808.0, + "step": 11712 + }, + { + "epoch": 1.286294750713815, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9898284673690796, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7451062202453613, + "num_tokens": 303030124.0, + "step": 11713 + }, + { + "epoch": 1.2864045684164287, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9952675104141235, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7295384407043457, + "num_tokens": 303062520.0, + "step": 11714 + }, + { + "epoch": 1.2865143861190425, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4020800590515137, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7176882028579712, + "num_tokens": 303087211.0, + "step": 11715 + }, + { + "epoch": 1.286624203821656, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.12977933883667, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7231623530387878, + "num_tokens": 303114010.0, + "step": 11716 + }, + { + "epoch": 1.2867340215242697, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.8216428756713867, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.739503026008606, + "num_tokens": 303132223.0, + "step": 11717 + }, + { + "epoch": 1.2868438392268833, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3924789428710938, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.72695392370224, + "num_tokens": 303154590.0, + "step": 11718 + }, + { + "epoch": 1.286953656929497, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1217970848083496, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7084975242614746, + "num_tokens": 303181959.0, + "step": 11719 + }, + { + "epoch": 1.2870634746321108, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4108731746673584, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7387369871139526, + "num_tokens": 303206582.0, + "step": 11720 + }, + { + "epoch": 1.2871732923347243, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.35196852684021, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7099300622940063, + "num_tokens": 303230746.0, + "step": 11721 + }, + { + "epoch": 1.287283110037338, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1497809886932373, + "learning_rate": 1e-06, + "loss": 0.8003, + "mean_token_accuracy": 0.7466717958450317, + "num_tokens": 303256952.0, + "step": 11722 + }, + { + "epoch": 1.2873929277399516, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.6377854347229004, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.735063910484314, + "num_tokens": 303276667.0, + "step": 11723 + }, + { + "epoch": 1.2875027454425654, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0994975566864014, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7350802421569824, + "num_tokens": 303302982.0, + "step": 11724 + }, + { + "epoch": 1.2876125631451791, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3885397911071777, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7092883586883545, + "num_tokens": 303330334.0, + "step": 11725 + }, + { + "epoch": 1.2877223808477927, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.228675127029419, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6944286823272705, + "num_tokens": 303358251.0, + "step": 11726 + }, + { + "epoch": 1.2878321985504062, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.260338068008423, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.742845892906189, + "num_tokens": 303381982.0, + "step": 11727 + }, + { + "epoch": 1.28794201625302, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.591456890106201, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7347973585128784, + "num_tokens": 303401346.0, + "step": 11728 + }, + { + "epoch": 1.2880518339556337, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1777167320251465, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7109307646751404, + "num_tokens": 303426493.0, + "step": 11729 + }, + { + "epoch": 1.2881616516582473, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.070772171020508, + "learning_rate": 1e-06, + "loss": 1.0521, + "mean_token_accuracy": 0.6817585825920105, + "num_tokens": 303458971.0, + "step": 11730 + }, + { + "epoch": 1.288271469360861, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3755617141723633, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7438899278640747, + "num_tokens": 303482654.0, + "step": 11731 + }, + { + "epoch": 1.2883812870634745, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9302854537963867, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7238578796386719, + "num_tokens": 303515731.0, + "step": 11732 + }, + { + "epoch": 1.2884911047660883, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2022368907928467, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7362931966781616, + "num_tokens": 303540606.0, + "step": 11733 + }, + { + "epoch": 1.288600922468702, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4028682708740234, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7236553430557251, + "num_tokens": 303562877.0, + "step": 11734 + }, + { + "epoch": 1.2887107401713156, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1738598346710205, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.703892707824707, + "num_tokens": 303590329.0, + "step": 11735 + }, + { + "epoch": 1.2888205578739294, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9770996570587158, + "learning_rate": 1e-06, + "loss": 1.0562, + "mean_token_accuracy": 0.6778947114944458, + "num_tokens": 303622275.0, + "step": 11736 + }, + { + "epoch": 1.2889303755765429, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.020359754562378, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7028526067733765, + "num_tokens": 303652835.0, + "step": 11737 + }, + { + "epoch": 1.2890401932791566, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.284284830093384, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7204517126083374, + "num_tokens": 303677543.0, + "step": 11738 + }, + { + "epoch": 1.2891500109817704, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1646780967712402, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7142423391342163, + "num_tokens": 303705579.0, + "step": 11739 + }, + { + "epoch": 1.289259828684384, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3211638927459717, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7325824499130249, + "num_tokens": 303727515.0, + "step": 11740 + }, + { + "epoch": 1.2893696463869975, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0933494567871094, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7168428897857666, + "num_tokens": 303757209.0, + "step": 11741 + }, + { + "epoch": 1.2894794640896112, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1798925399780273, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.701336145401001, + "num_tokens": 303785195.0, + "step": 11742 + }, + { + "epoch": 1.289589281792225, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.5976295471191406, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7166194319725037, + "num_tokens": 303804918.0, + "step": 11743 + }, + { + "epoch": 1.2896990994948385, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2609264850616455, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7092347145080566, + "num_tokens": 303829718.0, + "step": 11744 + }, + { + "epoch": 1.2898089171974523, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1355361938476562, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7100530862808228, + "num_tokens": 303858466.0, + "step": 11745 + }, + { + "epoch": 1.2899187349000658, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.318937063217163, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7458717823028564, + "num_tokens": 303881926.0, + "step": 11746 + }, + { + "epoch": 1.2900285526026796, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.193952798843384, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7297095060348511, + "num_tokens": 303908711.0, + "step": 11747 + }, + { + "epoch": 1.2901383703052933, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.228012800216675, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7069625854492188, + "num_tokens": 303934977.0, + "step": 11748 + }, + { + "epoch": 1.2902481880079069, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0714309215545654, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7236578464508057, + "num_tokens": 303962166.0, + "step": 11749 + }, + { + "epoch": 1.2903580057105206, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0321578979492188, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7120653986930847, + "num_tokens": 303993221.0, + "step": 11750 + }, + { + "epoch": 1.2904678234131342, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.414717674255371, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6966227293014526, + "num_tokens": 304018005.0, + "step": 11751 + }, + { + "epoch": 1.290577641115748, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.302528142929077, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7243232727050781, + "num_tokens": 304042860.0, + "step": 11752 + }, + { + "epoch": 1.2906874588183614, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0818703174591064, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7097511887550354, + "num_tokens": 304072245.0, + "step": 11753 + }, + { + "epoch": 1.2907972765209752, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.257326364517212, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7198294997215271, + "num_tokens": 304096862.0, + "step": 11754 + }, + { + "epoch": 1.2909070942235887, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.143129587173462, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7188516855239868, + "num_tokens": 304124158.0, + "step": 11755 + }, + { + "epoch": 1.2910169119262025, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2858288288116455, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6933398246765137, + "num_tokens": 304152468.0, + "step": 11756 + }, + { + "epoch": 1.2911267296288162, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3878774642944336, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7198473811149597, + "num_tokens": 304175812.0, + "step": 11757 + }, + { + "epoch": 1.2912365473314298, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3462085723876953, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7365753054618835, + "num_tokens": 304200028.0, + "step": 11758 + }, + { + "epoch": 1.2913463650340435, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.41412615776062, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7322326302528381, + "num_tokens": 304224531.0, + "step": 11759 + }, + { + "epoch": 1.291456182736657, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9832022190093994, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6935491561889648, + "num_tokens": 304256347.0, + "step": 11760 + }, + { + "epoch": 1.2915660004392708, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2648723125457764, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7040140628814697, + "num_tokens": 304281763.0, + "step": 11761 + }, + { + "epoch": 1.2916758181418846, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.419792890548706, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7561264038085938, + "num_tokens": 304304235.0, + "step": 11762 + }, + { + "epoch": 1.2917856358444981, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.329859733581543, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7251024842262268, + "num_tokens": 304329706.0, + "step": 11763 + }, + { + "epoch": 1.2918954535471117, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.096360445022583, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7336481213569641, + "num_tokens": 304360223.0, + "step": 11764 + }, + { + "epoch": 1.2920052712497254, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.328413248062134, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7297794818878174, + "num_tokens": 304382198.0, + "step": 11765 + }, + { + "epoch": 1.2921150889523392, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 1.9315261840820312, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7318341135978699, + "num_tokens": 304415819.0, + "step": 11766 + }, + { + "epoch": 1.2922249066549527, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2060041427612305, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7091357111930847, + "num_tokens": 304443269.0, + "step": 11767 + }, + { + "epoch": 1.2923347243575665, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.380286693572998, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7017019391059875, + "num_tokens": 304467737.0, + "step": 11768 + }, + { + "epoch": 1.29244454206018, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.0267789363861084, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7061423063278198, + "num_tokens": 304498891.0, + "step": 11769 + }, + { + "epoch": 1.2925543597627938, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.561429500579834, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.719157338142395, + "num_tokens": 304521902.0, + "step": 11770 + }, + { + "epoch": 1.2926641774654075, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.355950355529785, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7203937768936157, + "num_tokens": 304546070.0, + "step": 11771 + }, + { + "epoch": 1.292773995168021, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.5090184211730957, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7156026363372803, + "num_tokens": 304567669.0, + "step": 11772 + }, + { + "epoch": 1.2928838128706348, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.149808406829834, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7161002159118652, + "num_tokens": 304594245.0, + "step": 11773 + }, + { + "epoch": 1.2929936305732483, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1927480697631836, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7278620600700378, + "num_tokens": 304622361.0, + "step": 11774 + }, + { + "epoch": 1.293103448275862, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.6783947944641113, + "learning_rate": 1e-06, + "loss": 0.7626, + "mean_token_accuracy": 0.7588486075401306, + "num_tokens": 304641523.0, + "step": 11775 + }, + { + "epoch": 1.2932132659784759, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2372093200683594, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7291791439056396, + "num_tokens": 304666763.0, + "step": 11776 + }, + { + "epoch": 1.2933230836810894, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2394299507141113, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7114537954330444, + "num_tokens": 304691674.0, + "step": 11777 + }, + { + "epoch": 1.293432901383703, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4586334228515625, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7236682176589966, + "num_tokens": 304713565.0, + "step": 11778 + }, + { + "epoch": 1.2935427190863167, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.274049997329712, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7099625468254089, + "num_tokens": 304738147.0, + "step": 11779 + }, + { + "epoch": 1.2936525367889304, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.328242301940918, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7049301862716675, + "num_tokens": 304763544.0, + "step": 11780 + }, + { + "epoch": 1.293762354491544, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1639795303344727, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6997135281562805, + "num_tokens": 304793019.0, + "step": 11781 + }, + { + "epoch": 1.2938721721941577, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.115060567855835, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7078789472579956, + "num_tokens": 304821249.0, + "step": 11782 + }, + { + "epoch": 1.2939819898967713, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1277835369110107, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6989405751228333, + "num_tokens": 304849658.0, + "step": 11783 + }, + { + "epoch": 1.294091807599385, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4046733379364014, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7220003604888916, + "num_tokens": 304872325.0, + "step": 11784 + }, + { + "epoch": 1.2942016253019988, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.229058265686035, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7459594011306763, + "num_tokens": 304896426.0, + "step": 11785 + }, + { + "epoch": 1.2943114430046123, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.39290189743042, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7028246521949768, + "num_tokens": 304920851.0, + "step": 11786 + }, + { + "epoch": 1.294421260707226, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.350849151611328, + "learning_rate": 1e-06, + "loss": 0.7646, + "mean_token_accuracy": 0.766228199005127, + "num_tokens": 304943334.0, + "step": 11787 + }, + { + "epoch": 1.2945310784098396, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.272372245788574, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.6998612284660339, + "num_tokens": 304969500.0, + "step": 11788 + }, + { + "epoch": 1.2946408961124534, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4112231731414795, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7232286930084229, + "num_tokens": 304991110.0, + "step": 11789 + }, + { + "epoch": 1.2947507138150671, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4549498558044434, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7189807891845703, + "num_tokens": 305015955.0, + "step": 11790 + }, + { + "epoch": 1.2948605315176807, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.1626627445220947, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7154731750488281, + "num_tokens": 305042520.0, + "step": 11791 + }, + { + "epoch": 1.2949703492202942, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2466375827789307, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7066552639007568, + "num_tokens": 305071471.0, + "step": 11792 + }, + { + "epoch": 1.295080166922908, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2793781757354736, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7495021224021912, + "num_tokens": 305095514.0, + "step": 11793 + }, + { + "epoch": 1.2951899846255217, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4092631340026855, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7062212824821472, + "num_tokens": 305117920.0, + "step": 11794 + }, + { + "epoch": 1.2952998023281352, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.464357376098633, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7458175420761108, + "num_tokens": 305139672.0, + "step": 11795 + }, + { + "epoch": 1.295409620030749, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2689106464385986, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7286636233329773, + "num_tokens": 305164213.0, + "step": 11796 + }, + { + "epoch": 1.2955194377333625, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.366187334060669, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7127352356910706, + "num_tokens": 305188419.0, + "step": 11797 + }, + { + "epoch": 1.2956292554359763, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.127159357070923, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7083125114440918, + "num_tokens": 305217696.0, + "step": 11798 + }, + { + "epoch": 1.29573907313859, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.2270586490631104, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7203144431114197, + "num_tokens": 305244888.0, + "step": 11799 + }, + { + "epoch": 1.2958488908412036, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.128627300262451, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7488523721694946, + "num_tokens": 305271407.0, + "step": 11800 + }, + { + "epoch": 1.2959587085438173, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4546866416931152, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7204996347427368, + "num_tokens": 305293901.0, + "step": 11801 + }, + { + "epoch": 1.2960685262464309, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.247279405593872, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7163840532302856, + "num_tokens": 305319982.0, + "step": 11802 + }, + { + "epoch": 1.2961783439490446, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.4314098358154297, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7219822406768799, + "num_tokens": 305341512.0, + "step": 11803 + }, + { + "epoch": 1.2962881616516582, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.197059154510498, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7036950588226318, + "num_tokens": 305369414.0, + "step": 11804 + }, + { + "epoch": 1.296397979354272, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.3973567485809326, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7371026873588562, + "num_tokens": 305392887.0, + "step": 11805 + }, + { + "epoch": 1.2965077970568855, + "ewc_loss": 0.00010728836059570312, + "grad_norm": 2.005790948867798, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.692894458770752, + "num_tokens": 305423865.0, + "step": 11806 + }, + { + "epoch": 1.2966176147594992, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 1.9947872161865234, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7051029801368713, + "num_tokens": 305454469.0, + "step": 11807 + }, + { + "epoch": 1.296727432462113, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.091146945953369, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7184012532234192, + "num_tokens": 305483821.0, + "step": 11808 + }, + { + "epoch": 1.2968372501647265, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.402561664581299, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7347837686538696, + "num_tokens": 305506685.0, + "step": 11809 + }, + { + "epoch": 1.2969470678673403, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.35231614112854, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7350362539291382, + "num_tokens": 305530489.0, + "step": 11810 + }, + { + "epoch": 1.2970568855699538, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.3899896144866943, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7323159575462341, + "num_tokens": 305552683.0, + "step": 11811 + }, + { + "epoch": 1.2971667032725676, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.376476764678955, + "learning_rate": 1e-06, + "loss": 0.7321, + "mean_token_accuracy": 0.7647274732589722, + "num_tokens": 305575005.0, + "step": 11812 + }, + { + "epoch": 1.2972765209751813, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.3251168727874756, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7345167398452759, + "num_tokens": 305601169.0, + "step": 11813 + }, + { + "epoch": 1.2973863386777948, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.285759925842285, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7321797609329224, + "num_tokens": 305628213.0, + "step": 11814 + }, + { + "epoch": 1.2974961563804086, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.1864659786224365, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.711641252040863, + "num_tokens": 305655998.0, + "step": 11815 + }, + { + "epoch": 1.2976059740830221, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.186694383621216, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7066822052001953, + "num_tokens": 305682178.0, + "step": 11816 + }, + { + "epoch": 1.297715791785636, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.4565136432647705, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7080978155136108, + "num_tokens": 305706363.0, + "step": 11817 + }, + { + "epoch": 1.2978256094882494, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 1.9255869388580322, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7243545055389404, + "num_tokens": 305736478.0, + "step": 11818 + }, + { + "epoch": 1.2979354271908632, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.194885492324829, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7126734256744385, + "num_tokens": 305765312.0, + "step": 11819 + }, + { + "epoch": 1.2980452448934767, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.32094669342041, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7391626238822937, + "num_tokens": 305788731.0, + "step": 11820 + }, + { + "epoch": 1.2981550625960905, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.2591140270233154, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6933383345603943, + "num_tokens": 305814487.0, + "step": 11821 + }, + { + "epoch": 1.2982648802987042, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.258580207824707, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7383499145507812, + "num_tokens": 305840110.0, + "step": 11822 + }, + { + "epoch": 1.2983746980013178, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.0389883518218994, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7228201627731323, + "num_tokens": 305870635.0, + "step": 11823 + }, + { + "epoch": 1.2984845157039315, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.107943058013916, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.7053555250167847, + "num_tokens": 305900461.0, + "step": 11824 + }, + { + "epoch": 1.298594333406545, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.487668514251709, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7220267653465271, + "num_tokens": 305920842.0, + "step": 11825 + }, + { + "epoch": 1.2987041511091588, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.230252265930176, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7028539776802063, + "num_tokens": 305947118.0, + "step": 11826 + }, + { + "epoch": 1.2988139688117726, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.0863776206970215, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7232599854469299, + "num_tokens": 305973592.0, + "step": 11827 + }, + { + "epoch": 1.2989237865143861, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.2626633644104004, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7197322845458984, + "num_tokens": 306000339.0, + "step": 11828 + }, + { + "epoch": 1.2990336042169996, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 3.066340923309326, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7118489742279053, + "num_tokens": 306018889.0, + "step": 11829 + }, + { + "epoch": 1.2991434219196134, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.4837090969085693, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7223420739173889, + "num_tokens": 306041685.0, + "step": 11830 + }, + { + "epoch": 1.2992532396222272, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.1367030143737793, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7107328176498413, + "num_tokens": 306072866.0, + "step": 11831 + }, + { + "epoch": 1.2993630573248407, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.2329471111297607, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7167210578918457, + "num_tokens": 306100711.0, + "step": 11832 + }, + { + "epoch": 1.2994728750274545, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.55379581451416, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.6995375156402588, + "num_tokens": 306122537.0, + "step": 11833 + }, + { + "epoch": 1.299582692730068, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.4144020080566406, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7244486212730408, + "num_tokens": 306145470.0, + "step": 11834 + }, + { + "epoch": 1.2996925104326817, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.2690045833587646, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7102985382080078, + "num_tokens": 306171805.0, + "step": 11835 + }, + { + "epoch": 1.2998023281352955, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.1751046180725098, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7328392863273621, + "num_tokens": 306196865.0, + "step": 11836 + }, + { + "epoch": 1.299912145837909, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.0819294452667236, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7202502489089966, + "num_tokens": 306224739.0, + "step": 11837 + }, + { + "epoch": 1.3000219635405228, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.3683292865753174, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7026513814926147, + "num_tokens": 306248787.0, + "step": 11838 + }, + { + "epoch": 1.3001317812431363, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.180053949356079, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7113978862762451, + "num_tokens": 306274373.0, + "step": 11839 + }, + { + "epoch": 1.30024159894575, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.2063748836517334, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7176840305328369, + "num_tokens": 306299504.0, + "step": 11840 + }, + { + "epoch": 1.3003514166483638, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.0426745414733887, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.70757657289505, + "num_tokens": 306328299.0, + "step": 11841 + }, + { + "epoch": 1.3004612343509774, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.4658422470092773, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7362444996833801, + "num_tokens": 306351122.0, + "step": 11842 + }, + { + "epoch": 1.300571052053591, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.484773635864258, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7343958616256714, + "num_tokens": 306371039.0, + "step": 11843 + }, + { + "epoch": 1.3006808697562047, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.1476693153381348, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7156551480293274, + "num_tokens": 306398171.0, + "step": 11844 + }, + { + "epoch": 1.3007906874588184, + "ewc_loss": 0.00010776519775390625, + "grad_norm": 2.109529972076416, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7178124189376831, + "num_tokens": 306426726.0, + "step": 11845 + }, + { + "epoch": 1.300900505161432, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0838091373443604, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6965687870979309, + "num_tokens": 306459056.0, + "step": 11846 + }, + { + "epoch": 1.3010103228640457, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.056748151779175, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7037144899368286, + "num_tokens": 306488203.0, + "step": 11847 + }, + { + "epoch": 1.3011201405666593, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2384121417999268, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7388585209846497, + "num_tokens": 306514586.0, + "step": 11848 + }, + { + "epoch": 1.301229958269273, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2929720878601074, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7325873374938965, + "num_tokens": 306541307.0, + "step": 11849 + }, + { + "epoch": 1.3013397759718868, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.168287754058838, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7273610830307007, + "num_tokens": 306568190.0, + "step": 11850 + }, + { + "epoch": 1.3014495936745003, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.056671380996704, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.721537709236145, + "num_tokens": 306598005.0, + "step": 11851 + }, + { + "epoch": 1.301559411377114, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.156067132949829, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7243401408195496, + "num_tokens": 306626348.0, + "step": 11852 + }, + { + "epoch": 1.3016692290797276, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.5251424312591553, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7139247059822083, + "num_tokens": 306648604.0, + "step": 11853 + }, + { + "epoch": 1.3017790467823414, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.1487672328948975, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7071206569671631, + "num_tokens": 306677033.0, + "step": 11854 + }, + { + "epoch": 1.301888864484955, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.109473466873169, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7287074327468872, + "num_tokens": 306704591.0, + "step": 11855 + }, + { + "epoch": 1.3019986821875686, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.5661213397979736, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7319854497909546, + "num_tokens": 306722762.0, + "step": 11856 + }, + { + "epoch": 1.3021084998901822, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 1.9713622331619263, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7041444182395935, + "num_tokens": 306755451.0, + "step": 11857 + }, + { + "epoch": 1.302218317592796, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.4191596508026123, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7469682097434998, + "num_tokens": 306777213.0, + "step": 11858 + }, + { + "epoch": 1.3023281352954097, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 1.9440255165100098, + "learning_rate": 1e-06, + "loss": 1.036, + "mean_token_accuracy": 0.6893285512924194, + "num_tokens": 306811483.0, + "step": 11859 + }, + { + "epoch": 1.3024379529980232, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2738025188446045, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7164255380630493, + "num_tokens": 306835767.0, + "step": 11860 + }, + { + "epoch": 1.302547770700637, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.337156057357788, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.734081506729126, + "num_tokens": 306858057.0, + "step": 11861 + }, + { + "epoch": 1.3026575884032505, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.453587293624878, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7383694648742676, + "num_tokens": 306881002.0, + "step": 11862 + }, + { + "epoch": 1.3027674061058643, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2289185523986816, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7126408815383911, + "num_tokens": 306907122.0, + "step": 11863 + }, + { + "epoch": 1.302877223808478, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.298628568649292, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7232141494750977, + "num_tokens": 306931095.0, + "step": 11864 + }, + { + "epoch": 1.3029870415110916, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.4026706218719482, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7278810143470764, + "num_tokens": 306953333.0, + "step": 11865 + }, + { + "epoch": 1.3030968592137053, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.3463287353515625, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7128908634185791, + "num_tokens": 306979310.0, + "step": 11866 + }, + { + "epoch": 1.3032066769163189, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.278135299682617, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7252035140991211, + "num_tokens": 307005402.0, + "step": 11867 + }, + { + "epoch": 1.3033164946189326, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.259080171585083, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7176575064659119, + "num_tokens": 307031279.0, + "step": 11868 + }, + { + "epoch": 1.3034263123215462, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.4754979610443115, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7258110046386719, + "num_tokens": 307054553.0, + "step": 11869 + }, + { + "epoch": 1.30353613002416, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.1918277740478516, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7088371515274048, + "num_tokens": 307081607.0, + "step": 11870 + }, + { + "epoch": 1.3036459477267734, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.1391499042510986, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7161993980407715, + "num_tokens": 307108952.0, + "step": 11871 + }, + { + "epoch": 1.3037557654293872, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.5926833152770996, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7104243636131287, + "num_tokens": 307129054.0, + "step": 11872 + }, + { + "epoch": 1.303865583132001, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0820674896240234, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7442585229873657, + "num_tokens": 307155838.0, + "step": 11873 + }, + { + "epoch": 1.3039754008346145, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.1550068855285645, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7031915783882141, + "num_tokens": 307182441.0, + "step": 11874 + }, + { + "epoch": 1.3040852185372283, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.5808846950531006, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7263267636299133, + "num_tokens": 307204790.0, + "step": 11875 + }, + { + "epoch": 1.3041950362398418, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2747087478637695, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7106046676635742, + "num_tokens": 307230438.0, + "step": 11876 + }, + { + "epoch": 1.3043048539424555, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0840611457824707, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7072759866714478, + "num_tokens": 307262556.0, + "step": 11877 + }, + { + "epoch": 1.3044146716450693, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.564439058303833, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7032203078269958, + "num_tokens": 307283816.0, + "step": 11878 + }, + { + "epoch": 1.3045244893476828, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.649329423904419, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7102218866348267, + "num_tokens": 307304357.0, + "step": 11879 + }, + { + "epoch": 1.3046343070502964, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.2616658210754395, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.712005078792572, + "num_tokens": 307329642.0, + "step": 11880 + }, + { + "epoch": 1.3047441247529101, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0339577198028564, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7241654992103577, + "num_tokens": 307357631.0, + "step": 11881 + }, + { + "epoch": 1.3048539424555239, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.155163526535034, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7176980972290039, + "num_tokens": 307384204.0, + "step": 11882 + }, + { + "epoch": 1.3049637601581374, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.1380350589752197, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7009460926055908, + "num_tokens": 307411603.0, + "step": 11883 + }, + { + "epoch": 1.3050735778607512, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0983095169067383, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6858009099960327, + "num_tokens": 307440959.0, + "step": 11884 + }, + { + "epoch": 1.3051833955633647, + "ewc_loss": 0.00010824203491210938, + "grad_norm": 2.0269081592559814, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.6976010799407959, + "num_tokens": 307472194.0, + "step": 11885 + }, + { + "epoch": 1.3052932132659785, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.0389320850372314, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.719584584236145, + "num_tokens": 307502651.0, + "step": 11886 + }, + { + "epoch": 1.3054030309685922, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4750401973724365, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7286511063575745, + "num_tokens": 307525708.0, + "step": 11887 + }, + { + "epoch": 1.3055128486712058, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.3171212673187256, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7047629356384277, + "num_tokens": 307551774.0, + "step": 11888 + }, + { + "epoch": 1.3056226663738195, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.149578094482422, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7280470132827759, + "num_tokens": 307577389.0, + "step": 11889 + }, + { + "epoch": 1.305732484076433, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.131725311279297, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7123003005981445, + "num_tokens": 307605984.0, + "step": 11890 + }, + { + "epoch": 1.3058423017790468, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.5306975841522217, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7224723100662231, + "num_tokens": 307627950.0, + "step": 11891 + }, + { + "epoch": 1.3059521194816606, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.2552504539489746, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7153153419494629, + "num_tokens": 307653276.0, + "step": 11892 + }, + { + "epoch": 1.306061937184274, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.1308681964874268, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7055473327636719, + "num_tokens": 307681643.0, + "step": 11893 + }, + { + "epoch": 1.3061717548868876, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.1425864696502686, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7321821451187134, + "num_tokens": 307707993.0, + "step": 11894 + }, + { + "epoch": 1.3062815725895014, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4808287620544434, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7123449444770813, + "num_tokens": 307730284.0, + "step": 11895 + }, + { + "epoch": 1.3063913902921152, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.56412410736084, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7379684448242188, + "num_tokens": 307750876.0, + "step": 11896 + }, + { + "epoch": 1.3065012079947287, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4828147888183594, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7359920740127563, + "num_tokens": 307771352.0, + "step": 11897 + }, + { + "epoch": 1.3066110256973424, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.256673812866211, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6960406303405762, + "num_tokens": 307799490.0, + "step": 11898 + }, + { + "epoch": 1.306720843399956, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.478499412536621, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.741487979888916, + "num_tokens": 307822092.0, + "step": 11899 + }, + { + "epoch": 1.3068306611025697, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.238258123397827, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7197073101997375, + "num_tokens": 307847578.0, + "step": 11900 + }, + { + "epoch": 1.3069404788051835, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.02420973777771, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.708726167678833, + "num_tokens": 307879184.0, + "step": 11901 + }, + { + "epoch": 1.307050296507797, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.353515625, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7055224180221558, + "num_tokens": 307902946.0, + "step": 11902 + }, + { + "epoch": 1.3071601142104108, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.414454460144043, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7008653879165649, + "num_tokens": 307925745.0, + "step": 11903 + }, + { + "epoch": 1.3072699319130243, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.421755313873291, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7038456797599792, + "num_tokens": 307950738.0, + "step": 11904 + }, + { + "epoch": 1.307379749615638, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.083264112472534, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7087394595146179, + "num_tokens": 307980167.0, + "step": 11905 + }, + { + "epoch": 1.3074895673182518, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.5822718143463135, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7188035249710083, + "num_tokens": 308003106.0, + "step": 11906 + }, + { + "epoch": 1.3075993850208654, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.141366958618164, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7258819341659546, + "num_tokens": 308029924.0, + "step": 11907 + }, + { + "epoch": 1.307709202723479, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.3154001235961914, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7043800354003906, + "num_tokens": 308053551.0, + "step": 11908 + }, + { + "epoch": 1.3078190204260927, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4649667739868164, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7248021364212036, + "num_tokens": 308075839.0, + "step": 11909 + }, + { + "epoch": 1.3079288381287064, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.386974334716797, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7428616285324097, + "num_tokens": 308098003.0, + "step": 11910 + }, + { + "epoch": 1.30803865583132, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 1.9969615936279297, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7063287496566772, + "num_tokens": 308131484.0, + "step": 11911 + }, + { + "epoch": 1.3081484735339337, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4471449851989746, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.722994327545166, + "num_tokens": 308152707.0, + "step": 11912 + }, + { + "epoch": 1.3082582912365472, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.1714322566986084, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7156124114990234, + "num_tokens": 308183132.0, + "step": 11913 + }, + { + "epoch": 1.308368108939161, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.695950508117676, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7478945255279541, + "num_tokens": 308201717.0, + "step": 11914 + }, + { + "epoch": 1.3084779266417748, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.5677413940429688, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7230666875839233, + "num_tokens": 308221777.0, + "step": 11915 + }, + { + "epoch": 1.3085877443443883, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.0526857376098633, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7122257947921753, + "num_tokens": 308252221.0, + "step": 11916 + }, + { + "epoch": 1.308697562047002, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.3794054985046387, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7280374765396118, + "num_tokens": 308274420.0, + "step": 11917 + }, + { + "epoch": 1.3088073797496156, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 1.9284976720809937, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6908325552940369, + "num_tokens": 308309469.0, + "step": 11918 + }, + { + "epoch": 1.3089171974522293, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.162771224975586, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7373082637786865, + "num_tokens": 308334493.0, + "step": 11919 + }, + { + "epoch": 1.309027015154843, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.2945172786712646, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7118784785270691, + "num_tokens": 308362169.0, + "step": 11920 + }, + { + "epoch": 1.3091368328574566, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.450538396835327, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7319923043251038, + "num_tokens": 308383792.0, + "step": 11921 + }, + { + "epoch": 1.3092466505600702, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.3922646045684814, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7201412916183472, + "num_tokens": 308408941.0, + "step": 11922 + }, + { + "epoch": 1.309356468262684, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.2167749404907227, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.7028098106384277, + "num_tokens": 308437407.0, + "step": 11923 + }, + { + "epoch": 1.3094662859652977, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.1263654232025146, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7081534266471863, + "num_tokens": 308464422.0, + "step": 11924 + }, + { + "epoch": 1.3095761036679112, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.168257713317871, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6957510113716125, + "num_tokens": 308493901.0, + "step": 11925 + }, + { + "epoch": 1.309685921370525, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 1.9672033786773682, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7215564250946045, + "num_tokens": 308527056.0, + "step": 11926 + }, + { + "epoch": 1.3097957390731385, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.020991325378418, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7140195965766907, + "num_tokens": 308559454.0, + "step": 11927 + }, + { + "epoch": 1.3099055567757523, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.415128231048584, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.723331868648529, + "num_tokens": 308582588.0, + "step": 11928 + }, + { + "epoch": 1.310015374478366, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.359060287475586, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7133871912956238, + "num_tokens": 308607198.0, + "step": 11929 + }, + { + "epoch": 1.3101251921809796, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.269411563873291, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7127348184585571, + "num_tokens": 308633119.0, + "step": 11930 + }, + { + "epoch": 1.3102350098835933, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.3168444633483887, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7137984037399292, + "num_tokens": 308658039.0, + "step": 11931 + }, + { + "epoch": 1.3103448275862069, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.4154105186462402, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7418959736824036, + "num_tokens": 308678716.0, + "step": 11932 + }, + { + "epoch": 1.3104546452888206, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.0705273151397705, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7108013033866882, + "num_tokens": 308708883.0, + "step": 11933 + }, + { + "epoch": 1.3105644629914341, + "ewc_loss": 0.00010919570922851562, + "grad_norm": 2.2346630096435547, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7262197732925415, + "num_tokens": 308733671.0, + "step": 11934 + }, + { + "epoch": 1.310674280694048, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2693965435028076, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.711178183555603, + "num_tokens": 308760280.0, + "step": 11935 + }, + { + "epoch": 1.3107840983966614, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.516084909439087, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7343082427978516, + "num_tokens": 308780609.0, + "step": 11936 + }, + { + "epoch": 1.3108939160992752, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.362492322921753, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7030004262924194, + "num_tokens": 308804795.0, + "step": 11937 + }, + { + "epoch": 1.311003733801889, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.5679709911346436, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7265815734863281, + "num_tokens": 308827372.0, + "step": 11938 + }, + { + "epoch": 1.3111135515045025, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2363734245300293, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7297229766845703, + "num_tokens": 308851157.0, + "step": 11939 + }, + { + "epoch": 1.3112233692071162, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.327035427093506, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7167017459869385, + "num_tokens": 308876515.0, + "step": 11940 + }, + { + "epoch": 1.3113331869097298, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4561991691589355, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7028652429580688, + "num_tokens": 308898561.0, + "step": 11941 + }, + { + "epoch": 1.3114430046123435, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0323526859283447, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7160214185714722, + "num_tokens": 308928792.0, + "step": 11942 + }, + { + "epoch": 1.3115528223149573, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.107731580734253, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7356163263320923, + "num_tokens": 308955868.0, + "step": 11943 + }, + { + "epoch": 1.3116626400175708, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.316711187362671, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7182361483573914, + "num_tokens": 308979800.0, + "step": 11944 + }, + { + "epoch": 1.3117724577201844, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.132560968399048, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7103806734085083, + "num_tokens": 309009514.0, + "step": 11945 + }, + { + "epoch": 1.3118822754227981, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.5949184894561768, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7217139005661011, + "num_tokens": 309030237.0, + "step": 11946 + }, + { + "epoch": 1.3119920931254119, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2011055946350098, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7282027006149292, + "num_tokens": 309058352.0, + "step": 11947 + }, + { + "epoch": 1.3121019108280254, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.534611463546753, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7205191850662231, + "num_tokens": 309078411.0, + "step": 11948 + }, + { + "epoch": 1.3122117285306392, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4004180431365967, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7391726970672607, + "num_tokens": 309099362.0, + "step": 11949 + }, + { + "epoch": 1.3123215462332527, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.513734817504883, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7390328645706177, + "num_tokens": 309120392.0, + "step": 11950 + }, + { + "epoch": 1.3124313639358665, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.07308030128479, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7258514165878296, + "num_tokens": 309149883.0, + "step": 11951 + }, + { + "epoch": 1.3125411816384802, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2423906326293945, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7125017642974854, + "num_tokens": 309176404.0, + "step": 11952 + }, + { + "epoch": 1.3126509993410937, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9854410886764526, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7034251689910889, + "num_tokens": 309204974.0, + "step": 11953 + }, + { + "epoch": 1.3127608170437075, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.086130142211914, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7206164598464966, + "num_tokens": 309233538.0, + "step": 11954 + }, + { + "epoch": 1.312870634746321, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.979988694190979, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7076501846313477, + "num_tokens": 309265874.0, + "step": 11955 + }, + { + "epoch": 1.3129804524489348, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3486695289611816, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.693403959274292, + "num_tokens": 309292134.0, + "step": 11956 + }, + { + "epoch": 1.3130902701515486, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.470153331756592, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7391940355300903, + "num_tokens": 309314202.0, + "step": 11957 + }, + { + "epoch": 1.313200087854162, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9141803979873657, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7207566499710083, + "num_tokens": 309345545.0, + "step": 11958 + }, + { + "epoch": 1.3133099055567756, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.45228910446167, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7049276828765869, + "num_tokens": 309368933.0, + "step": 11959 + }, + { + "epoch": 1.3134197232593894, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1225481033325195, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7159165740013123, + "num_tokens": 309395945.0, + "step": 11960 + }, + { + "epoch": 1.3135295409620031, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.294497013092041, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.703882098197937, + "num_tokens": 309421344.0, + "step": 11961 + }, + { + "epoch": 1.3136393586646167, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.46266508102417, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7132430672645569, + "num_tokens": 309444483.0, + "step": 11962 + }, + { + "epoch": 1.3137491763672304, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3832802772521973, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7213387489318848, + "num_tokens": 309468167.0, + "step": 11963 + }, + { + "epoch": 1.313858994069844, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9438663721084595, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7065349817276001, + "num_tokens": 309503149.0, + "step": 11964 + }, + { + "epoch": 1.3139688117724577, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2376708984375, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7357281446456909, + "num_tokens": 309528160.0, + "step": 11965 + }, + { + "epoch": 1.3140786294750715, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.092543840408325, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.727087676525116, + "num_tokens": 309556572.0, + "step": 11966 + }, + { + "epoch": 1.314188447177685, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2576913833618164, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7229899764060974, + "num_tokens": 309582611.0, + "step": 11967 + }, + { + "epoch": 1.3142982648802988, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3034818172454834, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7409685850143433, + "num_tokens": 309609761.0, + "step": 11968 + }, + { + "epoch": 1.3144080825829123, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.156344413757324, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7163639068603516, + "num_tokens": 309635992.0, + "step": 11969 + }, + { + "epoch": 1.314517900285526, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4933745861053467, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.717300534248352, + "num_tokens": 309659002.0, + "step": 11970 + }, + { + "epoch": 1.3146277179881398, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1357855796813965, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7078607678413391, + "num_tokens": 309689015.0, + "step": 11971 + }, + { + "epoch": 1.3147375356907534, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2795069217681885, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6961880922317505, + "num_tokens": 309715038.0, + "step": 11972 + }, + { + "epoch": 1.314847353393367, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4034841060638428, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7135041952133179, + "num_tokens": 309739071.0, + "step": 11973 + }, + { + "epoch": 1.3149571710959806, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2036430835723877, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.693245530128479, + "num_tokens": 309769186.0, + "step": 11974 + }, + { + "epoch": 1.3150669887985944, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.102642774581909, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7190922498703003, + "num_tokens": 309800157.0, + "step": 11975 + }, + { + "epoch": 1.315176806501208, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.217026710510254, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.696490466594696, + "num_tokens": 309827522.0, + "step": 11976 + }, + { + "epoch": 1.3152866242038217, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.131438970565796, + "learning_rate": 1e-06, + "loss": 1.0203, + "mean_token_accuracy": 0.6929820775985718, + "num_tokens": 309855405.0, + "step": 11977 + }, + { + "epoch": 1.3153964419064352, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0250580310821533, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7145624756813049, + "num_tokens": 309885220.0, + "step": 11978 + }, + { + "epoch": 1.315506259609049, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2482755184173584, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7286806106567383, + "num_tokens": 309912524.0, + "step": 11979 + }, + { + "epoch": 1.3156160773116627, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.010817289352417, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6889747381210327, + "num_tokens": 309944880.0, + "step": 11980 + }, + { + "epoch": 1.3157258950142763, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3637619018554688, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7061085104942322, + "num_tokens": 309971327.0, + "step": 11981 + }, + { + "epoch": 1.31583571271689, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4806129932403564, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7035120129585266, + "num_tokens": 309996491.0, + "step": 11982 + }, + { + "epoch": 1.3159455304195036, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.638566493988037, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7369316220283508, + "num_tokens": 310016335.0, + "step": 11983 + }, + { + "epoch": 1.3160553481221173, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9247182607650757, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7038474678993225, + "num_tokens": 310051258.0, + "step": 11984 + }, + { + "epoch": 1.3161651658247309, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9885433912277222, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7121618986129761, + "num_tokens": 310084879.0, + "step": 11985 + }, + { + "epoch": 1.3162749835273446, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.193574905395508, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7391018867492676, + "num_tokens": 310110181.0, + "step": 11986 + }, + { + "epoch": 1.3163848012299582, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2449939250946045, + "learning_rate": 1e-06, + "loss": 0.7833, + "mean_token_accuracy": 0.7509248852729797, + "num_tokens": 310132279.0, + "step": 11987 + }, + { + "epoch": 1.316494618932572, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3061671257019043, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7325801253318787, + "num_tokens": 310154905.0, + "step": 11988 + }, + { + "epoch": 1.3166044366351857, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.232546806335449, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7201701402664185, + "num_tokens": 310182554.0, + "step": 11989 + }, + { + "epoch": 1.3167142543377992, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0424258708953857, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7199850082397461, + "num_tokens": 310216138.0, + "step": 11990 + }, + { + "epoch": 1.316824072040413, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3485333919525146, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7192003130912781, + "num_tokens": 310243173.0, + "step": 11991 + }, + { + "epoch": 1.3169338897430265, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3797707557678223, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7204256653785706, + "num_tokens": 310266769.0, + "step": 11992 + }, + { + "epoch": 1.3170437074456403, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0400750637054443, + "learning_rate": 1e-06, + "loss": 1.0285, + "mean_token_accuracy": 0.6942260265350342, + "num_tokens": 310299363.0, + "step": 11993 + }, + { + "epoch": 1.317153525148254, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.083514928817749, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6937092542648315, + "num_tokens": 310327603.0, + "step": 11994 + }, + { + "epoch": 1.3172633428508675, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3078227043151855, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7226540446281433, + "num_tokens": 310352373.0, + "step": 11995 + }, + { + "epoch": 1.3173731605534813, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2185118198394775, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7302337288856506, + "num_tokens": 310378023.0, + "step": 11996 + }, + { + "epoch": 1.3174829782560948, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9647554159164429, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7014097571372986, + "num_tokens": 310408772.0, + "step": 11997 + }, + { + "epoch": 1.3175927959587086, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1754214763641357, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7128466367721558, + "num_tokens": 310437185.0, + "step": 11998 + }, + { + "epoch": 1.3177026136613221, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.8813765048980713, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7219351530075073, + "num_tokens": 310472318.0, + "step": 11999 + }, + { + "epoch": 1.3178124313639359, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1162822246551514, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7360315918922424, + "num_tokens": 310501902.0, + "step": 12000 + }, + { + "epoch": 1.3179222490665494, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.454540729522705, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7514050602912903, + "num_tokens": 310523070.0, + "step": 12001 + }, + { + "epoch": 1.3180320667691632, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.180999517440796, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.717374861240387, + "num_tokens": 310548975.0, + "step": 12002 + }, + { + "epoch": 1.318141884471777, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.9628897905349731, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.754256546497345, + "num_tokens": 310578440.0, + "step": 12003 + }, + { + "epoch": 1.3182517021743905, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1580116748809814, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.7010806798934937, + "num_tokens": 310608050.0, + "step": 12004 + }, + { + "epoch": 1.3183615198770042, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.220855951309204, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.717886209487915, + "num_tokens": 310634990.0, + "step": 12005 + }, + { + "epoch": 1.3184713375796178, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.185213565826416, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7180936336517334, + "num_tokens": 310663671.0, + "step": 12006 + }, + { + "epoch": 1.3185811552822315, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3809807300567627, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.743966817855835, + "num_tokens": 310686032.0, + "step": 12007 + }, + { + "epoch": 1.3186909729848453, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.107100486755371, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.712617814540863, + "num_tokens": 310715174.0, + "step": 12008 + }, + { + "epoch": 1.3188007906874588, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.296198844909668, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.6974029541015625, + "num_tokens": 310741269.0, + "step": 12009 + }, + { + "epoch": 1.3189106083900723, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.5922985076904297, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7355015277862549, + "num_tokens": 310761641.0, + "step": 12010 + }, + { + "epoch": 1.319020426092686, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3080854415893555, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7296521663665771, + "num_tokens": 310784991.0, + "step": 12011 + }, + { + "epoch": 1.3191302437952999, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.5268924236297607, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6999639272689819, + "num_tokens": 310807560.0, + "step": 12012 + }, + { + "epoch": 1.3192400614979134, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.191681385040283, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7235209941864014, + "num_tokens": 310833251.0, + "step": 12013 + }, + { + "epoch": 1.3193498792005272, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 1.970452904701233, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7156498432159424, + "num_tokens": 310865625.0, + "step": 12014 + }, + { + "epoch": 1.3194596969031407, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.375638961791992, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7193269729614258, + "num_tokens": 310889811.0, + "step": 12015 + }, + { + "epoch": 1.3195695146057544, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.692800283432007, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7216047048568726, + "num_tokens": 310908529.0, + "step": 12016 + }, + { + "epoch": 1.3196793323083682, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0062382221221924, + "learning_rate": 1e-06, + "loss": 1.0596, + "mean_token_accuracy": 0.6816083192825317, + "num_tokens": 310940143.0, + "step": 12017 + }, + { + "epoch": 1.3197891500109817, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.54022216796875, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7240535020828247, + "num_tokens": 310960261.0, + "step": 12018 + }, + { + "epoch": 1.3198989677135955, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.465507745742798, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7160775661468506, + "num_tokens": 310983025.0, + "step": 12019 + }, + { + "epoch": 1.320008785416209, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2529633045196533, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7096136808395386, + "num_tokens": 311008978.0, + "step": 12020 + }, + { + "epoch": 1.3201186031188228, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1464998722076416, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7289639711380005, + "num_tokens": 311037576.0, + "step": 12021 + }, + { + "epoch": 1.3202284208214365, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.393091917037964, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.6990893483161926, + "num_tokens": 311061893.0, + "step": 12022 + }, + { + "epoch": 1.32033823852405, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.5128774642944336, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7380962371826172, + "num_tokens": 311084726.0, + "step": 12023 + }, + { + "epoch": 1.3204480562266636, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.1832478046417236, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7244314551353455, + "num_tokens": 311110099.0, + "step": 12024 + }, + { + "epoch": 1.3205578739292774, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.7039029598236084, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7374166250228882, + "num_tokens": 311130690.0, + "step": 12025 + }, + { + "epoch": 1.3206676916318911, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.2794482707977295, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6939818263053894, + "num_tokens": 311157366.0, + "step": 12026 + }, + { + "epoch": 1.3207775093345047, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4787673950195312, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7103732228279114, + "num_tokens": 311179686.0, + "step": 12027 + }, + { + "epoch": 1.3208873270371184, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.384284496307373, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7057867646217346, + "num_tokens": 311203656.0, + "step": 12028 + }, + { + "epoch": 1.320997144739732, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.170759439468384, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7160594463348389, + "num_tokens": 311230369.0, + "step": 12029 + }, + { + "epoch": 1.3211069624423457, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.0046908855438232, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7358483076095581, + "num_tokens": 311260256.0, + "step": 12030 + }, + { + "epoch": 1.3212167801449595, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.4876327514648438, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7024438381195068, + "num_tokens": 311282552.0, + "step": 12031 + }, + { + "epoch": 1.321326597847573, + "ewc_loss": 0.00010967254638671875, + "grad_norm": 2.3354666233062744, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7422313690185547, + "num_tokens": 311304267.0, + "step": 12032 + }, + { + "epoch": 1.3214364155501868, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2340824604034424, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7021785974502563, + "num_tokens": 311330001.0, + "step": 12033 + }, + { + "epoch": 1.3215462332528003, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.508404016494751, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7186006903648376, + "num_tokens": 311355827.0, + "step": 12034 + }, + { + "epoch": 1.321656050955414, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.3725500106811523, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7231173515319824, + "num_tokens": 311379536.0, + "step": 12035 + }, + { + "epoch": 1.3217658686580278, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.275259494781494, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7244246006011963, + "num_tokens": 311404203.0, + "step": 12036 + }, + { + "epoch": 1.3218756863606413, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2546679973602295, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7277169823646545, + "num_tokens": 311429686.0, + "step": 12037 + }, + { + "epoch": 1.3219855040632549, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.073469400405884, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.711919903755188, + "num_tokens": 311459298.0, + "step": 12038 + }, + { + "epoch": 1.3220953217658686, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.4683737754821777, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7476154565811157, + "num_tokens": 311479998.0, + "step": 12039 + }, + { + "epoch": 1.3222051394684824, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.1479012966156006, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7012981176376343, + "num_tokens": 311508414.0, + "step": 12040 + }, + { + "epoch": 1.322314957171096, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.345177412033081, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7334186434745789, + "num_tokens": 311531813.0, + "step": 12041 + }, + { + "epoch": 1.3224247748737097, + "ewc_loss": 0.000110626220703125, + "grad_norm": 1.998877763748169, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7196460962295532, + "num_tokens": 311565499.0, + "step": 12042 + }, + { + "epoch": 1.3225345925763232, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.6134419441223145, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7187728881835938, + "num_tokens": 311585634.0, + "step": 12043 + }, + { + "epoch": 1.322644410278937, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.113954782485962, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7193584442138672, + "num_tokens": 311612145.0, + "step": 12044 + }, + { + "epoch": 1.3227542279815507, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.1878812313079834, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7009973526000977, + "num_tokens": 311637952.0, + "step": 12045 + }, + { + "epoch": 1.3228640456841643, + "ewc_loss": 0.000110626220703125, + "grad_norm": 1.9507486820220947, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7295107245445251, + "num_tokens": 311667932.0, + "step": 12046 + }, + { + "epoch": 1.322973863386778, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2696995735168457, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7313175797462463, + "num_tokens": 311691471.0, + "step": 12047 + }, + { + "epoch": 1.3230836810893916, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.620002508163452, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7274290323257446, + "num_tokens": 311710910.0, + "step": 12048 + }, + { + "epoch": 1.3231934987920053, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.411050319671631, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7396289706230164, + "num_tokens": 311733474.0, + "step": 12049 + }, + { + "epoch": 1.3233033164946189, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.5941238403320312, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7145666480064392, + "num_tokens": 311754050.0, + "step": 12050 + }, + { + "epoch": 1.3234131341972326, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2895936965942383, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7034068703651428, + "num_tokens": 311782576.0, + "step": 12051 + }, + { + "epoch": 1.3235229518998461, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.131894826889038, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7306966781616211, + "num_tokens": 311810613.0, + "step": 12052 + }, + { + "epoch": 1.32363276960246, + "ewc_loss": 0.000110626220703125, + "grad_norm": 1.971976399421692, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6940065622329712, + "num_tokens": 311845004.0, + "step": 12053 + }, + { + "epoch": 1.3237425873050737, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.5268568992614746, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7425245046615601, + "num_tokens": 311865498.0, + "step": 12054 + }, + { + "epoch": 1.3238524050076872, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.023446798324585, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7108491659164429, + "num_tokens": 311895993.0, + "step": 12055 + }, + { + "epoch": 1.323962222710301, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2153215408325195, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7204939723014832, + "num_tokens": 311922674.0, + "step": 12056 + }, + { + "epoch": 1.3240720404129145, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.486673593521118, + "learning_rate": 1e-06, + "loss": 0.7545, + "mean_token_accuracy": 0.7582064867019653, + "num_tokens": 311944228.0, + "step": 12057 + }, + { + "epoch": 1.3241818581155282, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.3175742626190186, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7148958444595337, + "num_tokens": 311967283.0, + "step": 12058 + }, + { + "epoch": 1.324291675818142, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.6674070358276367, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7261946201324463, + "num_tokens": 311986322.0, + "step": 12059 + }, + { + "epoch": 1.3244014935207555, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.2969021797180176, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.717522919178009, + "num_tokens": 312011426.0, + "step": 12060 + }, + { + "epoch": 1.324511311223369, + "ewc_loss": 0.000110626220703125, + "grad_norm": 2.0734684467315674, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.696875810623169, + "num_tokens": 312043022.0, + "step": 12061 + }, + { + "epoch": 1.3246211289259828, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1462056636810303, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.695853590965271, + "num_tokens": 312073029.0, + "step": 12062 + }, + { + "epoch": 1.3247309466285966, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3371567726135254, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7301425933837891, + "num_tokens": 312098938.0, + "step": 12063 + }, + { + "epoch": 1.3248407643312101, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.484534978866577, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7182316780090332, + "num_tokens": 312122336.0, + "step": 12064 + }, + { + "epoch": 1.3249505820338239, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.377269983291626, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7078542113304138, + "num_tokens": 312145917.0, + "step": 12065 + }, + { + "epoch": 1.3250603997364374, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.5365242958068848, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.705060601234436, + "num_tokens": 312167339.0, + "step": 12066 + }, + { + "epoch": 1.3251702174390512, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3567922115325928, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7266991138458252, + "num_tokens": 312189994.0, + "step": 12067 + }, + { + "epoch": 1.325280035141665, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2022225856781006, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7158989906311035, + "num_tokens": 312220602.0, + "step": 12068 + }, + { + "epoch": 1.3253898528442785, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3245840072631836, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7144542932510376, + "num_tokens": 312244048.0, + "step": 12069 + }, + { + "epoch": 1.3254996705468922, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.125948905944824, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7118645310401917, + "num_tokens": 312272462.0, + "step": 12070 + }, + { + "epoch": 1.3256094882495058, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.0466725826263428, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6907999515533447, + "num_tokens": 312304011.0, + "step": 12071 + }, + { + "epoch": 1.3257193059521195, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 1.9732849597930908, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7076343297958374, + "num_tokens": 312339018.0, + "step": 12072 + }, + { + "epoch": 1.3258291236547333, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2516908645629883, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7092384099960327, + "num_tokens": 312365181.0, + "step": 12073 + }, + { + "epoch": 1.3259389413573468, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.954055070877075, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7431707382202148, + "num_tokens": 312380926.0, + "step": 12074 + }, + { + "epoch": 1.3260487590599603, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1204686164855957, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7235896587371826, + "num_tokens": 312408282.0, + "step": 12075 + }, + { + "epoch": 1.326158576762574, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.299281597137451, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7310534715652466, + "num_tokens": 312432842.0, + "step": 12076 + }, + { + "epoch": 1.3262683944651878, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1748430728912354, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7317183017730713, + "num_tokens": 312457932.0, + "step": 12077 + }, + { + "epoch": 1.3263782121678014, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3192873001098633, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7487039566040039, + "num_tokens": 312482510.0, + "step": 12078 + }, + { + "epoch": 1.3264880298704151, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 1.8474713563919067, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6943918466567993, + "num_tokens": 312522526.0, + "step": 12079 + }, + { + "epoch": 1.3265978475730287, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2638742923736572, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.70916748046875, + "num_tokens": 312547872.0, + "step": 12080 + }, + { + "epoch": 1.3267076652756424, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.548322916030884, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7481844425201416, + "num_tokens": 312567048.0, + "step": 12081 + }, + { + "epoch": 1.3268174829782562, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.456012010574341, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7207425832748413, + "num_tokens": 312590450.0, + "step": 12082 + }, + { + "epoch": 1.3269273006808697, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.068763256072998, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.702394425868988, + "num_tokens": 312621839.0, + "step": 12083 + }, + { + "epoch": 1.3270371183834835, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.193570852279663, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.6992151737213135, + "num_tokens": 312649493.0, + "step": 12084 + }, + { + "epoch": 1.327146936086097, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1726584434509277, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7080276608467102, + "num_tokens": 312678801.0, + "step": 12085 + }, + { + "epoch": 1.3272567537887108, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3035829067230225, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7320016026496887, + "num_tokens": 312702894.0, + "step": 12086 + }, + { + "epoch": 1.3273665714913245, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2606923580169678, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7169856429100037, + "num_tokens": 312726396.0, + "step": 12087 + }, + { + "epoch": 1.327476389193938, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.0427727699279785, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7040401697158813, + "num_tokens": 312755864.0, + "step": 12088 + }, + { + "epoch": 1.3275862068965516, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3521602153778076, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6974356174468994, + "num_tokens": 312780319.0, + "step": 12089 + }, + { + "epoch": 1.3276960245991654, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1318647861480713, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7166281938552856, + "num_tokens": 312808674.0, + "step": 12090 + }, + { + "epoch": 1.3278058423017791, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.25390887260437, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.702893853187561, + "num_tokens": 312834242.0, + "step": 12091 + }, + { + "epoch": 1.3279156600043927, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3927159309387207, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7325869202613831, + "num_tokens": 312856847.0, + "step": 12092 + }, + { + "epoch": 1.3280254777070064, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1188952922821045, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7025736570358276, + "num_tokens": 312885152.0, + "step": 12093 + }, + { + "epoch": 1.32813529540962, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.5114638805389404, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7113724946975708, + "num_tokens": 312907134.0, + "step": 12094 + }, + { + "epoch": 1.3282451131122337, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.606290340423584, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7405242919921875, + "num_tokens": 312929375.0, + "step": 12095 + }, + { + "epoch": 1.3283549308148475, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.152970552444458, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.6888421773910522, + "num_tokens": 312959592.0, + "step": 12096 + }, + { + "epoch": 1.328464748517461, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1982409954071045, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7275486588478088, + "num_tokens": 312984791.0, + "step": 12097 + }, + { + "epoch": 1.3285745662200747, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3777904510498047, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7367417216300964, + "num_tokens": 313008554.0, + "step": 12098 + }, + { + "epoch": 1.3286843839226883, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1393442153930664, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.711150050163269, + "num_tokens": 313036100.0, + "step": 12099 + }, + { + "epoch": 1.328794201625302, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3475513458251953, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7183842658996582, + "num_tokens": 313060742.0, + "step": 12100 + }, + { + "epoch": 1.3289040193279158, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2599775791168213, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7279502749443054, + "num_tokens": 313085469.0, + "step": 12101 + }, + { + "epoch": 1.3290138370305293, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.125331163406372, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7086440324783325, + "num_tokens": 313112793.0, + "step": 12102 + }, + { + "epoch": 1.3291236547331429, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.5547287464141846, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7359558343887329, + "num_tokens": 313134516.0, + "step": 12103 + }, + { + "epoch": 1.3292334724357566, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.1424036026000977, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7139856815338135, + "num_tokens": 313163023.0, + "step": 12104 + }, + { + "epoch": 1.3293432901383704, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 1.9564943313598633, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7221923470497131, + "num_tokens": 313196724.0, + "step": 12105 + }, + { + "epoch": 1.329453107840984, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 1.9866411685943604, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7143624424934387, + "num_tokens": 313227346.0, + "step": 12106 + }, + { + "epoch": 1.3295629255435977, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.166571855545044, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7097103595733643, + "num_tokens": 313255908.0, + "step": 12107 + }, + { + "epoch": 1.3296727432462112, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.2701656818389893, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7153701186180115, + "num_tokens": 313282227.0, + "step": 12108 + }, + { + "epoch": 1.329782560948825, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.429577589035034, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7309476137161255, + "num_tokens": 313303207.0, + "step": 12109 + }, + { + "epoch": 1.3298923786514387, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.4097795486450195, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7055422067642212, + "num_tokens": 313325047.0, + "step": 12110 + }, + { + "epoch": 1.3300021963540523, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.0572750568389893, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6888014078140259, + "num_tokens": 313356491.0, + "step": 12111 + }, + { + "epoch": 1.330112014056666, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 1.9873907566070557, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7044054269790649, + "num_tokens": 313389447.0, + "step": 12112 + }, + { + "epoch": 1.3302218317592795, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.323417901992798, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7153183817863464, + "num_tokens": 313415681.0, + "step": 12113 + }, + { + "epoch": 1.3303316494618933, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.312258720397949, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7252826690673828, + "num_tokens": 313440534.0, + "step": 12114 + }, + { + "epoch": 1.3304414671645068, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.0198025703430176, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7164533138275146, + "num_tokens": 313472238.0, + "step": 12115 + }, + { + "epoch": 1.3305512848671206, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.3100790977478027, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7108015418052673, + "num_tokens": 313496901.0, + "step": 12116 + }, + { + "epoch": 1.3306611025697341, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.33518385887146, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.6998692750930786, + "num_tokens": 313521690.0, + "step": 12117 + }, + { + "epoch": 1.330770920272348, + "ewc_loss": 0.00011110305786132812, + "grad_norm": 2.249262809753418, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.721468985080719, + "num_tokens": 313546813.0, + "step": 12118 + }, + { + "epoch": 1.3308807379749616, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.131847620010376, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7259808778762817, + "num_tokens": 313574997.0, + "step": 12119 + }, + { + "epoch": 1.3309905556775752, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 1.9977446794509888, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7056441307067871, + "num_tokens": 313607098.0, + "step": 12120 + }, + { + "epoch": 1.331100373380189, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.3346595764160156, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.747344970703125, + "num_tokens": 313630093.0, + "step": 12121 + }, + { + "epoch": 1.3312101910828025, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.1674954891204834, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7193763852119446, + "num_tokens": 313657207.0, + "step": 12122 + }, + { + "epoch": 1.3313200087854162, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.094165086746216, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7112786769866943, + "num_tokens": 313686686.0, + "step": 12123 + }, + { + "epoch": 1.33142982648803, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 1.9465413093566895, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7097779512405396, + "num_tokens": 313718020.0, + "step": 12124 + }, + { + "epoch": 1.3315396441906435, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.1551990509033203, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7266870737075806, + "num_tokens": 313744954.0, + "step": 12125 + }, + { + "epoch": 1.331649461893257, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.35451340675354, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7020623683929443, + "num_tokens": 313767651.0, + "step": 12126 + }, + { + "epoch": 1.3317592795958708, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.2201602458953857, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7203303575515747, + "num_tokens": 313793030.0, + "step": 12127 + }, + { + "epoch": 1.3318690972984846, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.2082982063293457, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7158767580986023, + "num_tokens": 313819612.0, + "step": 12128 + }, + { + "epoch": 1.331978915001098, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.1121511459350586, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7291178703308105, + "num_tokens": 313847069.0, + "step": 12129 + }, + { + "epoch": 1.3320887327037119, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.144636869430542, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7295331954956055, + "num_tokens": 313875458.0, + "step": 12130 + }, + { + "epoch": 1.3321985504063254, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.4817705154418945, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7306629419326782, + "num_tokens": 313896527.0, + "step": 12131 + }, + { + "epoch": 1.3323083681089392, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.4210753440856934, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7171750068664551, + "num_tokens": 313918697.0, + "step": 12132 + }, + { + "epoch": 1.332418185811553, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.147243022918701, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7182247042655945, + "num_tokens": 313946077.0, + "step": 12133 + }, + { + "epoch": 1.3325280035141664, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.2058374881744385, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.709181547164917, + "num_tokens": 313972542.0, + "step": 12134 + }, + { + "epoch": 1.3326378212167802, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 1.948254942893982, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7252941131591797, + "num_tokens": 314003809.0, + "step": 12135 + }, + { + "epoch": 1.3327476389193937, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.2457475662231445, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6930602788925171, + "num_tokens": 314030491.0, + "step": 12136 + }, + { + "epoch": 1.3328574566220075, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 1.9537054300308228, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7224112749099731, + "num_tokens": 314062339.0, + "step": 12137 + }, + { + "epoch": 1.3329672743246213, + "ewc_loss": 0.00011205673217773438, + "grad_norm": 2.1237988471984863, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7320135235786438, + "num_tokens": 314089700.0, + "step": 12138 + }, + { + "epoch": 1.3330770920272348, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1233816146850586, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7067634463310242, + "num_tokens": 314120020.0, + "step": 12139 + }, + { + "epoch": 1.3331869097298483, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2959024906158447, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7363621592521667, + "num_tokens": 314143549.0, + "step": 12140 + }, + { + "epoch": 1.333296727432462, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.0328800678253174, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7200571298599243, + "num_tokens": 314173193.0, + "step": 12141 + }, + { + "epoch": 1.3334065451350758, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.109196901321411, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7137932777404785, + "num_tokens": 314200701.0, + "step": 12142 + }, + { + "epoch": 1.3335163628376894, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2393417358398438, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7289683818817139, + "num_tokens": 314224618.0, + "step": 12143 + }, + { + "epoch": 1.3336261805403031, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.087409257888794, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7029706239700317, + "num_tokens": 314254988.0, + "step": 12144 + }, + { + "epoch": 1.3337359982429167, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.180691719055176, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7056782841682434, + "num_tokens": 314282456.0, + "step": 12145 + }, + { + "epoch": 1.3338458159455304, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.4461138248443604, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6999735832214355, + "num_tokens": 314305152.0, + "step": 12146 + }, + { + "epoch": 1.3339556336481442, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2052900791168213, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7137212753295898, + "num_tokens": 314333641.0, + "step": 12147 + }, + { + "epoch": 1.3340654513507577, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 1.9657461643218994, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.701264500617981, + "num_tokens": 314368978.0, + "step": 12148 + }, + { + "epoch": 1.3341752690533715, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1279563903808594, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7307727932929993, + "num_tokens": 314394530.0, + "step": 12149 + }, + { + "epoch": 1.334285086755985, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.4338746070861816, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7170254588127136, + "num_tokens": 314417140.0, + "step": 12150 + }, + { + "epoch": 1.3343949044585988, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3449361324310303, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7136087417602539, + "num_tokens": 314440596.0, + "step": 12151 + }, + { + "epoch": 1.3345047221612125, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2861945629119873, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7280318737030029, + "num_tokens": 314462690.0, + "step": 12152 + }, + { + "epoch": 1.334614539863826, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.4527153968811035, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7078915238380432, + "num_tokens": 314485297.0, + "step": 12153 + }, + { + "epoch": 1.3347243575664396, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3439505100250244, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.716771125793457, + "num_tokens": 314511086.0, + "step": 12154 + }, + { + "epoch": 1.3348341752690533, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3487565517425537, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7438557147979736, + "num_tokens": 314534220.0, + "step": 12155 + }, + { + "epoch": 1.334943992971667, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.235541820526123, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7039347290992737, + "num_tokens": 314559951.0, + "step": 12156 + }, + { + "epoch": 1.3350538106742806, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.303393602371216, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7217344045639038, + "num_tokens": 314584385.0, + "step": 12157 + }, + { + "epoch": 1.3351636283768944, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3544931411743164, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7315924167633057, + "num_tokens": 314608107.0, + "step": 12158 + }, + { + "epoch": 1.335273446079508, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.206235408782959, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6870906949043274, + "num_tokens": 314636606.0, + "step": 12159 + }, + { + "epoch": 1.3353832637821217, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.295665979385376, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7279431223869324, + "num_tokens": 314660635.0, + "step": 12160 + }, + { + "epoch": 1.3354930814847354, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1791062355041504, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7258565425872803, + "num_tokens": 314687780.0, + "step": 12161 + }, + { + "epoch": 1.335602899187349, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2289416790008545, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7337968349456787, + "num_tokens": 314713069.0, + "step": 12162 + }, + { + "epoch": 1.3357127168899627, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.275719404220581, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7153940200805664, + "num_tokens": 314737852.0, + "step": 12163 + }, + { + "epoch": 1.3358225345925763, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2488136291503906, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6890957951545715, + "num_tokens": 314764199.0, + "step": 12164 + }, + { + "epoch": 1.33593235229519, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2826783657073975, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7264602184295654, + "num_tokens": 314788466.0, + "step": 12165 + }, + { + "epoch": 1.3360421699978036, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.261871814727783, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7060263156890869, + "num_tokens": 314814228.0, + "step": 12166 + }, + { + "epoch": 1.3361519877004173, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 1.9187103509902954, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6870681047439575, + "num_tokens": 314847990.0, + "step": 12167 + }, + { + "epoch": 1.3362618054030309, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3752894401550293, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.737608790397644, + "num_tokens": 314871778.0, + "step": 12168 + }, + { + "epoch": 1.3363716231056446, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.5204813480377197, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7173438668251038, + "num_tokens": 314894790.0, + "step": 12169 + }, + { + "epoch": 1.3364814408082584, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.384652614593506, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7146559357643127, + "num_tokens": 314918890.0, + "step": 12170 + }, + { + "epoch": 1.336591258510872, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2530288696289062, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7276219129562378, + "num_tokens": 314944008.0, + "step": 12171 + }, + { + "epoch": 1.3367010762134857, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 1.9176286458969116, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7028367519378662, + "num_tokens": 314979013.0, + "step": 12172 + }, + { + "epoch": 1.3368108939160992, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.233140707015991, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7439395189285278, + "num_tokens": 315003209.0, + "step": 12173 + }, + { + "epoch": 1.336920711618713, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.166534185409546, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7348491549491882, + "num_tokens": 315029803.0, + "step": 12174 + }, + { + "epoch": 1.3370305293213267, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2775750160217285, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.7125491499900818, + "num_tokens": 315058104.0, + "step": 12175 + }, + { + "epoch": 1.3371403470239402, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.393246650695801, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7317009568214417, + "num_tokens": 315080546.0, + "step": 12176 + }, + { + "epoch": 1.337250164726554, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.3085055351257324, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7243491411209106, + "num_tokens": 315105139.0, + "step": 12177 + }, + { + "epoch": 1.3373599824291675, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.374047040939331, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7093217372894287, + "num_tokens": 315128254.0, + "step": 12178 + }, + { + "epoch": 1.3374698001317813, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.214555263519287, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7151365280151367, + "num_tokens": 315151826.0, + "step": 12179 + }, + { + "epoch": 1.3375796178343948, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1200520992279053, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7113925218582153, + "num_tokens": 315180659.0, + "step": 12180 + }, + { + "epoch": 1.3376894355370086, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2214667797088623, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7182797193527222, + "num_tokens": 315208419.0, + "step": 12181 + }, + { + "epoch": 1.3377992532396221, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.197390079498291, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.725800633430481, + "num_tokens": 315233736.0, + "step": 12182 + }, + { + "epoch": 1.3379090709422359, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.205249786376953, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7481224536895752, + "num_tokens": 315258558.0, + "step": 12183 + }, + { + "epoch": 1.3380188886448496, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.2746524810791016, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7153608202934265, + "num_tokens": 315284503.0, + "step": 12184 + }, + { + "epoch": 1.3381287063474632, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1296072006225586, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7128276228904724, + "num_tokens": 315314227.0, + "step": 12185 + }, + { + "epoch": 1.338238524050077, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.03662371635437, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7445876598358154, + "num_tokens": 315341861.0, + "step": 12186 + }, + { + "epoch": 1.3383483417526905, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.677497625350952, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7369206547737122, + "num_tokens": 315362198.0, + "step": 12187 + }, + { + "epoch": 1.3384581594553042, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.4276864528656006, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7129158973693848, + "num_tokens": 315384800.0, + "step": 12188 + }, + { + "epoch": 1.338567977157918, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.1670279502868652, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7206286787986755, + "num_tokens": 315412593.0, + "step": 12189 + }, + { + "epoch": 1.3386777948605315, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.26628041267395, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7211758494377136, + "num_tokens": 315437974.0, + "step": 12190 + }, + { + "epoch": 1.338787612563145, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1135811805725098, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6931784152984619, + "num_tokens": 315466659.0, + "step": 12191 + }, + { + "epoch": 1.3388974302657588, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 1.953405737876892, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7293776273727417, + "num_tokens": 315498402.0, + "step": 12192 + }, + { + "epoch": 1.3390072479683726, + "ewc_loss": 0.0001125335693359375, + "grad_norm": 2.427760124206543, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7283157706260681, + "num_tokens": 315521133.0, + "step": 12193 + }, + { + "epoch": 1.339117065670986, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2628204822540283, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7420822978019714, + "num_tokens": 315545156.0, + "step": 12194 + }, + { + "epoch": 1.3392268833735999, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1527020931243896, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7166646718978882, + "num_tokens": 315571491.0, + "step": 12195 + }, + { + "epoch": 1.3393367010762134, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.323394298553467, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7264612913131714, + "num_tokens": 315597388.0, + "step": 12196 + }, + { + "epoch": 1.3394465187788271, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.6018354892730713, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7257128357887268, + "num_tokens": 315618018.0, + "step": 12197 + }, + { + "epoch": 1.339556336481441, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2035937309265137, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7269711494445801, + "num_tokens": 315644214.0, + "step": 12198 + }, + { + "epoch": 1.3396661541840544, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.222395420074463, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7277045249938965, + "num_tokens": 315670775.0, + "step": 12199 + }, + { + "epoch": 1.3397759718866682, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5595757961273193, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7239885926246643, + "num_tokens": 315692463.0, + "step": 12200 + }, + { + "epoch": 1.3398857895892817, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3092405796051025, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6917808055877686, + "num_tokens": 315719338.0, + "step": 12201 + }, + { + "epoch": 1.3399956072918955, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1484782695770264, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6934831142425537, + "num_tokens": 315748717.0, + "step": 12202 + }, + { + "epoch": 1.3401054249945092, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.295377254486084, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7483852505683899, + "num_tokens": 315771304.0, + "step": 12203 + }, + { + "epoch": 1.3402152426971228, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.128187417984009, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7097380757331848, + "num_tokens": 315801733.0, + "step": 12204 + }, + { + "epoch": 1.3403250603997363, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2760279178619385, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.715160608291626, + "num_tokens": 315825516.0, + "step": 12205 + }, + { + "epoch": 1.34043487810235, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1298329830169678, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7073172330856323, + "num_tokens": 315854523.0, + "step": 12206 + }, + { + "epoch": 1.3405446958049638, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2317304611206055, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7218679785728455, + "num_tokens": 315881088.0, + "step": 12207 + }, + { + "epoch": 1.3406545135075774, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1238961219787598, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6825996041297913, + "num_tokens": 315913033.0, + "step": 12208 + }, + { + "epoch": 1.3407643312101911, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.229410171508789, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.708969235420227, + "num_tokens": 315939853.0, + "step": 12209 + }, + { + "epoch": 1.3408741489128047, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2201156616210938, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7118015289306641, + "num_tokens": 315964538.0, + "step": 12210 + }, + { + "epoch": 1.3409839666154184, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.8620268106460571, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7085481882095337, + "num_tokens": 316001696.0, + "step": 12211 + }, + { + "epoch": 1.3410937843180322, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.110865354537964, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7148237824440002, + "num_tokens": 316027487.0, + "step": 12212 + }, + { + "epoch": 1.3412036020206457, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.254347562789917, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7269866466522217, + "num_tokens": 316052502.0, + "step": 12213 + }, + { + "epoch": 1.3413134197232595, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3488502502441406, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7409712672233582, + "num_tokens": 316074290.0, + "step": 12214 + }, + { + "epoch": 1.341423237425873, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5955493450164795, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7140454053878784, + "num_tokens": 316095713.0, + "step": 12215 + }, + { + "epoch": 1.3415330551284868, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5140252113342285, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.713574230670929, + "num_tokens": 316119090.0, + "step": 12216 + }, + { + "epoch": 1.3416428728311005, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1794066429138184, + "learning_rate": 1e-06, + "loss": 1.0701, + "mean_token_accuracy": 0.6787354946136475, + "num_tokens": 316151908.0, + "step": 12217 + }, + { + "epoch": 1.341752690533714, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.8327763080596924, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7397940158843994, + "num_tokens": 316173112.0, + "step": 12218 + }, + { + "epoch": 1.3418625082363276, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.449910879135132, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7369875311851501, + "num_tokens": 316195721.0, + "step": 12219 + }, + { + "epoch": 1.3419723259389413, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.0656588077545166, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7115957140922546, + "num_tokens": 316224989.0, + "step": 12220 + }, + { + "epoch": 1.342082143641555, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.8998198509216309, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7078379988670349, + "num_tokens": 316259046.0, + "step": 12221 + }, + { + "epoch": 1.3421919613441686, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.341379404067993, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7330825328826904, + "num_tokens": 316281973.0, + "step": 12222 + }, + { + "epoch": 1.3423017790467824, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.309779644012451, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7124095559120178, + "num_tokens": 316307353.0, + "step": 12223 + }, + { + "epoch": 1.342411596749396, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3546805381774902, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7156333327293396, + "num_tokens": 316332042.0, + "step": 12224 + }, + { + "epoch": 1.3425214144520097, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.066071033477783, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7261860370635986, + "num_tokens": 316361949.0, + "step": 12225 + }, + { + "epoch": 1.3426312321546234, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.0845978260040283, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7415558099746704, + "num_tokens": 316389388.0, + "step": 12226 + }, + { + "epoch": 1.342741049857237, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.524174213409424, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7277417182922363, + "num_tokens": 316411616.0, + "step": 12227 + }, + { + "epoch": 1.3428508675598507, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2994463443756104, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7236589193344116, + "num_tokens": 316436210.0, + "step": 12228 + }, + { + "epoch": 1.3429606852624643, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3268861770629883, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.7030537128448486, + "num_tokens": 316460338.0, + "step": 12229 + }, + { + "epoch": 1.343070502965078, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.9464337825775146, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7227457761764526, + "num_tokens": 316492222.0, + "step": 12230 + }, + { + "epoch": 1.3431803206676916, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1357738971710205, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.729807436466217, + "num_tokens": 316520060.0, + "step": 12231 + }, + { + "epoch": 1.3432901383703053, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.9549360275268555, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6912683248519897, + "num_tokens": 316557495.0, + "step": 12232 + }, + { + "epoch": 1.3433999560729188, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.321903705596924, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7103639841079712, + "num_tokens": 316580479.0, + "step": 12233 + }, + { + "epoch": 1.3435097737755326, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2958920001983643, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7240327596664429, + "num_tokens": 316604763.0, + "step": 12234 + }, + { + "epoch": 1.3436195914781464, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2858328819274902, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.741652250289917, + "num_tokens": 316628819.0, + "step": 12235 + }, + { + "epoch": 1.34372940918076, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5158329010009766, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7281065583229065, + "num_tokens": 316649717.0, + "step": 12236 + }, + { + "epoch": 1.3438392268833736, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3002870082855225, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.716225802898407, + "num_tokens": 316673641.0, + "step": 12237 + }, + { + "epoch": 1.3439490445859872, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1222732067108154, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7114088535308838, + "num_tokens": 316703302.0, + "step": 12238 + }, + { + "epoch": 1.344058862288601, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.107065439224243, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7209091186523438, + "num_tokens": 316731974.0, + "step": 12239 + }, + { + "epoch": 1.3441686799912147, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3897557258605957, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7281975746154785, + "num_tokens": 316753735.0, + "step": 12240 + }, + { + "epoch": 1.3442784976938282, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3510990142822266, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7335139513015747, + "num_tokens": 316777376.0, + "step": 12241 + }, + { + "epoch": 1.3443883153964418, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.26841402053833, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7383609414100647, + "num_tokens": 316803200.0, + "step": 12242 + }, + { + "epoch": 1.3444981330990555, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.364104986190796, + "learning_rate": 1e-06, + "loss": 1.0659, + "mean_token_accuracy": 0.6724265813827515, + "num_tokens": 316830502.0, + "step": 12243 + }, + { + "epoch": 1.3446079508016693, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2000718116760254, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7289560437202454, + "num_tokens": 316858731.0, + "step": 12244 + }, + { + "epoch": 1.3447177685042828, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.043668746948242, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.7019685506820679, + "num_tokens": 316890304.0, + "step": 12245 + }, + { + "epoch": 1.3448275862068966, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.236929416656494, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7216687202453613, + "num_tokens": 316914171.0, + "step": 12246 + }, + { + "epoch": 1.34493740390951, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.073399543762207, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7262763381004333, + "num_tokens": 316945920.0, + "step": 12247 + }, + { + "epoch": 1.3450472216121239, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5496127605438232, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7304130792617798, + "num_tokens": 316967233.0, + "step": 12248 + }, + { + "epoch": 1.3451570393147376, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.009014844894409, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7077124714851379, + "num_tokens": 316996454.0, + "step": 12249 + }, + { + "epoch": 1.3452668570173512, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3097596168518066, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7150945663452148, + "num_tokens": 317021126.0, + "step": 12250 + }, + { + "epoch": 1.345376674719965, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1148228645324707, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7031288146972656, + "num_tokens": 317049288.0, + "step": 12251 + }, + { + "epoch": 1.3454864924225785, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.266430616378784, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7067269086837769, + "num_tokens": 317075763.0, + "step": 12252 + }, + { + "epoch": 1.3455963101251922, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1286919116973877, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7237696051597595, + "num_tokens": 317105490.0, + "step": 12253 + }, + { + "epoch": 1.345706127827806, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.330467462539673, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7228379845619202, + "num_tokens": 317128532.0, + "step": 12254 + }, + { + "epoch": 1.3458159455304195, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.490264415740967, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7243556976318359, + "num_tokens": 317149149.0, + "step": 12255 + }, + { + "epoch": 1.345925763233033, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.0808913707733154, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7302640080451965, + "num_tokens": 317179995.0, + "step": 12256 + }, + { + "epoch": 1.3460355809356468, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.189974308013916, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7213423252105713, + "num_tokens": 317208166.0, + "step": 12257 + }, + { + "epoch": 1.3461453986382605, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.372988700866699, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6911007165908813, + "num_tokens": 317235731.0, + "step": 12258 + }, + { + "epoch": 1.346255216340874, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3021717071533203, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7350132465362549, + "num_tokens": 317259161.0, + "step": 12259 + }, + { + "epoch": 1.3463650340434878, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.495004415512085, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7157371044158936, + "num_tokens": 317280911.0, + "step": 12260 + }, + { + "epoch": 1.3464748517461014, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.266843318939209, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7191066741943359, + "num_tokens": 317307150.0, + "step": 12261 + }, + { + "epoch": 1.3465846694487151, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1050164699554443, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7182971835136414, + "num_tokens": 317336466.0, + "step": 12262 + }, + { + "epoch": 1.346694487151329, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.523592948913574, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7429692149162292, + "num_tokens": 317358547.0, + "step": 12263 + }, + { + "epoch": 1.3468043048539424, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.8387924432754517, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7277567386627197, + "num_tokens": 317395401.0, + "step": 12264 + }, + { + "epoch": 1.3469141225565562, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1015257835388184, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.717836856842041, + "num_tokens": 317423197.0, + "step": 12265 + }, + { + "epoch": 1.3470239402591697, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.271124839782715, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7078307867050171, + "num_tokens": 317446824.0, + "step": 12266 + }, + { + "epoch": 1.3471337579617835, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2916126251220703, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7185048460960388, + "num_tokens": 317471013.0, + "step": 12267 + }, + { + "epoch": 1.3472435756643972, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3397750854492188, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7327955961227417, + "num_tokens": 317494213.0, + "step": 12268 + }, + { + "epoch": 1.3473533933670108, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1715381145477295, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7242437601089478, + "num_tokens": 317521586.0, + "step": 12269 + }, + { + "epoch": 1.3474632110696243, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.039062738418579, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7227004766464233, + "num_tokens": 317550203.0, + "step": 12270 + }, + { + "epoch": 1.347573028772238, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.0535600185394287, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7264090180397034, + "num_tokens": 317577786.0, + "step": 12271 + }, + { + "epoch": 1.3476828464748518, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1417276859283447, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7062204480171204, + "num_tokens": 317605820.0, + "step": 12272 + }, + { + "epoch": 1.3477926641774653, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1844162940979004, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6945732831954956, + "num_tokens": 317632019.0, + "step": 12273 + }, + { + "epoch": 1.347902481880079, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1602351665496826, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7145153880119324, + "num_tokens": 317660099.0, + "step": 12274 + }, + { + "epoch": 1.3480122995826926, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3058342933654785, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7090564370155334, + "num_tokens": 317686365.0, + "step": 12275 + }, + { + "epoch": 1.3481221172853064, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.5082430839538574, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7210938930511475, + "num_tokens": 317707757.0, + "step": 12276 + }, + { + "epoch": 1.3482319349879202, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3504273891448975, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7378371953964233, + "num_tokens": 317731141.0, + "step": 12277 + }, + { + "epoch": 1.3483417526905337, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2427802085876465, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7356412410736084, + "num_tokens": 317758618.0, + "step": 12278 + }, + { + "epoch": 1.3484515703931474, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3674516677856445, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7162444591522217, + "num_tokens": 317780188.0, + "step": 12279 + }, + { + "epoch": 1.348561388095761, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3096325397491455, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7435206770896912, + "num_tokens": 317803852.0, + "step": 12280 + }, + { + "epoch": 1.3486712057983747, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.9876874685287476, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.712299644947052, + "num_tokens": 317834475.0, + "step": 12281 + }, + { + "epoch": 1.3487810235009885, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1501803398132324, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7435474395751953, + "num_tokens": 317860962.0, + "step": 12282 + }, + { + "epoch": 1.348890841203602, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.213334798812866, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7234063744544983, + "num_tokens": 317887167.0, + "step": 12283 + }, + { + "epoch": 1.3490006589062156, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.658214569091797, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7381240129470825, + "num_tokens": 317907342.0, + "step": 12284 + }, + { + "epoch": 1.3491104766088293, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.65594220161438, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7176291942596436, + "num_tokens": 317929542.0, + "step": 12285 + }, + { + "epoch": 1.349220294311443, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.602564811706543, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7148162126541138, + "num_tokens": 317950297.0, + "step": 12286 + }, + { + "epoch": 1.3493301120140566, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.1277360916137695, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7041022777557373, + "num_tokens": 317978347.0, + "step": 12287 + }, + { + "epoch": 1.3494399297166704, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2535383701324463, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7365345358848572, + "num_tokens": 318003315.0, + "step": 12288 + }, + { + "epoch": 1.349549747419284, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.324808120727539, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7281485795974731, + "num_tokens": 318027089.0, + "step": 12289 + }, + { + "epoch": 1.3496595651218977, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.414623260498047, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7224605679512024, + "num_tokens": 318049983.0, + "step": 12290 + }, + { + "epoch": 1.3497693828245114, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.283005952835083, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7102965116500854, + "num_tokens": 318076446.0, + "step": 12291 + }, + { + "epoch": 1.349879200527125, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.3818697929382324, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7268568277359009, + "num_tokens": 318099282.0, + "step": 12292 + }, + { + "epoch": 1.3499890182297387, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 2.2891762256622314, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7133201360702515, + "num_tokens": 318125169.0, + "step": 12293 + }, + { + "epoch": 1.3500988359323522, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.9437110424041748, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7401940226554871, + "num_tokens": 318157779.0, + "step": 12294 + }, + { + "epoch": 1.350208653634966, + "ewc_loss": 0.00011301040649414062, + "grad_norm": 1.9686777591705322, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7458657026290894, + "num_tokens": 318187301.0, + "step": 12295 + }, + { + "epoch": 1.3503184713375795, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.1949703693389893, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.713519811630249, + "num_tokens": 318212475.0, + "step": 12296 + }, + { + "epoch": 1.3504282890401933, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.0170788764953613, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7073695659637451, + "num_tokens": 318245545.0, + "step": 12297 + }, + { + "epoch": 1.3505381067428068, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.338381767272949, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7346783876419067, + "num_tokens": 318268347.0, + "step": 12298 + }, + { + "epoch": 1.3506479244454206, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.0139505863189697, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7255011796951294, + "num_tokens": 318296588.0, + "step": 12299 + }, + { + "epoch": 1.3507577421480343, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.400071620941162, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7251724004745483, + "num_tokens": 318318387.0, + "step": 12300 + }, + { + "epoch": 1.3508675598506479, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.280238628387451, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7557424902915955, + "num_tokens": 318340938.0, + "step": 12301 + }, + { + "epoch": 1.3509773775532616, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.346564292907715, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7241683006286621, + "num_tokens": 318362842.0, + "step": 12302 + }, + { + "epoch": 1.3510871952558752, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.212610960006714, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7343483567237854, + "num_tokens": 318388176.0, + "step": 12303 + }, + { + "epoch": 1.351197012958489, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.406212329864502, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7344130873680115, + "num_tokens": 318411027.0, + "step": 12304 + }, + { + "epoch": 1.3513068306611027, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.4596173763275146, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7305055856704712, + "num_tokens": 318431454.0, + "step": 12305 + }, + { + "epoch": 1.3514166483637162, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.3835487365722656, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7091874480247498, + "num_tokens": 318456081.0, + "step": 12306 + }, + { + "epoch": 1.3515264660663298, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.1393473148345947, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.715530514717102, + "num_tokens": 318484595.0, + "step": 12307 + }, + { + "epoch": 1.3516362837689435, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.548774242401123, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7074457406997681, + "num_tokens": 318507267.0, + "step": 12308 + }, + { + "epoch": 1.3517461014715573, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.3308908939361572, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7339184284210205, + "num_tokens": 318530460.0, + "step": 12309 + }, + { + "epoch": 1.3518559191741708, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.153927803039551, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7148470282554626, + "num_tokens": 318557169.0, + "step": 12310 + }, + { + "epoch": 1.3519657368767846, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.1233527660369873, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7046223878860474, + "num_tokens": 318585845.0, + "step": 12311 + }, + { + "epoch": 1.352075554579398, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.481400728225708, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.704056978225708, + "num_tokens": 318607694.0, + "step": 12312 + }, + { + "epoch": 1.3521853722820119, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.4000537395477295, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7145659923553467, + "num_tokens": 318630886.0, + "step": 12313 + }, + { + "epoch": 1.3522951899846256, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.5456790924072266, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7350283861160278, + "num_tokens": 318651318.0, + "step": 12314 + }, + { + "epoch": 1.3524050076872391, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.0799810886383057, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7089893221855164, + "num_tokens": 318681369.0, + "step": 12315 + }, + { + "epoch": 1.352514825389853, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.303069591522217, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7071346044540405, + "num_tokens": 318708443.0, + "step": 12316 + }, + { + "epoch": 1.3526246430924664, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.5043601989746094, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7220295667648315, + "num_tokens": 318728753.0, + "step": 12317 + }, + { + "epoch": 1.3527344607950802, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.253187417984009, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7096652984619141, + "num_tokens": 318752398.0, + "step": 12318 + }, + { + "epoch": 1.352844278497694, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.288043975830078, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6987138986587524, + "num_tokens": 318777523.0, + "step": 12319 + }, + { + "epoch": 1.3529540962003075, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.1819379329681396, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7149645090103149, + "num_tokens": 318804092.0, + "step": 12320 + }, + { + "epoch": 1.353063913902921, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.27768874168396, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7260432243347168, + "num_tokens": 318830073.0, + "step": 12321 + }, + { + "epoch": 1.3531737316055348, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.239889621734619, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7201613783836365, + "num_tokens": 318855429.0, + "step": 12322 + }, + { + "epoch": 1.3532835493081485, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.2904398441314697, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7309523224830627, + "num_tokens": 318878567.0, + "step": 12323 + }, + { + "epoch": 1.353393367010762, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.112879514694214, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7388404607772827, + "num_tokens": 318905612.0, + "step": 12324 + }, + { + "epoch": 1.3535031847133758, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.206718921661377, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7147011756896973, + "num_tokens": 318932808.0, + "step": 12325 + }, + { + "epoch": 1.3536130024159894, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.2693893909454346, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7125564217567444, + "num_tokens": 318958431.0, + "step": 12326 + }, + { + "epoch": 1.3537228201186031, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.216780185699463, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.732784628868103, + "num_tokens": 318981995.0, + "step": 12327 + }, + { + "epoch": 1.3538326378212169, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.185307741165161, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.697666347026825, + "num_tokens": 319011387.0, + "step": 12328 + }, + { + "epoch": 1.3539424555238304, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.0306036472320557, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6909881234169006, + "num_tokens": 319044796.0, + "step": 12329 + }, + { + "epoch": 1.3540522732264442, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.034817695617676, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7136573195457458, + "num_tokens": 319074717.0, + "step": 12330 + }, + { + "epoch": 1.3541620909290577, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.3624050617218018, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.6998500227928162, + "num_tokens": 319099557.0, + "step": 12331 + }, + { + "epoch": 1.3542719086316715, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.4424262046813965, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7292680144309998, + "num_tokens": 319123101.0, + "step": 12332 + }, + { + "epoch": 1.3543817263342852, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.4742746353149414, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7223929762840271, + "num_tokens": 319145466.0, + "step": 12333 + }, + { + "epoch": 1.3544915440368988, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 1.7795050144195557, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.688448429107666, + "num_tokens": 319184665.0, + "step": 12334 + }, + { + "epoch": 1.3546013617395123, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.541625738143921, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7056652307510376, + "num_tokens": 319205907.0, + "step": 12335 + }, + { + "epoch": 1.354711179442126, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.445828914642334, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7229366302490234, + "num_tokens": 319227066.0, + "step": 12336 + }, + { + "epoch": 1.3548209971447398, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.2967145442962646, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7020360827445984, + "num_tokens": 319252101.0, + "step": 12337 + }, + { + "epoch": 1.3549308148473533, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.361482858657837, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.719677209854126, + "num_tokens": 319275052.0, + "step": 12338 + }, + { + "epoch": 1.355040632549967, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.5147511959075928, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.7507981061935425, + "num_tokens": 319295313.0, + "step": 12339 + }, + { + "epoch": 1.3551504502525806, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.2611494064331055, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.711478590965271, + "num_tokens": 319321658.0, + "step": 12340 + }, + { + "epoch": 1.3552602679551944, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.408336639404297, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7209208607673645, + "num_tokens": 319344828.0, + "step": 12341 + }, + { + "epoch": 1.3553700856578081, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1586694717407227, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7199862003326416, + "num_tokens": 319372964.0, + "step": 12342 + }, + { + "epoch": 1.3554799033604217, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.3934617042541504, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7227107286453247, + "num_tokens": 319397019.0, + "step": 12343 + }, + { + "epoch": 1.3555897210630354, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.263047695159912, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6952991485595703, + "num_tokens": 319423150.0, + "step": 12344 + }, + { + "epoch": 1.355699538765649, + "ewc_loss": 0.00011348724365234375, + "grad_norm": 2.5063376426696777, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7392257452011108, + "num_tokens": 319444028.0, + "step": 12345 + }, + { + "epoch": 1.3558093564682627, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.23764967918396, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7082076072692871, + "num_tokens": 319471726.0, + "step": 12346 + }, + { + "epoch": 1.3559191741708763, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.2626445293426514, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7111110091209412, + "num_tokens": 319496920.0, + "step": 12347 + }, + { + "epoch": 1.35602899187349, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.158623456954956, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.7427107095718384, + "num_tokens": 319522394.0, + "step": 12348 + }, + { + "epoch": 1.3561388095761036, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.062105178833008, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7369805574417114, + "num_tokens": 319549366.0, + "step": 12349 + }, + { + "epoch": 1.3562486272787173, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.2780416011810303, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7389992475509644, + "num_tokens": 319573807.0, + "step": 12350 + }, + { + "epoch": 1.356358444981331, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.389315605163574, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7176914811134338, + "num_tokens": 319597163.0, + "step": 12351 + }, + { + "epoch": 1.3564682626839446, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1186437606811523, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7329146862030029, + "num_tokens": 319624851.0, + "step": 12352 + }, + { + "epoch": 1.3565780803865584, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3773841857910156, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7046957015991211, + "num_tokens": 319648659.0, + "step": 12353 + }, + { + "epoch": 1.356687898089172, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.6635901927948, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.712098240852356, + "num_tokens": 319668633.0, + "step": 12354 + }, + { + "epoch": 1.3567977157917857, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.040915012359619, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7172508239746094, + "num_tokens": 319697321.0, + "step": 12355 + }, + { + "epoch": 1.3569075334943994, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.0655179023742676, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6934045553207397, + "num_tokens": 319727386.0, + "step": 12356 + }, + { + "epoch": 1.357017351197013, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.300607919692993, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7214450836181641, + "num_tokens": 319751705.0, + "step": 12357 + }, + { + "epoch": 1.3571271688996267, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3455512523651123, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7296156287193298, + "num_tokens": 319777094.0, + "step": 12358 + }, + { + "epoch": 1.3572369866022402, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.0178511142730713, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7152380347251892, + "num_tokens": 319806021.0, + "step": 12359 + }, + { + "epoch": 1.357346804304854, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1728832721710205, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7114365696907043, + "num_tokens": 319834341.0, + "step": 12360 + }, + { + "epoch": 1.3574566220074675, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4443209171295166, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6868392825126648, + "num_tokens": 319863895.0, + "step": 12361 + }, + { + "epoch": 1.3575664397100813, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1036665439605713, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7057336568832397, + "num_tokens": 319893533.0, + "step": 12362 + }, + { + "epoch": 1.3576762574126948, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.074537515640259, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.6981185674667358, + "num_tokens": 319925237.0, + "step": 12363 + }, + { + "epoch": 1.3577860751153086, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.242384672164917, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7109707593917847, + "num_tokens": 319953740.0, + "step": 12364 + }, + { + "epoch": 1.3578958928179223, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4977219104766846, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.740139365196228, + "num_tokens": 319975003.0, + "step": 12365 + }, + { + "epoch": 1.3580057105205359, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.0234737396240234, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7158017754554749, + "num_tokens": 320005624.0, + "step": 12366 + }, + { + "epoch": 1.3581155282231496, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.496368169784546, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7186509966850281, + "num_tokens": 320030292.0, + "step": 12367 + }, + { + "epoch": 1.3582253459257632, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3072969913482666, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7081387639045715, + "num_tokens": 320054812.0, + "step": 12368 + }, + { + "epoch": 1.358335163628377, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1162631511688232, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7172275185585022, + "num_tokens": 320083315.0, + "step": 12369 + }, + { + "epoch": 1.3584449813309907, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.357464551925659, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7102400064468384, + "num_tokens": 320107574.0, + "step": 12370 + }, + { + "epoch": 1.3585547990336042, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.063199758529663, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7013375759124756, + "num_tokens": 320137930.0, + "step": 12371 + }, + { + "epoch": 1.3586646167362177, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.521711587905884, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7142395377159119, + "num_tokens": 320160452.0, + "step": 12372 + }, + { + "epoch": 1.3587744344388315, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.390801191329956, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7066213488578796, + "num_tokens": 320186376.0, + "step": 12373 + }, + { + "epoch": 1.3588842521414453, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3404059410095215, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.7063112258911133, + "num_tokens": 320212668.0, + "step": 12374 + }, + { + "epoch": 1.3589940698440588, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.01861310005188, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7213035821914673, + "num_tokens": 320242462.0, + "step": 12375 + }, + { + "epoch": 1.3591038875466726, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.094050884246826, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7301663756370544, + "num_tokens": 320270761.0, + "step": 12376 + }, + { + "epoch": 1.359213705249286, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.165876626968384, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7304685711860657, + "num_tokens": 320299388.0, + "step": 12377 + }, + { + "epoch": 1.3593235229518998, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 1.970129132270813, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7239216566085815, + "num_tokens": 320329246.0, + "step": 12378 + }, + { + "epoch": 1.3594333406545136, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1809778213500977, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7157020568847656, + "num_tokens": 320357784.0, + "step": 12379 + }, + { + "epoch": 1.3595431583571271, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.105583429336548, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7194976806640625, + "num_tokens": 320388559.0, + "step": 12380 + }, + { + "epoch": 1.359652976059741, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.17172908782959, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7083730101585388, + "num_tokens": 320413902.0, + "step": 12381 + }, + { + "epoch": 1.3597627937623544, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.6162681579589844, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7188875675201416, + "num_tokens": 320433528.0, + "step": 12382 + }, + { + "epoch": 1.3598726114649682, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.2113609313964844, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6994137763977051, + "num_tokens": 320461910.0, + "step": 12383 + }, + { + "epoch": 1.359982429167582, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.378095865249634, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7143850326538086, + "num_tokens": 320488243.0, + "step": 12384 + }, + { + "epoch": 1.3600922468701955, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.338059186935425, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7012463808059692, + "num_tokens": 320513492.0, + "step": 12385 + }, + { + "epoch": 1.360202064572809, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1937625408172607, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.703508198261261, + "num_tokens": 320541935.0, + "step": 12386 + }, + { + "epoch": 1.3603118822754228, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.6990549564361572, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7475612163543701, + "num_tokens": 320559125.0, + "step": 12387 + }, + { + "epoch": 1.3604216999780365, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.2760984897613525, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7170522212982178, + "num_tokens": 320583098.0, + "step": 12388 + }, + { + "epoch": 1.36053151768065, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4692318439483643, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7181410789489746, + "num_tokens": 320603576.0, + "step": 12389 + }, + { + "epoch": 1.3606413353832638, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 1.788105845451355, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6960859894752502, + "num_tokens": 320644237.0, + "step": 12390 + }, + { + "epoch": 1.3607511530858774, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1686782836914062, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7005763649940491, + "num_tokens": 320673042.0, + "step": 12391 + }, + { + "epoch": 1.360860970788491, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 3.1773502826690674, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7199447154998779, + "num_tokens": 320696520.0, + "step": 12392 + }, + { + "epoch": 1.3609707884911049, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3241002559661865, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7412621974945068, + "num_tokens": 320719902.0, + "step": 12393 + }, + { + "epoch": 1.3610806061937184, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 1.933220624923706, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7190694212913513, + "num_tokens": 320753425.0, + "step": 12394 + }, + { + "epoch": 1.3611904238963322, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.176440954208374, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7169967293739319, + "num_tokens": 320780438.0, + "step": 12395 + }, + { + "epoch": 1.3613002415989457, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4264214038848877, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7344520092010498, + "num_tokens": 320801567.0, + "step": 12396 + }, + { + "epoch": 1.3614100593015594, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.405142068862915, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7249525785446167, + "num_tokens": 320825226.0, + "step": 12397 + }, + { + "epoch": 1.3615198770041732, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.494302988052368, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7372037768363953, + "num_tokens": 320845317.0, + "step": 12398 + }, + { + "epoch": 1.3616296947067867, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.7789199352264404, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7261976003646851, + "num_tokens": 320863226.0, + "step": 12399 + }, + { + "epoch": 1.3617395124094003, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.5603389739990234, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7325266599655151, + "num_tokens": 320884355.0, + "step": 12400 + }, + { + "epoch": 1.361849330112014, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.3941566944122314, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7172579765319824, + "num_tokens": 320909532.0, + "step": 12401 + }, + { + "epoch": 1.3619591478146278, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.299168586730957, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7104511260986328, + "num_tokens": 320937048.0, + "step": 12402 + }, + { + "epoch": 1.3620689655172413, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.282182216644287, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7207697629928589, + "num_tokens": 320960287.0, + "step": 12403 + }, + { + "epoch": 1.362178783219855, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.6840739250183105, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7069729566574097, + "num_tokens": 320979821.0, + "step": 12404 + }, + { + "epoch": 1.3622886009224686, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.428159475326538, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7027184963226318, + "num_tokens": 321002289.0, + "step": 12405 + }, + { + "epoch": 1.3623984186250824, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.12933349609375, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.6970779299736023, + "num_tokens": 321031301.0, + "step": 12406 + }, + { + "epoch": 1.3625082363276961, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4345040321350098, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7083483934402466, + "num_tokens": 321054918.0, + "step": 12407 + }, + { + "epoch": 1.3626180540303097, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.16444730758667, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7298005819320679, + "num_tokens": 321081261.0, + "step": 12408 + }, + { + "epoch": 1.3627278717329234, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.1283130645751953, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7315808534622192, + "num_tokens": 321109810.0, + "step": 12409 + }, + { + "epoch": 1.362837689435537, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.4266748428344727, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7227673530578613, + "num_tokens": 321132982.0, + "step": 12410 + }, + { + "epoch": 1.3629475071381507, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 1.9863442182540894, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7238336205482483, + "num_tokens": 321162779.0, + "step": 12411 + }, + { + "epoch": 1.3630573248407643, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.5616796016693115, + "learning_rate": 1e-06, + "loss": 0.7719, + "mean_token_accuracy": 0.7557836771011353, + "num_tokens": 321181944.0, + "step": 12412 + }, + { + "epoch": 1.363167142543378, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.0690274238586426, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7071028351783752, + "num_tokens": 321210380.0, + "step": 12413 + }, + { + "epoch": 1.3632769602459915, + "ewc_loss": 0.00011396408081054688, + "grad_norm": 2.259845733642578, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7175199389457703, + "num_tokens": 321235617.0, + "step": 12414 + }, + { + "epoch": 1.3633867779486053, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.3263680934906006, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7213543653488159, + "num_tokens": 321260493.0, + "step": 12415 + }, + { + "epoch": 1.363496595651219, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.468975305557251, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7387754321098328, + "num_tokens": 321282094.0, + "step": 12416 + }, + { + "epoch": 1.3636064133538326, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.0100581645965576, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7109640836715698, + "num_tokens": 321313841.0, + "step": 12417 + }, + { + "epoch": 1.3637162310564463, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.4362504482269287, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7227051258087158, + "num_tokens": 321336344.0, + "step": 12418 + }, + { + "epoch": 1.3638260487590599, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.0198633670806885, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.7027099132537842, + "num_tokens": 321369387.0, + "step": 12419 + }, + { + "epoch": 1.3639358664616736, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.2572391033172607, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7281648516654968, + "num_tokens": 321395461.0, + "step": 12420 + }, + { + "epoch": 1.3640456841642874, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.2087934017181396, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7326135635375977, + "num_tokens": 321420665.0, + "step": 12421 + }, + { + "epoch": 1.364155501866901, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.2004904747009277, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7282173037528992, + "num_tokens": 321446602.0, + "step": 12422 + }, + { + "epoch": 1.3642653195695145, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.259795665740967, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7084957361221313, + "num_tokens": 321472037.0, + "step": 12423 + }, + { + "epoch": 1.3643751372721282, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.1575446128845215, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7202119827270508, + "num_tokens": 321498650.0, + "step": 12424 + }, + { + "epoch": 1.364484954974742, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.23488712310791, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.7015862464904785, + "num_tokens": 321524056.0, + "step": 12425 + }, + { + "epoch": 1.3645947726773555, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.3560664653778076, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7009907960891724, + "num_tokens": 321548443.0, + "step": 12426 + }, + { + "epoch": 1.3647045903799693, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.1337873935699463, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7049670815467834, + "num_tokens": 321578386.0, + "step": 12427 + }, + { + "epoch": 1.3648144080825828, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.0357537269592285, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7191907167434692, + "num_tokens": 321607798.0, + "step": 12428 + }, + { + "epoch": 1.3649242257851966, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.0675885677337646, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6899735927581787, + "num_tokens": 321637720.0, + "step": 12429 + }, + { + "epoch": 1.3650340434878103, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.297369956970215, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.6981366872787476, + "num_tokens": 321665072.0, + "step": 12430 + }, + { + "epoch": 1.3651438611904239, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.230008840560913, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7300553321838379, + "num_tokens": 321690054.0, + "step": 12431 + }, + { + "epoch": 1.3652536788930376, + "ewc_loss": 0.00011444091796875, + "grad_norm": 1.9579116106033325, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7291640639305115, + "num_tokens": 321722282.0, + "step": 12432 + }, + { + "epoch": 1.3653634965956511, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.0357704162597656, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6978318691253662, + "num_tokens": 321754184.0, + "step": 12433 + }, + { + "epoch": 1.365473314298265, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.5423715114593506, + "learning_rate": 1e-06, + "loss": 0.7499, + "mean_token_accuracy": 0.7610922455787659, + "num_tokens": 321774303.0, + "step": 12434 + }, + { + "epoch": 1.3655831320008787, + "ewc_loss": 0.00011444091796875, + "grad_norm": 2.2258927822113037, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.719173789024353, + "num_tokens": 321799656.0, + "step": 12435 + }, + { + "epoch": 1.3656929497034922, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.7714905738830566, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7175410985946655, + "num_tokens": 321818801.0, + "step": 12436 + }, + { + "epoch": 1.3658027674061057, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.25144624710083, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7071126699447632, + "num_tokens": 321844992.0, + "step": 12437 + }, + { + "epoch": 1.3659125851087195, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.510286331176758, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7223249673843384, + "num_tokens": 321866970.0, + "step": 12438 + }, + { + "epoch": 1.3660224028113332, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2367286682128906, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7545101642608643, + "num_tokens": 321891675.0, + "step": 12439 + }, + { + "epoch": 1.3661322205139468, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.86726713180542, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.722137451171875, + "num_tokens": 321926398.0, + "step": 12440 + }, + { + "epoch": 1.3662420382165605, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2981042861938477, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7332906126976013, + "num_tokens": 321950293.0, + "step": 12441 + }, + { + "epoch": 1.366351855919174, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1094300746917725, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7225207090377808, + "num_tokens": 321979849.0, + "step": 12442 + }, + { + "epoch": 1.3664616736217878, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4975130558013916, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7218362092971802, + "num_tokens": 322001928.0, + "step": 12443 + }, + { + "epoch": 1.3665714913244016, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2269885540008545, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7251809239387512, + "num_tokens": 322029905.0, + "step": 12444 + }, + { + "epoch": 1.3666813090270151, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2645633220672607, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7115591764450073, + "num_tokens": 322055755.0, + "step": 12445 + }, + { + "epoch": 1.3667911267296289, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3090405464172363, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6986018419265747, + "num_tokens": 322081185.0, + "step": 12446 + }, + { + "epoch": 1.3669009444322424, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.368664264678955, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7137447595596313, + "num_tokens": 322107051.0, + "step": 12447 + }, + { + "epoch": 1.3670107621348562, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.373805522918701, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.735918402671814, + "num_tokens": 322130687.0, + "step": 12448 + }, + { + "epoch": 1.36712057983747, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.008474826812744, + "learning_rate": 1e-06, + "loss": 1.0661, + "mean_token_accuracy": 0.6774170398712158, + "num_tokens": 322164725.0, + "step": 12449 + }, + { + "epoch": 1.3672303975400835, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9881399869918823, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7203073501586914, + "num_tokens": 322193166.0, + "step": 12450 + }, + { + "epoch": 1.367340215242697, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4526572227478027, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7485769987106323, + "num_tokens": 322212005.0, + "step": 12451 + }, + { + "epoch": 1.3674500329453108, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2473387718200684, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7011523246765137, + "num_tokens": 322239553.0, + "step": 12452 + }, + { + "epoch": 1.3675598506479245, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0992987155914307, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7197892069816589, + "num_tokens": 322269202.0, + "step": 12453 + }, + { + "epoch": 1.367669668350538, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.217897415161133, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7030772566795349, + "num_tokens": 322297129.0, + "step": 12454 + }, + { + "epoch": 1.3677794860531518, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9956343173980713, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7035214304924011, + "num_tokens": 322331302.0, + "step": 12455 + }, + { + "epoch": 1.3678893037557653, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2890491485595703, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7157469391822815, + "num_tokens": 322356109.0, + "step": 12456 + }, + { + "epoch": 1.367999121458379, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2427403926849365, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6990892887115479, + "num_tokens": 322382372.0, + "step": 12457 + }, + { + "epoch": 1.3681089391609929, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0922703742980957, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7385174036026001, + "num_tokens": 322408429.0, + "step": 12458 + }, + { + "epoch": 1.3682187568636064, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2718112468719482, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7239678502082825, + "num_tokens": 322433048.0, + "step": 12459 + }, + { + "epoch": 1.3683285745662201, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.214373826980591, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7284641861915588, + "num_tokens": 322458545.0, + "step": 12460 + }, + { + "epoch": 1.3684383922688337, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.7288897037506104, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7395509481430054, + "num_tokens": 322478183.0, + "step": 12461 + }, + { + "epoch": 1.3685482099714474, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.493433713912964, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7161288857460022, + "num_tokens": 322499421.0, + "step": 12462 + }, + { + "epoch": 1.3686580276740612, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.095428466796875, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.722813606262207, + "num_tokens": 322528571.0, + "step": 12463 + }, + { + "epoch": 1.3687678453766747, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.427669048309326, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7057297229766846, + "num_tokens": 322552969.0, + "step": 12464 + }, + { + "epoch": 1.3688776630792883, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4635367393493652, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.717546284198761, + "num_tokens": 322576081.0, + "step": 12465 + }, + { + "epoch": 1.368987480781902, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2083733081817627, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7198330163955688, + "num_tokens": 322605259.0, + "step": 12466 + }, + { + "epoch": 1.3690972984845158, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.195180892944336, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7274982333183289, + "num_tokens": 322632366.0, + "step": 12467 + }, + { + "epoch": 1.3692071161871293, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.319565773010254, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.6855988502502441, + "num_tokens": 322658014.0, + "step": 12468 + }, + { + "epoch": 1.369316933889743, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.254096031188965, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7184658050537109, + "num_tokens": 322685064.0, + "step": 12469 + }, + { + "epoch": 1.3694267515923566, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2453346252441406, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7304600477218628, + "num_tokens": 322711501.0, + "step": 12470 + }, + { + "epoch": 1.3695365692949704, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2036359310150146, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7340829372406006, + "num_tokens": 322738897.0, + "step": 12471 + }, + { + "epoch": 1.3696463869975841, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3428003787994385, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7259914875030518, + "num_tokens": 322762890.0, + "step": 12472 + }, + { + "epoch": 1.3697562047001977, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.285531520843506, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7068395614624023, + "num_tokens": 322788541.0, + "step": 12473 + }, + { + "epoch": 1.3698660224028114, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.289698600769043, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7145770192146301, + "num_tokens": 322812278.0, + "step": 12474 + }, + { + "epoch": 1.369975840105425, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0780117511749268, + "learning_rate": 1e-06, + "loss": 1.0535, + "mean_token_accuracy": 0.6875118613243103, + "num_tokens": 322844266.0, + "step": 12475 + }, + { + "epoch": 1.3700856578080387, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.374093770980835, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6982908248901367, + "num_tokens": 322867518.0, + "step": 12476 + }, + { + "epoch": 1.3701954755106522, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.45350980758667, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7230733036994934, + "num_tokens": 322891856.0, + "step": 12477 + }, + { + "epoch": 1.370305293213266, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3522448539733887, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6981673240661621, + "num_tokens": 322916582.0, + "step": 12478 + }, + { + "epoch": 1.3704151109158795, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.6124424934387207, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7148937582969666, + "num_tokens": 322938281.0, + "step": 12479 + }, + { + "epoch": 1.3705249286184933, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3780629634857178, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7194450497627258, + "num_tokens": 322961516.0, + "step": 12480 + }, + { + "epoch": 1.370634746321107, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.7346956729888916, + "learning_rate": 1e-06, + "loss": 0.792, + "mean_token_accuracy": 0.7511088848114014, + "num_tokens": 322979819.0, + "step": 12481 + }, + { + "epoch": 1.3707445640237206, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.8946573734283447, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7099772095680237, + "num_tokens": 323013915.0, + "step": 12482 + }, + { + "epoch": 1.3708543817263343, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.431159257888794, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7395668625831604, + "num_tokens": 323036826.0, + "step": 12483 + }, + { + "epoch": 1.3709641994289479, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0665090084075928, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7085849046707153, + "num_tokens": 323067041.0, + "step": 12484 + }, + { + "epoch": 1.3710740171315616, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.391781806945801, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7181580066680908, + "num_tokens": 323090266.0, + "step": 12485 + }, + { + "epoch": 1.3711838348341754, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3645577430725098, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7031421661376953, + "num_tokens": 323116084.0, + "step": 12486 + }, + { + "epoch": 1.371293652536789, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0832958221435547, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7209063768386841, + "num_tokens": 323145061.0, + "step": 12487 + }, + { + "epoch": 1.3714034702394025, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4069838523864746, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.723020076751709, + "num_tokens": 323167555.0, + "step": 12488 + }, + { + "epoch": 1.3715132879420162, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.430079221725464, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7304795384407043, + "num_tokens": 323189204.0, + "step": 12489 + }, + { + "epoch": 1.37162310564463, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3430910110473633, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7305639982223511, + "num_tokens": 323212801.0, + "step": 12490 + }, + { + "epoch": 1.3717329233472435, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0444822311401367, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7314010858535767, + "num_tokens": 323239307.0, + "step": 12491 + }, + { + "epoch": 1.3718427410498573, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.329329252243042, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7100894451141357, + "num_tokens": 323264180.0, + "step": 12492 + }, + { + "epoch": 1.3719525587524708, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1661336421966553, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7036976218223572, + "num_tokens": 323292103.0, + "step": 12493 + }, + { + "epoch": 1.3720623764550846, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.320540189743042, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7158858776092529, + "num_tokens": 323315127.0, + "step": 12494 + }, + { + "epoch": 1.3721721941576983, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.328646659851074, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6959725022315979, + "num_tokens": 323340495.0, + "step": 12495 + }, + { + "epoch": 1.3722820118603118, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3603429794311523, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6953006982803345, + "num_tokens": 323366720.0, + "step": 12496 + }, + { + "epoch": 1.3723918295629256, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.367792844772339, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7277155518531799, + "num_tokens": 323390323.0, + "step": 12497 + }, + { + "epoch": 1.3725016472655391, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3426690101623535, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7276384234428406, + "num_tokens": 323412899.0, + "step": 12498 + }, + { + "epoch": 1.372611464968153, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.859872341156006, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7323411703109741, + "num_tokens": 323430627.0, + "step": 12499 + }, + { + "epoch": 1.3727212826707667, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.5828371047973633, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7339474558830261, + "num_tokens": 323451115.0, + "step": 12500 + }, + { + "epoch": 1.3728311003733802, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.301792621612549, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7134436964988708, + "num_tokens": 323477932.0, + "step": 12501 + }, + { + "epoch": 1.3729409180759937, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0608530044555664, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7106692790985107, + "num_tokens": 323507554.0, + "step": 12502 + }, + { + "epoch": 1.3730507357786075, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.188354969024658, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7047913670539856, + "num_tokens": 323536096.0, + "step": 12503 + }, + { + "epoch": 1.3731605534812212, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0617856979370117, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6864354014396667, + "num_tokens": 323568211.0, + "step": 12504 + }, + { + "epoch": 1.3732703711838348, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.107780933380127, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7026581764221191, + "num_tokens": 323597033.0, + "step": 12505 + }, + { + "epoch": 1.3733801888864485, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1707534790039062, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7338898777961731, + "num_tokens": 323623987.0, + "step": 12506 + }, + { + "epoch": 1.373490006589062, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.108083486557007, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7196489572525024, + "num_tokens": 323651896.0, + "step": 12507 + }, + { + "epoch": 1.3735998242916758, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.196046829223633, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7032774090766907, + "num_tokens": 323679655.0, + "step": 12508 + }, + { + "epoch": 1.3737096419942896, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2786974906921387, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7115061283111572, + "num_tokens": 323704418.0, + "step": 12509 + }, + { + "epoch": 1.3738194596969031, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.578382968902588, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7126233577728271, + "num_tokens": 323727290.0, + "step": 12510 + }, + { + "epoch": 1.3739292773995169, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.140463352203369, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6992788314819336, + "num_tokens": 323756261.0, + "step": 12511 + }, + { + "epoch": 1.3740390951021304, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0024054050445557, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7008838653564453, + "num_tokens": 323785724.0, + "step": 12512 + }, + { + "epoch": 1.3741489128047442, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1914186477661133, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6996281147003174, + "num_tokens": 323812965.0, + "step": 12513 + }, + { + "epoch": 1.374258730507358, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.205368995666504, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7387015223503113, + "num_tokens": 323838244.0, + "step": 12514 + }, + { + "epoch": 1.3743685482099715, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.475435972213745, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7468221187591553, + "num_tokens": 323858223.0, + "step": 12515 + }, + { + "epoch": 1.374478365912585, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9716391563415527, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7072497606277466, + "num_tokens": 323890650.0, + "step": 12516 + }, + { + "epoch": 1.3745881836151987, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2046823501586914, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7177023887634277, + "num_tokens": 323915138.0, + "step": 12517 + }, + { + "epoch": 1.3746980013178125, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4270524978637695, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7003827691078186, + "num_tokens": 323939118.0, + "step": 12518 + }, + { + "epoch": 1.374807819020426, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9925485849380493, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7151641845703125, + "num_tokens": 323967959.0, + "step": 12519 + }, + { + "epoch": 1.3749176367230398, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4722940921783447, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.719721794128418, + "num_tokens": 323989323.0, + "step": 12520 + }, + { + "epoch": 1.3750274544256533, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.499053716659546, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7275671362876892, + "num_tokens": 324010929.0, + "step": 12521 + }, + { + "epoch": 1.375137272128267, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.657050848007202, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7483788132667542, + "num_tokens": 324029764.0, + "step": 12522 + }, + { + "epoch": 1.3752470898308808, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4539763927459717, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7370750308036804, + "num_tokens": 324051254.0, + "step": 12523 + }, + { + "epoch": 1.3753569075334944, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.453521490097046, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7245303392410278, + "num_tokens": 324073161.0, + "step": 12524 + }, + { + "epoch": 1.3754667252361081, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2667253017425537, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7224906086921692, + "num_tokens": 324096249.0, + "step": 12525 + }, + { + "epoch": 1.3755765429387217, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3678369522094727, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7152891159057617, + "num_tokens": 324119618.0, + "step": 12526 + }, + { + "epoch": 1.3756863606413354, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1355535984039307, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.715324878692627, + "num_tokens": 324149285.0, + "step": 12527 + }, + { + "epoch": 1.3757961783439492, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.5213944911956787, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7169681787490845, + "num_tokens": 324170406.0, + "step": 12528 + }, + { + "epoch": 1.3759059960465627, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.156832695007324, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7010275721549988, + "num_tokens": 324198163.0, + "step": 12529 + }, + { + "epoch": 1.3760158137491763, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1708555221557617, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7146288752555847, + "num_tokens": 324226223.0, + "step": 12530 + }, + { + "epoch": 1.37612563145179, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2027785778045654, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7026662826538086, + "num_tokens": 324252547.0, + "step": 12531 + }, + { + "epoch": 1.3762354491544038, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.8717129230499268, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7290007472038269, + "num_tokens": 324269247.0, + "step": 12532 + }, + { + "epoch": 1.3763452668570173, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.155667304992676, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.707586407661438, + "num_tokens": 324294598.0, + "step": 12533 + }, + { + "epoch": 1.376455084559631, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.176948070526123, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7078797817230225, + "num_tokens": 324322960.0, + "step": 12534 + }, + { + "epoch": 1.3765649022622446, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4124042987823486, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.745918869972229, + "num_tokens": 324345051.0, + "step": 12535 + }, + { + "epoch": 1.3766747199648584, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.27323579788208, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7055853605270386, + "num_tokens": 324370134.0, + "step": 12536 + }, + { + "epoch": 1.376784537667472, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.178236722946167, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6935265064239502, + "num_tokens": 324396827.0, + "step": 12537 + }, + { + "epoch": 1.3768943553700856, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4004626274108887, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7128983736038208, + "num_tokens": 324419711.0, + "step": 12538 + }, + { + "epoch": 1.3770041730726994, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3839964866638184, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7180052995681763, + "num_tokens": 324442608.0, + "step": 12539 + }, + { + "epoch": 1.377113990775313, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4180119037628174, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7189391255378723, + "num_tokens": 324463862.0, + "step": 12540 + }, + { + "epoch": 1.3772238084779267, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4651002883911133, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7237054109573364, + "num_tokens": 324485779.0, + "step": 12541 + }, + { + "epoch": 1.3773336261805402, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.229356050491333, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6995378732681274, + "num_tokens": 324513413.0, + "step": 12542 + }, + { + "epoch": 1.377443443883154, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9718915224075317, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7275772094726562, + "num_tokens": 324543901.0, + "step": 12543 + }, + { + "epoch": 1.3775532615857675, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.485823154449463, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7056964635848999, + "num_tokens": 324566134.0, + "step": 12544 + }, + { + "epoch": 1.3776630792883813, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1558632850646973, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6933919787406921, + "num_tokens": 324594770.0, + "step": 12545 + }, + { + "epoch": 1.377772896990995, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.21955943107605, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7082089185714722, + "num_tokens": 324619443.0, + "step": 12546 + }, + { + "epoch": 1.3778827146936086, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.550034999847412, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7269062995910645, + "num_tokens": 324639903.0, + "step": 12547 + }, + { + "epoch": 1.3779925323962223, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.32647442817688, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7300204634666443, + "num_tokens": 324661797.0, + "step": 12548 + }, + { + "epoch": 1.3781023500988359, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.171441078186035, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7369842529296875, + "num_tokens": 324686076.0, + "step": 12549 + }, + { + "epoch": 1.3782121678014496, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3141143321990967, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7374634146690369, + "num_tokens": 324709315.0, + "step": 12550 + }, + { + "epoch": 1.3783219855040634, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2477755546569824, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7323168516159058, + "num_tokens": 324733142.0, + "step": 12551 + }, + { + "epoch": 1.378431803206677, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.338064193725586, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7381024956703186, + "num_tokens": 324756808.0, + "step": 12552 + }, + { + "epoch": 1.3785416209092904, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.63982892036438, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7159128189086914, + "num_tokens": 324777993.0, + "step": 12553 + }, + { + "epoch": 1.3786514386119042, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.07383131980896, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7430799603462219, + "num_tokens": 324806322.0, + "step": 12554 + }, + { + "epoch": 1.378761256314518, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 1.9894286394119263, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7026163339614868, + "num_tokens": 324837557.0, + "step": 12555 + }, + { + "epoch": 1.3788710740171315, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.6141316890716553, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7021956443786621, + "num_tokens": 324857980.0, + "step": 12556 + }, + { + "epoch": 1.3789808917197452, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.245490550994873, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7086519002914429, + "num_tokens": 324887487.0, + "step": 12557 + }, + { + "epoch": 1.3790907094223588, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.4701361656188965, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7433335781097412, + "num_tokens": 324908476.0, + "step": 12558 + }, + { + "epoch": 1.3792005271249725, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3642046451568604, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7245600819587708, + "num_tokens": 324934794.0, + "step": 12559 + }, + { + "epoch": 1.3793103448275863, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.044992446899414, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7347320318222046, + "num_tokens": 324965560.0, + "step": 12560 + }, + { + "epoch": 1.3794201625301998, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.206209421157837, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.718230664730072, + "num_tokens": 324992511.0, + "step": 12561 + }, + { + "epoch": 1.3795299802328136, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2216107845306396, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7262727618217468, + "num_tokens": 325019487.0, + "step": 12562 + }, + { + "epoch": 1.3796397979354271, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3518996238708496, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.718932032585144, + "num_tokens": 325043901.0, + "step": 12563 + }, + { + "epoch": 1.3797496156380409, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.227872848510742, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7294623255729675, + "num_tokens": 325068013.0, + "step": 12564 + }, + { + "epoch": 1.3798594333406546, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.3000965118408203, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7068670988082886, + "num_tokens": 325094329.0, + "step": 12565 + }, + { + "epoch": 1.3799692510432682, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.106275796890259, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7214053869247437, + "num_tokens": 325121449.0, + "step": 12566 + }, + { + "epoch": 1.3800790687458817, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1634130477905273, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7241158485412598, + "num_tokens": 325146118.0, + "step": 12567 + }, + { + "epoch": 1.3801888864484955, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.066678047180176, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7099413275718689, + "num_tokens": 325177571.0, + "step": 12568 + }, + { + "epoch": 1.3802987041511092, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.091331958770752, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.6950563192367554, + "num_tokens": 325207045.0, + "step": 12569 + }, + { + "epoch": 1.3804085218537228, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2258188724517822, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.714847207069397, + "num_tokens": 325233060.0, + "step": 12570 + }, + { + "epoch": 1.3805183395563365, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.362971782684326, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7320058345794678, + "num_tokens": 325255279.0, + "step": 12571 + }, + { + "epoch": 1.38062815725895, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.364367723464966, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7146058678627014, + "num_tokens": 325279494.0, + "step": 12572 + }, + { + "epoch": 1.3807379749615638, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1533291339874268, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7505224943161011, + "num_tokens": 325303432.0, + "step": 12573 + }, + { + "epoch": 1.3808477926641776, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2689208984375, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7020242810249329, + "num_tokens": 325329242.0, + "step": 12574 + }, + { + "epoch": 1.380957610366791, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2062907218933105, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7135980725288391, + "num_tokens": 325355446.0, + "step": 12575 + }, + { + "epoch": 1.3810674280694049, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.115492582321167, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7117342352867126, + "num_tokens": 325382954.0, + "step": 12576 + }, + { + "epoch": 1.3811772457720184, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.2603931427001953, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7048096656799316, + "num_tokens": 325407603.0, + "step": 12577 + }, + { + "epoch": 1.3812870634746321, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1435725688934326, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7005041837692261, + "num_tokens": 325434185.0, + "step": 12578 + }, + { + "epoch": 1.381396881177246, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.127309560775757, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7200573682785034, + "num_tokens": 325461343.0, + "step": 12579 + }, + { + "epoch": 1.3815066988798594, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.456712484359741, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7094899415969849, + "num_tokens": 325483556.0, + "step": 12580 + }, + { + "epoch": 1.381616516582473, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.24906849861145, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7089706063270569, + "num_tokens": 325510162.0, + "step": 12581 + }, + { + "epoch": 1.3817263342850867, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.350590467453003, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7512879371643066, + "num_tokens": 325533463.0, + "step": 12582 + }, + { + "epoch": 1.3818361519877005, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.239095687866211, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7460970878601074, + "num_tokens": 325558235.0, + "step": 12583 + }, + { + "epoch": 1.381945969690314, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.466049909591675, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7236170768737793, + "num_tokens": 325579757.0, + "step": 12584 + }, + { + "epoch": 1.3820557873929278, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.284864902496338, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6976801753044128, + "num_tokens": 325605573.0, + "step": 12585 + }, + { + "epoch": 1.3821656050955413, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.346740961074829, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7179343700408936, + "num_tokens": 325629148.0, + "step": 12586 + }, + { + "epoch": 1.382275422798155, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.221606492996216, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7165511846542358, + "num_tokens": 325655848.0, + "step": 12587 + }, + { + "epoch": 1.3823852405007688, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.378138780593872, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7213309407234192, + "num_tokens": 325679168.0, + "step": 12588 + }, + { + "epoch": 1.3824950582033824, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0947067737579346, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6996339559555054, + "num_tokens": 325711489.0, + "step": 12589 + }, + { + "epoch": 1.3826048759059961, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1985180377960205, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.6900741457939148, + "num_tokens": 325740922.0, + "step": 12590 + }, + { + "epoch": 1.3827146936086097, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1009230613708496, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7044151425361633, + "num_tokens": 325770207.0, + "step": 12591 + }, + { + "epoch": 1.3828245113112234, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.0589418411254883, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7415886521339417, + "num_tokens": 325798524.0, + "step": 12592 + }, + { + "epoch": 1.382934329013837, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.272385835647583, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6995394229888916, + "num_tokens": 325825687.0, + "step": 12593 + }, + { + "epoch": 1.3830441467164507, + "ewc_loss": 0.00011491775512695312, + "grad_norm": 2.1454386711120605, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7175387144088745, + "num_tokens": 325854967.0, + "step": 12594 + }, + { + "epoch": 1.3831539644190642, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 1.989880084991455, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6933728456497192, + "num_tokens": 325887490.0, + "step": 12595 + }, + { + "epoch": 1.383263782121678, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.5934364795684814, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.705251932144165, + "num_tokens": 325908570.0, + "step": 12596 + }, + { + "epoch": 1.3833735998242918, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.587401866912842, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7205631732940674, + "num_tokens": 325928528.0, + "step": 12597 + }, + { + "epoch": 1.3834834175269053, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.1747796535491943, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7058817148208618, + "num_tokens": 325956713.0, + "step": 12598 + }, + { + "epoch": 1.383593235229519, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.134045362472534, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7139197587966919, + "num_tokens": 325984839.0, + "step": 12599 + }, + { + "epoch": 1.3837030529321326, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.101304531097412, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7275996208190918, + "num_tokens": 326011963.0, + "step": 12600 + }, + { + "epoch": 1.3838128706347463, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.275416612625122, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7245426774024963, + "num_tokens": 326037627.0, + "step": 12601 + }, + { + "epoch": 1.38392268833736, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 1.9399311542510986, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7162982225418091, + "num_tokens": 326068689.0, + "step": 12602 + }, + { + "epoch": 1.3840325060399736, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.1478915214538574, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7108224630355835, + "num_tokens": 326099103.0, + "step": 12603 + }, + { + "epoch": 1.3841423237425874, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.282867193222046, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6882556676864624, + "num_tokens": 326125592.0, + "step": 12604 + }, + { + "epoch": 1.384252141445201, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.0556752681732178, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7169850468635559, + "num_tokens": 326154108.0, + "step": 12605 + }, + { + "epoch": 1.3843619591478147, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.32985520362854, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7245607376098633, + "num_tokens": 326178614.0, + "step": 12606 + }, + { + "epoch": 1.3844717768504282, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.147266387939453, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7020529508590698, + "num_tokens": 326205864.0, + "step": 12607 + }, + { + "epoch": 1.384581594553042, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.4043545722961426, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7115590572357178, + "num_tokens": 326229546.0, + "step": 12608 + }, + { + "epoch": 1.3846914122556555, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0868980884552, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.704492449760437, + "num_tokens": 326257447.0, + "step": 12609 + }, + { + "epoch": 1.3848012299582693, + "ewc_loss": 0.00011539459228515625, + "grad_norm": 2.359158754348755, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7233843803405762, + "num_tokens": 326279636.0, + "step": 12610 + }, + { + "epoch": 1.384911047660883, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.139406681060791, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7054147720336914, + "num_tokens": 326307847.0, + "step": 12611 + }, + { + "epoch": 1.3850208653634966, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.323380947113037, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7110595703125, + "num_tokens": 326334739.0, + "step": 12612 + }, + { + "epoch": 1.3851306830661103, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.651639938354492, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7428621053695679, + "num_tokens": 326354685.0, + "step": 12613 + }, + { + "epoch": 1.3852405007687238, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.12335467338562, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.6996492743492126, + "num_tokens": 326384519.0, + "step": 12614 + }, + { + "epoch": 1.3853503184713376, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4482102394104004, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7388997077941895, + "num_tokens": 326407163.0, + "step": 12615 + }, + { + "epoch": 1.3854601361739514, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.246682643890381, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7120780348777771, + "num_tokens": 326432569.0, + "step": 12616 + }, + { + "epoch": 1.385569953876565, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.146436929702759, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7005083560943604, + "num_tokens": 326462060.0, + "step": 12617 + }, + { + "epoch": 1.3856797715791784, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2548556327819824, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7306252717971802, + "num_tokens": 326487727.0, + "step": 12618 + }, + { + "epoch": 1.3857895892817922, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4452202320098877, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7325407266616821, + "num_tokens": 326510858.0, + "step": 12619 + }, + { + "epoch": 1.385899406984406, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4119009971618652, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7284946441650391, + "num_tokens": 326532694.0, + "step": 12620 + }, + { + "epoch": 1.3860092246870195, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.18410325050354, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7278652191162109, + "num_tokens": 326560597.0, + "step": 12621 + }, + { + "epoch": 1.3861190423896332, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.5191352367401123, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.690321683883667, + "num_tokens": 326582964.0, + "step": 12622 + }, + { + "epoch": 1.3862288600922468, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1950855255126953, + "learning_rate": 1e-06, + "loss": 1.0291, + "mean_token_accuracy": 0.7024773359298706, + "num_tokens": 326609534.0, + "step": 12623 + }, + { + "epoch": 1.3863386777948605, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2722482681274414, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7346827983856201, + "num_tokens": 326634252.0, + "step": 12624 + }, + { + "epoch": 1.3864484954974743, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3330881595611572, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7331098318099976, + "num_tokens": 326658361.0, + "step": 12625 + }, + { + "epoch": 1.3865583132000878, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.487762689590454, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7349145412445068, + "num_tokens": 326678597.0, + "step": 12626 + }, + { + "epoch": 1.3866681309027016, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1711537837982178, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7068488597869873, + "num_tokens": 326707463.0, + "step": 12627 + }, + { + "epoch": 1.3867779486053151, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.031102418899536, + "learning_rate": 1e-06, + "loss": 1.0409, + "mean_token_accuracy": 0.6846640110015869, + "num_tokens": 326740808.0, + "step": 12628 + }, + { + "epoch": 1.3868877663079289, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0425169467926025, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.710475504398346, + "num_tokens": 326770351.0, + "step": 12629 + }, + { + "epoch": 1.3869975840105426, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2126777172088623, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.727523148059845, + "num_tokens": 326795628.0, + "step": 12630 + }, + { + "epoch": 1.3871074017131562, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.022373676300049, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7099207043647766, + "num_tokens": 326828612.0, + "step": 12631 + }, + { + "epoch": 1.3872172194157697, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3643689155578613, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.728753924369812, + "num_tokens": 326852073.0, + "step": 12632 + }, + { + "epoch": 1.3873270371183835, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.344928503036499, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7441606521606445, + "num_tokens": 326877643.0, + "step": 12633 + }, + { + "epoch": 1.3874368548209972, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.296109914779663, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7084331512451172, + "num_tokens": 326902042.0, + "step": 12634 + }, + { + "epoch": 1.3875466725236107, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2616162300109863, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.717180073261261, + "num_tokens": 326927492.0, + "step": 12635 + }, + { + "epoch": 1.3876564902262245, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.169710159301758, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7169329524040222, + "num_tokens": 326955640.0, + "step": 12636 + }, + { + "epoch": 1.387766307928838, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.131239891052246, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7123850584030151, + "num_tokens": 326982934.0, + "step": 12637 + }, + { + "epoch": 1.3878761256314518, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.6716554164886475, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7104390859603882, + "num_tokens": 327003284.0, + "step": 12638 + }, + { + "epoch": 1.3879859433340656, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9285715818405151, + "learning_rate": 1e-06, + "loss": 1.0351, + "mean_token_accuracy": 0.6844144463539124, + "num_tokens": 327039297.0, + "step": 12639 + }, + { + "epoch": 1.388095761036679, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1588242053985596, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6911706924438477, + "num_tokens": 327068485.0, + "step": 12640 + }, + { + "epoch": 1.3882055787392928, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2239410877227783, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7147279381752014, + "num_tokens": 327094322.0, + "step": 12641 + }, + { + "epoch": 1.3883153964419064, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.349281072616577, + "learning_rate": 1e-06, + "loss": 0.7969, + "mean_token_accuracy": 0.7499645948410034, + "num_tokens": 327118567.0, + "step": 12642 + }, + { + "epoch": 1.3884252141445201, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1847126483917236, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7198992371559143, + "num_tokens": 327144865.0, + "step": 12643 + }, + { + "epoch": 1.388535031847134, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9922376871109009, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7092795372009277, + "num_tokens": 327174998.0, + "step": 12644 + }, + { + "epoch": 1.3886448495497474, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2344889640808105, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7028192281723022, + "num_tokens": 327200509.0, + "step": 12645 + }, + { + "epoch": 1.388754667252361, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.191486120223999, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7293825745582581, + "num_tokens": 327225745.0, + "step": 12646 + }, + { + "epoch": 1.3888644849549747, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.092297077178955, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.6959173083305359, + "num_tokens": 327254984.0, + "step": 12647 + }, + { + "epoch": 1.3889743026575885, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1889212131500244, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7268491983413696, + "num_tokens": 327279223.0, + "step": 12648 + }, + { + "epoch": 1.389084120360202, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4667296409606934, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7201112508773804, + "num_tokens": 327303319.0, + "step": 12649 + }, + { + "epoch": 1.3891939380628158, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.089768886566162, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.71146559715271, + "num_tokens": 327331675.0, + "step": 12650 + }, + { + "epoch": 1.3893037557654293, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.321085214614868, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7185813188552856, + "num_tokens": 327355990.0, + "step": 12651 + }, + { + "epoch": 1.389413573468043, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4046401977539062, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7298611402511597, + "num_tokens": 327379324.0, + "step": 12652 + }, + { + "epoch": 1.3895233911706568, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4035985469818115, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7083348631858826, + "num_tokens": 327402498.0, + "step": 12653 + }, + { + "epoch": 1.3896332088732704, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3570194244384766, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7108963131904602, + "num_tokens": 327426181.0, + "step": 12654 + }, + { + "epoch": 1.389743026575884, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0310027599334717, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.711712121963501, + "num_tokens": 327455846.0, + "step": 12655 + }, + { + "epoch": 1.3898528442784976, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.194880247116089, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6866884231567383, + "num_tokens": 327486203.0, + "step": 12656 + }, + { + "epoch": 1.3899626619811114, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2590789794921875, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7408409118652344, + "num_tokens": 327510687.0, + "step": 12657 + }, + { + "epoch": 1.390072479683725, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2358598709106445, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.719916045665741, + "num_tokens": 327536849.0, + "step": 12658 + }, + { + "epoch": 1.3901822973863387, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3557920455932617, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.712660014629364, + "num_tokens": 327559433.0, + "step": 12659 + }, + { + "epoch": 1.3902921150889522, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4306607246398926, + "learning_rate": 1e-06, + "loss": 0.7358, + "mean_token_accuracy": 0.7640535831451416, + "num_tokens": 327580154.0, + "step": 12660 + }, + { + "epoch": 1.390401932791566, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4362123012542725, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7355491518974304, + "num_tokens": 327602813.0, + "step": 12661 + }, + { + "epoch": 1.3905117504941797, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2940406799316406, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7334584593772888, + "num_tokens": 327629603.0, + "step": 12662 + }, + { + "epoch": 1.3906215681967933, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3977108001708984, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.710636556148529, + "num_tokens": 327654648.0, + "step": 12663 + }, + { + "epoch": 1.390731385899407, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0909435749053955, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7033208012580872, + "num_tokens": 327684120.0, + "step": 12664 + }, + { + "epoch": 1.3908412036020206, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.461113452911377, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7128622531890869, + "num_tokens": 327707176.0, + "step": 12665 + }, + { + "epoch": 1.3909510213046343, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.16060209274292, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7001765966415405, + "num_tokens": 327734720.0, + "step": 12666 + }, + { + "epoch": 1.391060839007248, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2205960750579834, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7193962335586548, + "num_tokens": 327760150.0, + "step": 12667 + }, + { + "epoch": 1.3911706567098616, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2293636798858643, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.719231367111206, + "num_tokens": 327784149.0, + "step": 12668 + }, + { + "epoch": 1.3912804744124752, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3305811882019043, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7298595905303955, + "num_tokens": 327808125.0, + "step": 12669 + }, + { + "epoch": 1.391390292115089, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3794760704040527, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7355644702911377, + "num_tokens": 327830077.0, + "step": 12670 + }, + { + "epoch": 1.3915001098177027, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.202136993408203, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7166441679000854, + "num_tokens": 327857461.0, + "step": 12671 + }, + { + "epoch": 1.3916099275203162, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.595287561416626, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7247301936149597, + "num_tokens": 327877176.0, + "step": 12672 + }, + { + "epoch": 1.39171974522293, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.7247982025146484, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7430557012557983, + "num_tokens": 327895395.0, + "step": 12673 + }, + { + "epoch": 1.3918295629255435, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9609429836273193, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7255611419677734, + "num_tokens": 327927128.0, + "step": 12674 + }, + { + "epoch": 1.3919393806281573, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2879011631011963, + "learning_rate": 1e-06, + "loss": 0.828, + "mean_token_accuracy": 0.7425498962402344, + "num_tokens": 327951863.0, + "step": 12675 + }, + { + "epoch": 1.392049198330771, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1893839836120605, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7330331802368164, + "num_tokens": 327975457.0, + "step": 12676 + }, + { + "epoch": 1.3921590160333845, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.238114833831787, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7103655338287354, + "num_tokens": 328001139.0, + "step": 12677 + }, + { + "epoch": 1.3922688337359983, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2352709770202637, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7084609270095825, + "num_tokens": 328027431.0, + "step": 12678 + }, + { + "epoch": 1.3923786514386118, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2028889656066895, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7008824348449707, + "num_tokens": 328054469.0, + "step": 12679 + }, + { + "epoch": 1.3924884691412256, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.999617338180542, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7077207565307617, + "num_tokens": 328086474.0, + "step": 12680 + }, + { + "epoch": 1.3925982868438394, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.178732395172119, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7063239216804504, + "num_tokens": 328114272.0, + "step": 12681 + }, + { + "epoch": 1.3927081045464529, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3352103233337402, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7145791053771973, + "num_tokens": 328139879.0, + "step": 12682 + }, + { + "epoch": 1.3928179222490664, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.594172954559326, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7061682939529419, + "num_tokens": 328163458.0, + "step": 12683 + }, + { + "epoch": 1.3929277399516802, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2236719131469727, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7178260684013367, + "num_tokens": 328189557.0, + "step": 12684 + }, + { + "epoch": 1.393037557654294, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.203423023223877, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.710239827632904, + "num_tokens": 328216781.0, + "step": 12685 + }, + { + "epoch": 1.3931473753569075, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.068206548690796, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7277210354804993, + "num_tokens": 328248094.0, + "step": 12686 + }, + { + "epoch": 1.3932571930595212, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9790375232696533, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6982195377349854, + "num_tokens": 328282358.0, + "step": 12687 + }, + { + "epoch": 1.3933670107621348, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9809610843658447, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.7004519701004028, + "num_tokens": 328318452.0, + "step": 12688 + }, + { + "epoch": 1.3934768284647485, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.244460105895996, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7038180828094482, + "num_tokens": 328344819.0, + "step": 12689 + }, + { + "epoch": 1.3935866461673623, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.5818519592285156, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6990265250205994, + "num_tokens": 328366894.0, + "step": 12690 + }, + { + "epoch": 1.3936964638699758, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.333611488342285, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7113442420959473, + "num_tokens": 328393601.0, + "step": 12691 + }, + { + "epoch": 1.3938062815725896, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.189352035522461, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7244210243225098, + "num_tokens": 328419230.0, + "step": 12692 + }, + { + "epoch": 1.393916099275203, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4394338130950928, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7107480764389038, + "num_tokens": 328443372.0, + "step": 12693 + }, + { + "epoch": 1.3940259169778169, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.364079713821411, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7285566329956055, + "num_tokens": 328465375.0, + "step": 12694 + }, + { + "epoch": 1.3941357346804306, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.028740167617798, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7151296138763428, + "num_tokens": 328497099.0, + "step": 12695 + }, + { + "epoch": 1.3942455523830442, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.389580249786377, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7426310777664185, + "num_tokens": 328517455.0, + "step": 12696 + }, + { + "epoch": 1.3943553700856577, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2406301498413086, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7038648128509521, + "num_tokens": 328542932.0, + "step": 12697 + }, + { + "epoch": 1.3944651877882714, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4334359169006348, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.729283332824707, + "num_tokens": 328566209.0, + "step": 12698 + }, + { + "epoch": 1.3945750054908852, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.139775276184082, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7192130088806152, + "num_tokens": 328591339.0, + "step": 12699 + }, + { + "epoch": 1.3946848231934987, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 1.9683666229248047, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6922863721847534, + "num_tokens": 328624959.0, + "step": 12700 + }, + { + "epoch": 1.3947946408961125, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.266334056854248, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6898456811904907, + "num_tokens": 328653922.0, + "step": 12701 + }, + { + "epoch": 1.394904458598726, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.482715606689453, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.73015958070755, + "num_tokens": 328674331.0, + "step": 12702 + }, + { + "epoch": 1.3950142763013398, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2265567779541016, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7330612540245056, + "num_tokens": 328698819.0, + "step": 12703 + }, + { + "epoch": 1.3951240940039535, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0833520889282227, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7162015438079834, + "num_tokens": 328727751.0, + "step": 12704 + }, + { + "epoch": 1.395233911706567, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4556047916412354, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7353129982948303, + "num_tokens": 328749952.0, + "step": 12705 + }, + { + "epoch": 1.3953437294091808, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.216348648071289, + "learning_rate": 1e-06, + "loss": 0.799, + "mean_token_accuracy": 0.7418886423110962, + "num_tokens": 328773941.0, + "step": 12706 + }, + { + "epoch": 1.3954535471117944, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3494174480438232, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6948763728141785, + "num_tokens": 328799129.0, + "step": 12707 + }, + { + "epoch": 1.3955633648144081, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.102003574371338, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7167274951934814, + "num_tokens": 328826607.0, + "step": 12708 + }, + { + "epoch": 1.3956731825170219, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.322430372238159, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7424657940864563, + "num_tokens": 328851609.0, + "step": 12709 + }, + { + "epoch": 1.3957830002196354, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1455466747283936, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7162736058235168, + "num_tokens": 328878492.0, + "step": 12710 + }, + { + "epoch": 1.395892817922249, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.336186647415161, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7221972346305847, + "num_tokens": 328903482.0, + "step": 12711 + }, + { + "epoch": 1.3960026356248627, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.394835948944092, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7262654900550842, + "num_tokens": 328925219.0, + "step": 12712 + }, + { + "epoch": 1.3961124533274765, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.255571126937866, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7142894268035889, + "num_tokens": 328949992.0, + "step": 12713 + }, + { + "epoch": 1.39622227103009, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.528618574142456, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7310593724250793, + "num_tokens": 328972712.0, + "step": 12714 + }, + { + "epoch": 1.3963320887327038, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.253257989883423, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.6987987160682678, + "num_tokens": 329000792.0, + "step": 12715 + }, + { + "epoch": 1.3964419064353173, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.14050555229187, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7252662181854248, + "num_tokens": 329028808.0, + "step": 12716 + }, + { + "epoch": 1.396551724137931, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2417171001434326, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7239845395088196, + "num_tokens": 329056023.0, + "step": 12717 + }, + { + "epoch": 1.3966615418405448, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.0642638206481934, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7006279230117798, + "num_tokens": 329087466.0, + "step": 12718 + }, + { + "epoch": 1.3967713595431583, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.2694132328033447, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7165892720222473, + "num_tokens": 329113202.0, + "step": 12719 + }, + { + "epoch": 1.396881177245772, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.3223507404327393, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7042282223701477, + "num_tokens": 329138724.0, + "step": 12720 + }, + { + "epoch": 1.3969909949483856, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.250762939453125, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.7021064162254333, + "num_tokens": 329165168.0, + "step": 12721 + }, + { + "epoch": 1.3971008126509994, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.3458855152130127, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7075166702270508, + "num_tokens": 329189652.0, + "step": 12722 + }, + { + "epoch": 1.397210630353613, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4032094478607178, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7530286312103271, + "num_tokens": 329211107.0, + "step": 12723 + }, + { + "epoch": 1.3973204480562267, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.484160900115967, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7112383842468262, + "num_tokens": 329233763.0, + "step": 12724 + }, + { + "epoch": 1.3974302657588402, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.4282500743865967, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.716544508934021, + "num_tokens": 329256232.0, + "step": 12725 + }, + { + "epoch": 1.397540083461454, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.4727509021759033, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7190876007080078, + "num_tokens": 329278118.0, + "step": 12726 + }, + { + "epoch": 1.3976499011640677, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.22432279586792, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7422827482223511, + "num_tokens": 329306441.0, + "step": 12727 + }, + { + "epoch": 1.3977597188666813, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.0835602283477783, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7065722346305847, + "num_tokens": 329338047.0, + "step": 12728 + }, + { + "epoch": 1.397869536569295, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.4581804275512695, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7319105267524719, + "num_tokens": 329361361.0, + "step": 12729 + }, + { + "epoch": 1.3979793542719086, + "ewc_loss": 0.0001163482666015625, + "grad_norm": 2.551039218902588, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7291811108589172, + "num_tokens": 329382056.0, + "step": 12730 + }, + { + "epoch": 1.3980891719745223, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.5027999877929688, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7272443771362305, + "num_tokens": 329404022.0, + "step": 12731 + }, + { + "epoch": 1.398198989677136, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2001712322235107, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7103297114372253, + "num_tokens": 329430905.0, + "step": 12732 + }, + { + "epoch": 1.3983088073797496, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3784964084625244, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7235589027404785, + "num_tokens": 329453502.0, + "step": 12733 + }, + { + "epoch": 1.3984186250823631, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1882219314575195, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7055724859237671, + "num_tokens": 329482453.0, + "step": 12734 + }, + { + "epoch": 1.398528442784977, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.044689416885376, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6909829378128052, + "num_tokens": 329514440.0, + "step": 12735 + }, + { + "epoch": 1.3986382604875907, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.5817859172821045, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7603070735931396, + "num_tokens": 329534295.0, + "step": 12736 + }, + { + "epoch": 1.3987480781902042, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.5284996032714844, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7092698812484741, + "num_tokens": 329555527.0, + "step": 12737 + }, + { + "epoch": 1.398857895892818, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1206459999084473, + "learning_rate": 1e-06, + "loss": 1.0314, + "mean_token_accuracy": 0.6829700469970703, + "num_tokens": 329586322.0, + "step": 12738 + }, + { + "epoch": 1.3989677135954315, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.3011810779571533, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7180251479148865, + "num_tokens": 329611497.0, + "step": 12739 + }, + { + "epoch": 1.3990775312980452, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.261350154876709, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.723272442817688, + "num_tokens": 329636836.0, + "step": 12740 + }, + { + "epoch": 1.399187349000659, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.4070656299591064, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.720307469367981, + "num_tokens": 329661010.0, + "step": 12741 + }, + { + "epoch": 1.3992971667032725, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.514448404312134, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.721289873123169, + "num_tokens": 329682644.0, + "step": 12742 + }, + { + "epoch": 1.3994069844058863, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.2902016639709473, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7268933057785034, + "num_tokens": 329707774.0, + "step": 12743 + }, + { + "epoch": 1.3995168021084998, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1043851375579834, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7209304571151733, + "num_tokens": 329735999.0, + "step": 12744 + }, + { + "epoch": 1.3996266198111136, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.160297393798828, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.697662353515625, + "num_tokens": 329765151.0, + "step": 12745 + }, + { + "epoch": 1.3997364375137273, + "ewc_loss": 0.00011587142944335938, + "grad_norm": 2.1902666091918945, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7388900518417358, + "num_tokens": 329791205.0, + "step": 12746 + }, + { + "epoch": 1.3998462552163409, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.3222413063049316, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.69527268409729, + "num_tokens": 329815303.0, + "step": 12747 + }, + { + "epoch": 1.3999560729189544, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.342637538909912, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7339518666267395, + "num_tokens": 329838161.0, + "step": 12748 + }, + { + "epoch": 1.4000658906215682, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1200058460235596, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6963179707527161, + "num_tokens": 329869874.0, + "step": 12749 + }, + { + "epoch": 1.400175708324182, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.535374402999878, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7416229248046875, + "num_tokens": 329890223.0, + "step": 12750 + }, + { + "epoch": 1.4002855260267955, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1210858821868896, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7165910005569458, + "num_tokens": 329918639.0, + "step": 12751 + }, + { + "epoch": 1.4003953437294092, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.48942494392395, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7093757390975952, + "num_tokens": 329943018.0, + "step": 12752 + }, + { + "epoch": 1.4005051614320227, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.3088631629943848, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7234670519828796, + "num_tokens": 329968118.0, + "step": 12753 + }, + { + "epoch": 1.4006149791346365, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1436405181884766, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7283519506454468, + "num_tokens": 329996407.0, + "step": 12754 + }, + { + "epoch": 1.4007247968372503, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2575523853302, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7040784955024719, + "num_tokens": 330020797.0, + "step": 12755 + }, + { + "epoch": 1.4008346145398638, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.314622163772583, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.71169114112854, + "num_tokens": 330046151.0, + "step": 12756 + }, + { + "epoch": 1.4009444322424776, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.5276193618774414, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.737834632396698, + "num_tokens": 330067459.0, + "step": 12757 + }, + { + "epoch": 1.401054249945091, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.3450071811676025, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7339398860931396, + "num_tokens": 330090836.0, + "step": 12758 + }, + { + "epoch": 1.4011640676477048, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.5813119411468506, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6950780153274536, + "num_tokens": 330112581.0, + "step": 12759 + }, + { + "epoch": 1.4012738853503186, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.412287473678589, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7171638011932373, + "num_tokens": 330134770.0, + "step": 12760 + }, + { + "epoch": 1.4013837030529321, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2579705715179443, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.688562273979187, + "num_tokens": 330162305.0, + "step": 12761 + }, + { + "epoch": 1.4014935207555457, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 1.954623818397522, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.6943332552909851, + "num_tokens": 330195513.0, + "step": 12762 + }, + { + "epoch": 1.4016033384581594, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1623330116271973, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7016823291778564, + "num_tokens": 330224612.0, + "step": 12763 + }, + { + "epoch": 1.4017131561607732, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.4119927883148193, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7029628157615662, + "num_tokens": 330245872.0, + "step": 12764 + }, + { + "epoch": 1.4018229738633867, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.886207342147827, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7294000387191772, + "num_tokens": 330263478.0, + "step": 12765 + }, + { + "epoch": 1.4019327915660005, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1704797744750977, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7198446989059448, + "num_tokens": 330292579.0, + "step": 12766 + }, + { + "epoch": 1.402042609268614, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.4210307598114014, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7155616283416748, + "num_tokens": 330315212.0, + "step": 12767 + }, + { + "epoch": 1.4021524269712278, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.12888503074646, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7139269709587097, + "num_tokens": 330343511.0, + "step": 12768 + }, + { + "epoch": 1.4022622446738415, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2121434211730957, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7276373505592346, + "num_tokens": 330371013.0, + "step": 12769 + }, + { + "epoch": 1.402372062376455, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.498573064804077, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7495828866958618, + "num_tokens": 330390565.0, + "step": 12770 + }, + { + "epoch": 1.4024818800790688, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.3860225677490234, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7411555051803589, + "num_tokens": 330414047.0, + "step": 12771 + }, + { + "epoch": 1.4025916977816824, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.204439401626587, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7220145463943481, + "num_tokens": 330440831.0, + "step": 12772 + }, + { + "epoch": 1.4027015154842961, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.5386364459991455, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7462339401245117, + "num_tokens": 330460607.0, + "step": 12773 + }, + { + "epoch": 1.4028113331869096, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 1.9823834896087646, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.69597327709198, + "num_tokens": 330493053.0, + "step": 12774 + }, + { + "epoch": 1.4029211508895234, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 1.8651386499404907, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7186516523361206, + "num_tokens": 330525499.0, + "step": 12775 + }, + { + "epoch": 1.403030968592137, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2485527992248535, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6871473789215088, + "num_tokens": 330552529.0, + "step": 12776 + }, + { + "epoch": 1.4031407862947507, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.027644634246826, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6960486769676208, + "num_tokens": 330584224.0, + "step": 12777 + }, + { + "epoch": 1.4032506039973645, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.336714267730713, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.718437910079956, + "num_tokens": 330608743.0, + "step": 12778 + }, + { + "epoch": 1.403360421699978, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.122495174407959, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7153081893920898, + "num_tokens": 330637403.0, + "step": 12779 + }, + { + "epoch": 1.4034702394025917, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.0985002517700195, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7156873941421509, + "num_tokens": 330669501.0, + "step": 12780 + }, + { + "epoch": 1.4035800571052053, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.0550971031188965, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7082564830780029, + "num_tokens": 330701549.0, + "step": 12781 + }, + { + "epoch": 1.403689874807819, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.1679718494415283, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.7030175924301147, + "num_tokens": 330729181.0, + "step": 12782 + }, + { + "epoch": 1.4037996925104328, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.157768487930298, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.7002131938934326, + "num_tokens": 330757862.0, + "step": 12783 + }, + { + "epoch": 1.4039095102130463, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2765953540802, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7143832445144653, + "num_tokens": 330783695.0, + "step": 12784 + }, + { + "epoch": 1.40401932791566, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 1.882082223892212, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7070887684822083, + "num_tokens": 330817814.0, + "step": 12785 + }, + { + "epoch": 1.4041291456182736, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2229957580566406, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7270681858062744, + "num_tokens": 330844176.0, + "step": 12786 + }, + { + "epoch": 1.4042389633208874, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.2360031604766846, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7071254253387451, + "num_tokens": 330870796.0, + "step": 12787 + }, + { + "epoch": 1.404348781023501, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.365727186203003, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7178714871406555, + "num_tokens": 330894438.0, + "step": 12788 + }, + { + "epoch": 1.4044585987261147, + "ewc_loss": 0.00011682510375976562, + "grad_norm": 2.336678981781006, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7058799862861633, + "num_tokens": 330921480.0, + "step": 12789 + }, + { + "epoch": 1.4045684164287282, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.333512783050537, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7073960304260254, + "num_tokens": 330946772.0, + "step": 12790 + }, + { + "epoch": 1.404678234131342, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0399277210235596, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7172774076461792, + "num_tokens": 330979370.0, + "step": 12791 + }, + { + "epoch": 1.4047880518339557, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.316020965576172, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7025342583656311, + "num_tokens": 331002954.0, + "step": 12792 + }, + { + "epoch": 1.4048978695365693, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.957016110420227, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6934885382652283, + "num_tokens": 331035129.0, + "step": 12793 + }, + { + "epoch": 1.405007687239183, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.363776922225952, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7117450833320618, + "num_tokens": 331059434.0, + "step": 12794 + }, + { + "epoch": 1.4051175049417965, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.576082706451416, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7037053108215332, + "num_tokens": 331080140.0, + "step": 12795 + }, + { + "epoch": 1.4052273226444103, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1923394203186035, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7265419960021973, + "num_tokens": 331106893.0, + "step": 12796 + }, + { + "epoch": 1.405337140347024, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.58689284324646, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7466609477996826, + "num_tokens": 331126529.0, + "step": 12797 + }, + { + "epoch": 1.4054469580496376, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0469324588775635, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7044067978858948, + "num_tokens": 331158566.0, + "step": 12798 + }, + { + "epoch": 1.4055567757522511, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.107490062713623, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7150895595550537, + "num_tokens": 331186311.0, + "step": 12799 + }, + { + "epoch": 1.4056665934548649, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0349936485290527, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7412716746330261, + "num_tokens": 331215102.0, + "step": 12800 + }, + { + "epoch": 1.4057764111574786, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.505103349685669, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7101367712020874, + "num_tokens": 331236980.0, + "step": 12801 + }, + { + "epoch": 1.4058862288600922, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4081037044525146, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6978124380111694, + "num_tokens": 331261948.0, + "step": 12802 + }, + { + "epoch": 1.405996046562706, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2602152824401855, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7102739810943604, + "num_tokens": 331288192.0, + "step": 12803 + }, + { + "epoch": 1.4061058642653195, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0488076210021973, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7192826271057129, + "num_tokens": 331317084.0, + "step": 12804 + }, + { + "epoch": 1.4062156819679332, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.033942937850952, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7358622550964355, + "num_tokens": 331345302.0, + "step": 12805 + }, + { + "epoch": 1.406325499670547, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1656787395477295, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7204233407974243, + "num_tokens": 331372262.0, + "step": 12806 + }, + { + "epoch": 1.4064353173731605, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.191779851913452, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7284473776817322, + "num_tokens": 331398490.0, + "step": 12807 + }, + { + "epoch": 1.4065451350757743, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.204315662384033, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6956596374511719, + "num_tokens": 331424470.0, + "step": 12808 + }, + { + "epoch": 1.4066549527783878, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3536863327026367, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7100658416748047, + "num_tokens": 331447838.0, + "step": 12809 + }, + { + "epoch": 1.4067647704810016, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.526304244995117, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7133049964904785, + "num_tokens": 331470010.0, + "step": 12810 + }, + { + "epoch": 1.4068745881836153, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1640419960021973, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7010625004768372, + "num_tokens": 331496230.0, + "step": 12811 + }, + { + "epoch": 1.4069844058862289, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1576733589172363, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.692333459854126, + "num_tokens": 331525604.0, + "step": 12812 + }, + { + "epoch": 1.4070942235888424, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1221630573272705, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7155827283859253, + "num_tokens": 331553471.0, + "step": 12813 + }, + { + "epoch": 1.4072040412914562, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4087398052215576, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7365750074386597, + "num_tokens": 331575373.0, + "step": 12814 + }, + { + "epoch": 1.40731385899407, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.056030750274658, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7134853601455688, + "num_tokens": 331604945.0, + "step": 12815 + }, + { + "epoch": 1.4074236766966834, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1693742275238037, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6999003887176514, + "num_tokens": 331632112.0, + "step": 12816 + }, + { + "epoch": 1.4075334943992972, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0378458499908447, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7233544588088989, + "num_tokens": 331659631.0, + "step": 12817 + }, + { + "epoch": 1.4076433121019107, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3850200176239014, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7394764423370361, + "num_tokens": 331681471.0, + "step": 12818 + }, + { + "epoch": 1.4077531298045245, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.357009172439575, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7373207211494446, + "num_tokens": 331704106.0, + "step": 12819 + }, + { + "epoch": 1.4078629475071383, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.029794216156006, + "learning_rate": 1e-06, + "loss": 1.0035, + "mean_token_accuracy": 0.6952673196792603, + "num_tokens": 331736036.0, + "step": 12820 + }, + { + "epoch": 1.4079727652097518, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3173294067382812, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7162114977836609, + "num_tokens": 331761772.0, + "step": 12821 + }, + { + "epoch": 1.4080825829123655, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3515186309814453, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7228957414627075, + "num_tokens": 331785723.0, + "step": 12822 + }, + { + "epoch": 1.408192400614979, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3833725452423096, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.722097635269165, + "num_tokens": 331809006.0, + "step": 12823 + }, + { + "epoch": 1.4083022183175928, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5635573863983154, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7375954389572144, + "num_tokens": 331830095.0, + "step": 12824 + }, + { + "epoch": 1.4084120360202066, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.389223575592041, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7451034784317017, + "num_tokens": 331852970.0, + "step": 12825 + }, + { + "epoch": 1.4085218537228201, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.03644061088562, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7327495813369751, + "num_tokens": 331880021.0, + "step": 12826 + }, + { + "epoch": 1.4086316714254337, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.091002941131592, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6916876435279846, + "num_tokens": 331909539.0, + "step": 12827 + }, + { + "epoch": 1.4087414891280474, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.092820882797241, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6907393336296082, + "num_tokens": 331938496.0, + "step": 12828 + }, + { + "epoch": 1.4088513068306612, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0147523880004883, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7312825918197632, + "num_tokens": 331970350.0, + "step": 12829 + }, + { + "epoch": 1.4089611245332747, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4802286624908447, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7191355228424072, + "num_tokens": 331991869.0, + "step": 12830 + }, + { + "epoch": 1.4090709422358885, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5261929035186768, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7223358750343323, + "num_tokens": 332013233.0, + "step": 12831 + }, + { + "epoch": 1.409180759938502, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1934993267059326, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7111254930496216, + "num_tokens": 332039007.0, + "step": 12832 + }, + { + "epoch": 1.4092905776411158, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.24051570892334, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7136034965515137, + "num_tokens": 332066044.0, + "step": 12833 + }, + { + "epoch": 1.4094003953437295, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.166571617126465, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7450447678565979, + "num_tokens": 332089319.0, + "step": 12834 + }, + { + "epoch": 1.409510213046343, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.60200572013855, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7212460041046143, + "num_tokens": 332109374.0, + "step": 12835 + }, + { + "epoch": 1.4096200307489568, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1456241607666016, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7121332287788391, + "num_tokens": 332137601.0, + "step": 12836 + }, + { + "epoch": 1.4097298484515703, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3437867164611816, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7144857048988342, + "num_tokens": 332162374.0, + "step": 12837 + }, + { + "epoch": 1.409839666154184, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0776796340942383, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7114245891571045, + "num_tokens": 332191028.0, + "step": 12838 + }, + { + "epoch": 1.4099494838567976, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4483065605163574, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7248222827911377, + "num_tokens": 332212579.0, + "step": 12839 + }, + { + "epoch": 1.4100593015594114, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.244767189025879, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7250961065292358, + "num_tokens": 332240660.0, + "step": 12840 + }, + { + "epoch": 1.410169119262025, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.086209535598755, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7083165645599365, + "num_tokens": 332271878.0, + "step": 12841 + }, + { + "epoch": 1.4102789369646387, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.064946174621582, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7224043607711792, + "num_tokens": 332300385.0, + "step": 12842 + }, + { + "epoch": 1.4103887546672524, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3020670413970947, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7227402329444885, + "num_tokens": 332324927.0, + "step": 12843 + }, + { + "epoch": 1.410498572369866, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.525036096572876, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7205098867416382, + "num_tokens": 332347441.0, + "step": 12844 + }, + { + "epoch": 1.4106083900724797, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.6681180000305176, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7227441668510437, + "num_tokens": 332367849.0, + "step": 12845 + }, + { + "epoch": 1.4107182077750933, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2620768547058105, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7084643840789795, + "num_tokens": 332394262.0, + "step": 12846 + }, + { + "epoch": 1.410828025477707, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4005448818206787, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.7001018524169922, + "num_tokens": 332417902.0, + "step": 12847 + }, + { + "epoch": 1.4109378431803208, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.9664615392684937, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7070949077606201, + "num_tokens": 332450973.0, + "step": 12848 + }, + { + "epoch": 1.4110476608829343, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3289921283721924, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7191462516784668, + "num_tokens": 332474891.0, + "step": 12849 + }, + { + "epoch": 1.4111574785855479, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.196920156478882, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7148916721343994, + "num_tokens": 332501341.0, + "step": 12850 + }, + { + "epoch": 1.4112672962881616, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2089571952819824, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.709354817867279, + "num_tokens": 332529082.0, + "step": 12851 + }, + { + "epoch": 1.4113771139907754, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.274604320526123, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6891990900039673, + "num_tokens": 332556574.0, + "step": 12852 + }, + { + "epoch": 1.411486931693389, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.353564977645874, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7135011553764343, + "num_tokens": 332579569.0, + "step": 12853 + }, + { + "epoch": 1.4115967493960027, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0666606426239014, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7040327191352844, + "num_tokens": 332611878.0, + "step": 12854 + }, + { + "epoch": 1.4117065670986162, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2940948009490967, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7064619064331055, + "num_tokens": 332638058.0, + "step": 12855 + }, + { + "epoch": 1.41181638480123, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1753156185150146, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.6971303224563599, + "num_tokens": 332666000.0, + "step": 12856 + }, + { + "epoch": 1.4119262025038437, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0347886085510254, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7258076667785645, + "num_tokens": 332694149.0, + "step": 12857 + }, + { + "epoch": 1.4120360202064572, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2214248180389404, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6870010495185852, + "num_tokens": 332724188.0, + "step": 12858 + }, + { + "epoch": 1.412145837909071, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0930349826812744, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.698352038860321, + "num_tokens": 332752306.0, + "step": 12859 + }, + { + "epoch": 1.4122556556116845, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2834689617156982, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.706672191619873, + "num_tokens": 332780115.0, + "step": 12860 + }, + { + "epoch": 1.4123654733142983, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.591167688369751, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7152434587478638, + "num_tokens": 332800506.0, + "step": 12861 + }, + { + "epoch": 1.412475291016912, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.157651424407959, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7119227647781372, + "num_tokens": 332830867.0, + "step": 12862 + }, + { + "epoch": 1.4125851087195256, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.170538902282715, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6837958097457886, + "num_tokens": 332860462.0, + "step": 12863 + }, + { + "epoch": 1.4126949264221391, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4856417179107666, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7285006642341614, + "num_tokens": 332880760.0, + "step": 12864 + }, + { + "epoch": 1.4128047441247529, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0705785751342773, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.702283501625061, + "num_tokens": 332910473.0, + "step": 12865 + }, + { + "epoch": 1.4129145618273666, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.453903913497925, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6965922117233276, + "num_tokens": 332936802.0, + "step": 12866 + }, + { + "epoch": 1.4130243795299802, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.475250482559204, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7363491058349609, + "num_tokens": 332959432.0, + "step": 12867 + }, + { + "epoch": 1.413134197232594, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1021902561187744, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6959757208824158, + "num_tokens": 332987546.0, + "step": 12868 + }, + { + "epoch": 1.4132440149352075, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2580344676971436, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.6978647708892822, + "num_tokens": 333014776.0, + "step": 12869 + }, + { + "epoch": 1.4133538326378212, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.290351629257202, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7142529487609863, + "num_tokens": 333040984.0, + "step": 12870 + }, + { + "epoch": 1.413463650340435, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3315749168395996, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7273015379905701, + "num_tokens": 333064260.0, + "step": 12871 + }, + { + "epoch": 1.4135734680430485, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.561476469039917, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7271969318389893, + "num_tokens": 333085230.0, + "step": 12872 + }, + { + "epoch": 1.4136832857456623, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.210908889770508, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7047740817070007, + "num_tokens": 333111202.0, + "step": 12873 + }, + { + "epoch": 1.4137931034482758, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2236721515655518, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7108409404754639, + "num_tokens": 333135562.0, + "step": 12874 + }, + { + "epoch": 1.4139029211508896, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.229159116744995, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7154386639595032, + "num_tokens": 333162233.0, + "step": 12875 + }, + { + "epoch": 1.4140127388535033, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.271479368209839, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7186954617500305, + "num_tokens": 333187389.0, + "step": 12876 + }, + { + "epoch": 1.4141225565561168, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1964330673217773, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7194938659667969, + "num_tokens": 333216193.0, + "step": 12877 + }, + { + "epoch": 1.4142323742587304, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.495811939239502, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7275539636611938, + "num_tokens": 333237036.0, + "step": 12878 + }, + { + "epoch": 1.4143421919613441, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2274045944213867, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7275636196136475, + "num_tokens": 333261687.0, + "step": 12879 + }, + { + "epoch": 1.414452009663958, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.492086410522461, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7144285440444946, + "num_tokens": 333285475.0, + "step": 12880 + }, + { + "epoch": 1.4145618273665714, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.407226800918579, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.7083125114440918, + "num_tokens": 333308813.0, + "step": 12881 + }, + { + "epoch": 1.4146716450691852, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1719629764556885, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7150297164916992, + "num_tokens": 333335279.0, + "step": 12882 + }, + { + "epoch": 1.4147814627717987, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1048412322998047, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.709314227104187, + "num_tokens": 333364811.0, + "step": 12883 + }, + { + "epoch": 1.4148912804744125, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.6033220291137695, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7268245816230774, + "num_tokens": 333385365.0, + "step": 12884 + }, + { + "epoch": 1.4150010981770262, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5627596378326416, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7277078628540039, + "num_tokens": 333405477.0, + "step": 12885 + }, + { + "epoch": 1.4151109158796398, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3331775665283203, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7160495519638062, + "num_tokens": 333429113.0, + "step": 12886 + }, + { + "epoch": 1.4152207335822535, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3795015811920166, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7311584949493408, + "num_tokens": 333451628.0, + "step": 12887 + }, + { + "epoch": 1.415330551284867, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.6718568801879883, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.742912232875824, + "num_tokens": 333469094.0, + "step": 12888 + }, + { + "epoch": 1.4154403689874808, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2809388637542725, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7460846304893494, + "num_tokens": 333492280.0, + "step": 12889 + }, + { + "epoch": 1.4155501866900946, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.214531898498535, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7019813060760498, + "num_tokens": 333518127.0, + "step": 12890 + }, + { + "epoch": 1.4156600043927081, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.171664237976074, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6823990345001221, + "num_tokens": 333546384.0, + "step": 12891 + }, + { + "epoch": 1.4157698220953217, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3497824668884277, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7147511839866638, + "num_tokens": 333571273.0, + "step": 12892 + }, + { + "epoch": 1.4158796397979354, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3498053550720215, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7262861728668213, + "num_tokens": 333594588.0, + "step": 12893 + }, + { + "epoch": 1.4159894575005492, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0485336780548096, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7384065389633179, + "num_tokens": 333622666.0, + "step": 12894 + }, + { + "epoch": 1.4160992752031627, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3691670894622803, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7322722673416138, + "num_tokens": 333644976.0, + "step": 12895 + }, + { + "epoch": 1.4162090929057765, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.190262794494629, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.7129355669021606, + "num_tokens": 333673260.0, + "step": 12896 + }, + { + "epoch": 1.41631891060839, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.28051495552063, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6888874769210815, + "num_tokens": 333701124.0, + "step": 12897 + }, + { + "epoch": 1.4164287283110037, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1768136024475098, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7157502770423889, + "num_tokens": 333728031.0, + "step": 12898 + }, + { + "epoch": 1.4165385460136175, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2509779930114746, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7334134578704834, + "num_tokens": 333753218.0, + "step": 12899 + }, + { + "epoch": 1.416648363716231, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.266608953475952, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7181458473205566, + "num_tokens": 333780133.0, + "step": 12900 + }, + { + "epoch": 1.4167581814188448, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.300340175628662, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7077347040176392, + "num_tokens": 333805585.0, + "step": 12901 + }, + { + "epoch": 1.4168679991214583, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.180826425552368, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7071204781532288, + "num_tokens": 333832788.0, + "step": 12902 + }, + { + "epoch": 1.416977816824072, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0993545055389404, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7244641780853271, + "num_tokens": 333861074.0, + "step": 12903 + }, + { + "epoch": 1.4170876345266856, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5473973751068115, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7147969007492065, + "num_tokens": 333883421.0, + "step": 12904 + }, + { + "epoch": 1.4171974522292994, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.544509172439575, + "learning_rate": 1e-06, + "loss": 0.7511, + "mean_token_accuracy": 0.757186233997345, + "num_tokens": 333902876.0, + "step": 12905 + }, + { + "epoch": 1.417307269931913, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0759024620056152, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7151256799697876, + "num_tokens": 333932878.0, + "step": 12906 + }, + { + "epoch": 1.4174170876345267, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.036717176437378, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7132773399353027, + "num_tokens": 333964610.0, + "step": 12907 + }, + { + "epoch": 1.4175269053371404, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3341426849365234, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.705832302570343, + "num_tokens": 333990179.0, + "step": 12908 + }, + { + "epoch": 1.417636723039754, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.420118570327759, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7286814451217651, + "num_tokens": 334014098.0, + "step": 12909 + }, + { + "epoch": 1.4177465407423677, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5122318267822266, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7205445766448975, + "num_tokens": 334035745.0, + "step": 12910 + }, + { + "epoch": 1.4178563584449813, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0916037559509277, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7193946838378906, + "num_tokens": 334063259.0, + "step": 12911 + }, + { + "epoch": 1.417966176147595, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.035687208175659, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7158703804016113, + "num_tokens": 334092798.0, + "step": 12912 + }, + { + "epoch": 1.4180759938502088, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.055548906326294, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7248462438583374, + "num_tokens": 334121983.0, + "step": 12913 + }, + { + "epoch": 1.4181858115528223, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1829674243927, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.697799801826477, + "num_tokens": 334150366.0, + "step": 12914 + }, + { + "epoch": 1.4182956292554358, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.332075595855713, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7294496297836304, + "num_tokens": 334173217.0, + "step": 12915 + }, + { + "epoch": 1.4184054469580496, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0689618587493896, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6929877400398254, + "num_tokens": 334203261.0, + "step": 12916 + }, + { + "epoch": 1.4185152646606634, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.360835552215576, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7012941837310791, + "num_tokens": 334229166.0, + "step": 12917 + }, + { + "epoch": 1.418625082363277, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1192374229431152, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6926166415214539, + "num_tokens": 334258769.0, + "step": 12918 + }, + { + "epoch": 1.4187349000658906, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.088334083557129, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7128376364707947, + "num_tokens": 334287469.0, + "step": 12919 + }, + { + "epoch": 1.4188447177685042, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1747725009918213, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7163203954696655, + "num_tokens": 334314416.0, + "step": 12920 + }, + { + "epoch": 1.418954535471118, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.142078399658203, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7277147769927979, + "num_tokens": 334339914.0, + "step": 12921 + }, + { + "epoch": 1.4190643531737317, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3732991218566895, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7193083763122559, + "num_tokens": 334362130.0, + "step": 12922 + }, + { + "epoch": 1.4191741708763452, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.9030787944793701, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7040654420852661, + "num_tokens": 334394745.0, + "step": 12923 + }, + { + "epoch": 1.419283988578959, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.129798173904419, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7220233678817749, + "num_tokens": 334422706.0, + "step": 12924 + }, + { + "epoch": 1.4193938062815725, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2690436840057373, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.708871066570282, + "num_tokens": 334448475.0, + "step": 12925 + }, + { + "epoch": 1.4195036239841863, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.265967607498169, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7296009659767151, + "num_tokens": 334473179.0, + "step": 12926 + }, + { + "epoch": 1.4196134416868, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.162217617034912, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7177767753601074, + "num_tokens": 334499427.0, + "step": 12927 + }, + { + "epoch": 1.4197232593894136, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0877389907836914, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7034382820129395, + "num_tokens": 334531301.0, + "step": 12928 + }, + { + "epoch": 1.419833077092027, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.273186445236206, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7180020809173584, + "num_tokens": 334555507.0, + "step": 12929 + }, + { + "epoch": 1.4199428947946409, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.960556983947754, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7125605344772339, + "num_tokens": 334587652.0, + "step": 12930 + }, + { + "epoch": 1.4200527124972546, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.482579231262207, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7203278541564941, + "num_tokens": 334610022.0, + "step": 12931 + }, + { + "epoch": 1.4201625301998682, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0389411449432373, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7029922008514404, + "num_tokens": 334639176.0, + "step": 12932 + }, + { + "epoch": 1.420272347902482, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.5783019065856934, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7369250655174255, + "num_tokens": 334658933.0, + "step": 12933 + }, + { + "epoch": 1.4203821656050954, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2270145416259766, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7028262615203857, + "num_tokens": 334687502.0, + "step": 12934 + }, + { + "epoch": 1.4204919833077092, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.967417597770691, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.7057287693023682, + "num_tokens": 334719409.0, + "step": 12935 + }, + { + "epoch": 1.420601801010323, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 1.9912583827972412, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.705691397190094, + "num_tokens": 334749363.0, + "step": 12936 + }, + { + "epoch": 1.4207116187129365, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0779526233673096, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6912906765937805, + "num_tokens": 334780552.0, + "step": 12937 + }, + { + "epoch": 1.4208214364155503, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1968846321105957, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7012466192245483, + "num_tokens": 334808206.0, + "step": 12938 + }, + { + "epoch": 1.4209312541181638, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3566553592681885, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7175891399383545, + "num_tokens": 334831561.0, + "step": 12939 + }, + { + "epoch": 1.4210410718207775, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1784706115722656, + "learning_rate": 1e-06, + "loss": 1.0484, + "mean_token_accuracy": 0.6850284337997437, + "num_tokens": 334859188.0, + "step": 12940 + }, + { + "epoch": 1.4211508895233913, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2783493995666504, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7242459654808044, + "num_tokens": 334884809.0, + "step": 12941 + }, + { + "epoch": 1.4212607072260048, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.217912435531616, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7138098478317261, + "num_tokens": 334910764.0, + "step": 12942 + }, + { + "epoch": 1.4213705249286184, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4597065448760986, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7207082509994507, + "num_tokens": 334935069.0, + "step": 12943 + }, + { + "epoch": 1.4214803426312321, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2091267108917236, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7139320373535156, + "num_tokens": 334961762.0, + "step": 12944 + }, + { + "epoch": 1.4215901603338459, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2863802909851074, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7232239246368408, + "num_tokens": 334988095.0, + "step": 12945 + }, + { + "epoch": 1.4216999780364594, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.1292383670806885, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7298116683959961, + "num_tokens": 335015639.0, + "step": 12946 + }, + { + "epoch": 1.4218097957390732, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4966325759887695, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.726529598236084, + "num_tokens": 335038331.0, + "step": 12947 + }, + { + "epoch": 1.4219196134416867, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3562424182891846, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6983758211135864, + "num_tokens": 335064612.0, + "step": 12948 + }, + { + "epoch": 1.4220294311443005, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.322268486022949, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.723909854888916, + "num_tokens": 335088936.0, + "step": 12949 + }, + { + "epoch": 1.4221392488469142, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.325793981552124, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7379484176635742, + "num_tokens": 335111531.0, + "step": 12950 + }, + { + "epoch": 1.4222490665495278, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2046616077423096, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7167307734489441, + "num_tokens": 335138758.0, + "step": 12951 + }, + { + "epoch": 1.4223588842521415, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2660045623779297, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7225108742713928, + "num_tokens": 335165080.0, + "step": 12952 + }, + { + "epoch": 1.422468701954755, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4170117378234863, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7038974165916443, + "num_tokens": 335189641.0, + "step": 12953 + }, + { + "epoch": 1.4225785196573688, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.051243305206299, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7066588401794434, + "num_tokens": 335219361.0, + "step": 12954 + }, + { + "epoch": 1.4226883373599823, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.3410074710845947, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.71671462059021, + "num_tokens": 335242980.0, + "step": 12955 + }, + { + "epoch": 1.422798155062596, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4735910892486572, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7036833167076111, + "num_tokens": 335266402.0, + "step": 12956 + }, + { + "epoch": 1.4229079727652096, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4436869621276855, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7285727262496948, + "num_tokens": 335287588.0, + "step": 12957 + }, + { + "epoch": 1.4230177904678234, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.275162935256958, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7274423241615295, + "num_tokens": 335312230.0, + "step": 12958 + }, + { + "epoch": 1.4231276081704372, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0452356338500977, + "learning_rate": 1e-06, + "loss": 0.7735, + "mean_token_accuracy": 0.7505232095718384, + "num_tokens": 335339674.0, + "step": 12959 + }, + { + "epoch": 1.4232374258730507, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.301536798477173, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7146519422531128, + "num_tokens": 335366323.0, + "step": 12960 + }, + { + "epoch": 1.4233472435756644, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.419140577316284, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7362087965011597, + "num_tokens": 335388086.0, + "step": 12961 + }, + { + "epoch": 1.423457061278278, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.056182622909546, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7321764230728149, + "num_tokens": 335416717.0, + "step": 12962 + }, + { + "epoch": 1.4235668789808917, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0840179920196533, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7106077671051025, + "num_tokens": 335448787.0, + "step": 12963 + }, + { + "epoch": 1.4236766966835055, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.261388063430786, + "learning_rate": 1e-06, + "loss": 0.7534, + "mean_token_accuracy": 0.7592905759811401, + "num_tokens": 335471561.0, + "step": 12964 + }, + { + "epoch": 1.423786514386119, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2298872470855713, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7188389301300049, + "num_tokens": 335498861.0, + "step": 12965 + }, + { + "epoch": 1.4238963320887328, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.725902557373047, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7275570631027222, + "num_tokens": 335518377.0, + "step": 12966 + }, + { + "epoch": 1.4240061497913463, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.8545584678649902, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7329787015914917, + "num_tokens": 335539438.0, + "step": 12967 + }, + { + "epoch": 1.42411596749396, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.2007994651794434, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7575571537017822, + "num_tokens": 335563992.0, + "step": 12968 + }, + { + "epoch": 1.4242257851965736, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.113821029663086, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6971309781074524, + "num_tokens": 335595342.0, + "step": 12969 + }, + { + "epoch": 1.4243356028991874, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.0993194580078125, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.6992602348327637, + "num_tokens": 335623167.0, + "step": 12970 + }, + { + "epoch": 1.424445420601801, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.470531940460205, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7347393035888672, + "num_tokens": 335645301.0, + "step": 12971 + }, + { + "epoch": 1.4245552383044147, + "ewc_loss": 0.00011730194091796875, + "grad_norm": 2.4827582836151123, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7314730286598206, + "num_tokens": 335668265.0, + "step": 12972 + }, + { + "epoch": 1.4246650560070284, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 1.9683200120925903, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6837464570999146, + "num_tokens": 335702380.0, + "step": 12973 + }, + { + "epoch": 1.424774873709642, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.3089888095855713, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7090736627578735, + "num_tokens": 335728746.0, + "step": 12974 + }, + { + "epoch": 1.4248846914122557, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.049264430999756, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7139600515365601, + "num_tokens": 335760911.0, + "step": 12975 + }, + { + "epoch": 1.4249945091148692, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2185845375061035, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7329832911491394, + "num_tokens": 335787964.0, + "step": 12976 + }, + { + "epoch": 1.425104326817483, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.5166523456573486, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7208303809165955, + "num_tokens": 335808830.0, + "step": 12977 + }, + { + "epoch": 1.4252141445200968, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.0246779918670654, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7064228057861328, + "num_tokens": 335838235.0, + "step": 12978 + }, + { + "epoch": 1.4253239622227103, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.333087205886841, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7187781929969788, + "num_tokens": 335861638.0, + "step": 12979 + }, + { + "epoch": 1.4254337799253238, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2643978595733643, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7095088362693787, + "num_tokens": 335887560.0, + "step": 12980 + }, + { + "epoch": 1.4255435976279376, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.0454728603363037, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.710893988609314, + "num_tokens": 335916147.0, + "step": 12981 + }, + { + "epoch": 1.4256534153305513, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.08625864982605, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6953766345977783, + "num_tokens": 335945961.0, + "step": 12982 + }, + { + "epoch": 1.4257632330331649, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.398512601852417, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7256899476051331, + "num_tokens": 335968366.0, + "step": 12983 + }, + { + "epoch": 1.4258730507357786, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.4266457557678223, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.737604022026062, + "num_tokens": 335989642.0, + "step": 12984 + }, + { + "epoch": 1.4259828684383922, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.176307439804077, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7113440036773682, + "num_tokens": 336017457.0, + "step": 12985 + }, + { + "epoch": 1.426092686141006, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.223642587661743, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7169293165206909, + "num_tokens": 336043602.0, + "step": 12986 + }, + { + "epoch": 1.4262025038436197, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.185441255569458, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7253406047821045, + "num_tokens": 336070210.0, + "step": 12987 + }, + { + "epoch": 1.4263123215462332, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.198219060897827, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7223296165466309, + "num_tokens": 336097055.0, + "step": 12988 + }, + { + "epoch": 1.426422139248847, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2275795936584473, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7093973159790039, + "num_tokens": 336124255.0, + "step": 12989 + }, + { + "epoch": 1.4265319569514605, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.3370168209075928, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7150915265083313, + "num_tokens": 336151343.0, + "step": 12990 + }, + { + "epoch": 1.4266417746540743, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 1.9952775239944458, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.716172456741333, + "num_tokens": 336183647.0, + "step": 12991 + }, + { + "epoch": 1.426751592356688, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.4985499382019043, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7274436950683594, + "num_tokens": 336204677.0, + "step": 12992 + }, + { + "epoch": 1.4268614100593016, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.103940963745117, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7092383503913879, + "num_tokens": 336232008.0, + "step": 12993 + }, + { + "epoch": 1.426971227761915, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 1.9946526288986206, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6979771256446838, + "num_tokens": 336265650.0, + "step": 12994 + }, + { + "epoch": 1.4270810454645289, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.1825292110443115, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.6989625692367554, + "num_tokens": 336293078.0, + "step": 12995 + }, + { + "epoch": 1.4271908631671426, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.0134615898132324, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7004297971725464, + "num_tokens": 336325403.0, + "step": 12996 + }, + { + "epoch": 1.4273006808697561, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.321484088897705, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7142605781555176, + "num_tokens": 336350773.0, + "step": 12997 + }, + { + "epoch": 1.42741049857237, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.243018388748169, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7229213714599609, + "num_tokens": 336375432.0, + "step": 12998 + }, + { + "epoch": 1.4275203162749834, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.1700384616851807, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7234138250350952, + "num_tokens": 336402467.0, + "step": 12999 + }, + { + "epoch": 1.4276301339775972, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.263070821762085, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7233200073242188, + "num_tokens": 336426745.0, + "step": 13000 + }, + { + "epoch": 1.427739951680211, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.9476737976074219, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7173457145690918, + "num_tokens": 336456691.0, + "step": 13001 + }, + { + "epoch": 1.4278497693828245, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1402032375335693, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7373807430267334, + "num_tokens": 336481532.0, + "step": 13002 + }, + { + "epoch": 1.4279595870854382, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3724584579467773, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7467643022537231, + "num_tokens": 336502282.0, + "step": 13003 + }, + { + "epoch": 1.4280694047880518, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.9408236742019653, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7031123638153076, + "num_tokens": 336536836.0, + "step": 13004 + }, + { + "epoch": 1.4281792224906655, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3557231426239014, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.731442928314209, + "num_tokens": 336559622.0, + "step": 13005 + }, + { + "epoch": 1.4282890401932793, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1465420722961426, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7303990125656128, + "num_tokens": 336586090.0, + "step": 13006 + }, + { + "epoch": 1.4283988578958928, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.376335382461548, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7258085012435913, + "num_tokens": 336610788.0, + "step": 13007 + }, + { + "epoch": 1.4285086755985064, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.1180827617645264, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7255206108093262, + "num_tokens": 336637114.0, + "step": 13008 + }, + { + "epoch": 1.4286184933011201, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2890868186950684, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7161446213722229, + "num_tokens": 336662254.0, + "step": 13009 + }, + { + "epoch": 1.4287283110037339, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.0296194553375244, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7033056616783142, + "num_tokens": 336692324.0, + "step": 13010 + }, + { + "epoch": 1.4288381287063474, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2967336177825928, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7067718505859375, + "num_tokens": 336719362.0, + "step": 13011 + }, + { + "epoch": 1.4289479464089612, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.588441848754883, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7204094529151917, + "num_tokens": 336739122.0, + "step": 13012 + }, + { + "epoch": 1.4290577641115747, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.024883270263672, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6968172788619995, + "num_tokens": 336772121.0, + "step": 13013 + }, + { + "epoch": 1.4291675818141885, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.3666138648986816, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7054559588432312, + "num_tokens": 336796925.0, + "step": 13014 + }, + { + "epoch": 1.4292773995168022, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.348428726196289, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7105693817138672, + "num_tokens": 336821568.0, + "step": 13015 + }, + { + "epoch": 1.4293872172194158, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2927753925323486, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7148251533508301, + "num_tokens": 336847377.0, + "step": 13016 + }, + { + "epoch": 1.4294970349220295, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.2153501510620117, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6921355128288269, + "num_tokens": 336872968.0, + "step": 13017 + }, + { + "epoch": 1.429606852624643, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.326329469680786, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.704551637172699, + "num_tokens": 336897484.0, + "step": 13018 + }, + { + "epoch": 1.4297166703272568, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.0261309146881104, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7374125123023987, + "num_tokens": 336925818.0, + "step": 13019 + }, + { + "epoch": 1.4298264880298703, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 3.068084716796875, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7318854331970215, + "num_tokens": 336949216.0, + "step": 13020 + }, + { + "epoch": 1.429936305732484, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 2.526155710220337, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7082030773162842, + "num_tokens": 336970982.0, + "step": 13021 + }, + { + "epoch": 1.4300461234350976, + "ewc_loss": 0.00011777877807617188, + "grad_norm": 1.9378777742385864, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.7045977115631104, + "num_tokens": 337003995.0, + "step": 13022 + }, + { + "epoch": 1.4301559411377114, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.8309144973754883, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7435084581375122, + "num_tokens": 337022070.0, + "step": 13023 + }, + { + "epoch": 1.4302657588403251, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2855498790740967, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7231640815734863, + "num_tokens": 337044754.0, + "step": 13024 + }, + { + "epoch": 1.4303755765429387, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2197065353393555, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7014578580856323, + "num_tokens": 337071397.0, + "step": 13025 + }, + { + "epoch": 1.4304853942455524, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2148311138153076, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.7005125284194946, + "num_tokens": 337097440.0, + "step": 13026 + }, + { + "epoch": 1.430595211948166, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3971927165985107, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7165609002113342, + "num_tokens": 337122037.0, + "step": 13027 + }, + { + "epoch": 1.4307050296507797, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.236703395843506, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7175470590591431, + "num_tokens": 337147695.0, + "step": 13028 + }, + { + "epoch": 1.4308148473533935, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2976763248443604, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7242720127105713, + "num_tokens": 337174018.0, + "step": 13029 + }, + { + "epoch": 1.430924665056007, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2064812183380127, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7119067311286926, + "num_tokens": 337201036.0, + "step": 13030 + }, + { + "epoch": 1.4310344827586206, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.189894199371338, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7232621312141418, + "num_tokens": 337227358.0, + "step": 13031 + }, + { + "epoch": 1.4311443004612343, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3113534450531006, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7144598960876465, + "num_tokens": 337251693.0, + "step": 13032 + }, + { + "epoch": 1.431254118163848, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1833274364471436, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7244096398353577, + "num_tokens": 337278719.0, + "step": 13033 + }, + { + "epoch": 1.4313639358664616, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.0802001953125, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6945048570632935, + "num_tokens": 337310203.0, + "step": 13034 + }, + { + "epoch": 1.4314737535690754, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.463336229324341, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7177630662918091, + "num_tokens": 337332944.0, + "step": 13035 + }, + { + "epoch": 1.431583571271689, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.5291504859924316, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7176205515861511, + "num_tokens": 337353080.0, + "step": 13036 + }, + { + "epoch": 1.4316933889743026, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.9118287563323975, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6849692463874817, + "num_tokens": 337388005.0, + "step": 13037 + }, + { + "epoch": 1.4318032066769164, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.332810163497925, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7151097059249878, + "num_tokens": 337412486.0, + "step": 13038 + }, + { + "epoch": 1.43191302437953, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.0913243293762207, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7044309377670288, + "num_tokens": 337442562.0, + "step": 13039 + }, + { + "epoch": 1.4320228420821437, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.6191141605377197, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7375977039337158, + "num_tokens": 337462022.0, + "step": 13040 + }, + { + "epoch": 1.4321326597847572, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1479876041412354, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6857566237449646, + "num_tokens": 337489573.0, + "step": 13041 + }, + { + "epoch": 1.432242477487371, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.223351240158081, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7060738205909729, + "num_tokens": 337517170.0, + "step": 13042 + }, + { + "epoch": 1.4323522951899847, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.9337142705917358, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7121396660804749, + "num_tokens": 337550404.0, + "step": 13043 + }, + { + "epoch": 1.4324621128925983, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2858893871307373, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7244398593902588, + "num_tokens": 337573747.0, + "step": 13044 + }, + { + "epoch": 1.4325719305952118, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.4119515419006348, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7290803790092468, + "num_tokens": 337596783.0, + "step": 13045 + }, + { + "epoch": 1.4326817482978256, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.29744815826416, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7239292860031128, + "num_tokens": 337619950.0, + "step": 13046 + }, + { + "epoch": 1.4327915660004393, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.9582105875015259, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7358156442642212, + "num_tokens": 337650293.0, + "step": 13047 + }, + { + "epoch": 1.4329013837030529, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.199775457382202, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.717136561870575, + "num_tokens": 337680476.0, + "step": 13048 + }, + { + "epoch": 1.4330112014056666, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.4044055938720703, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7197403311729431, + "num_tokens": 337702464.0, + "step": 13049 + }, + { + "epoch": 1.4331210191082802, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1801912784576416, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7172521352767944, + "num_tokens": 337729715.0, + "step": 13050 + }, + { + "epoch": 1.433230836810894, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.109980583190918, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7018718719482422, + "num_tokens": 337757375.0, + "step": 13051 + }, + { + "epoch": 1.4333406545135077, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.196654796600342, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7287614941596985, + "num_tokens": 337784525.0, + "step": 13052 + }, + { + "epoch": 1.4334504722161212, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.5674753189086914, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7311957478523254, + "num_tokens": 337807150.0, + "step": 13053 + }, + { + "epoch": 1.433560289918735, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.464114189147949, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7087900638580322, + "num_tokens": 337830400.0, + "step": 13054 + }, + { + "epoch": 1.4336701076213485, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2093570232391357, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7275058031082153, + "num_tokens": 337854411.0, + "step": 13055 + }, + { + "epoch": 1.4337799253239623, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1406962871551514, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7301695942878723, + "num_tokens": 337880853.0, + "step": 13056 + }, + { + "epoch": 1.433889743026576, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.2173514366149902, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7224243879318237, + "num_tokens": 337906239.0, + "step": 13057 + }, + { + "epoch": 1.4339995607291895, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3611936569213867, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7478482723236084, + "num_tokens": 337928118.0, + "step": 13058 + }, + { + "epoch": 1.434109378431803, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1716909408569336, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7090701460838318, + "num_tokens": 337957483.0, + "step": 13059 + }, + { + "epoch": 1.4342191961344168, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.269667387008667, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.722771406173706, + "num_tokens": 337986256.0, + "step": 13060 + }, + { + "epoch": 1.4343290138370306, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.069952964782715, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7289886474609375, + "num_tokens": 338016447.0, + "step": 13061 + }, + { + "epoch": 1.4344388315396441, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3969969749450684, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.694894552230835, + "num_tokens": 338041049.0, + "step": 13062 + }, + { + "epoch": 1.434548649242258, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.0552141666412354, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.73973548412323, + "num_tokens": 338069754.0, + "step": 13063 + }, + { + "epoch": 1.4346584669448714, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.4430718421936035, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7298157811164856, + "num_tokens": 338090336.0, + "step": 13064 + }, + { + "epoch": 1.4347682846474852, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.0912342071533203, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7296463251113892, + "num_tokens": 338118398.0, + "step": 13065 + }, + { + "epoch": 1.434878102350099, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1501731872558594, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7310196161270142, + "num_tokens": 338144635.0, + "step": 13066 + }, + { + "epoch": 1.4349879200527125, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1628994941711426, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7051945924758911, + "num_tokens": 338172175.0, + "step": 13067 + }, + { + "epoch": 1.4350977377553262, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.13020658493042, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7143787741661072, + "num_tokens": 338200321.0, + "step": 13068 + }, + { + "epoch": 1.4352075554579398, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3984317779541016, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7247928380966187, + "num_tokens": 338224147.0, + "step": 13069 + }, + { + "epoch": 1.4353173731605535, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.0024664402008057, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.726923942565918, + "num_tokens": 338253884.0, + "step": 13070 + }, + { + "epoch": 1.4354271908631673, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1356184482574463, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7116438150405884, + "num_tokens": 338281900.0, + "step": 13071 + }, + { + "epoch": 1.4355370085657808, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.4081358909606934, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7200152277946472, + "num_tokens": 338305899.0, + "step": 13072 + }, + { + "epoch": 1.4356468262683943, + "ewc_loss": 0.000118255615234375, + "grad_norm": 1.771561861038208, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7070227265357971, + "num_tokens": 338344227.0, + "step": 13073 + }, + { + "epoch": 1.435756643971008, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.378084659576416, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7129542231559753, + "num_tokens": 338369871.0, + "step": 13074 + }, + { + "epoch": 1.4358664616736219, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.409188747406006, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7138875126838684, + "num_tokens": 338396378.0, + "step": 13075 + }, + { + "epoch": 1.4359762793762354, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3838584423065186, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7168781757354736, + "num_tokens": 338418917.0, + "step": 13076 + }, + { + "epoch": 1.4360860970788492, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.355666399002075, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.721297562122345, + "num_tokens": 338443366.0, + "step": 13077 + }, + { + "epoch": 1.4361959147814627, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.438619375228882, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7437085509300232, + "num_tokens": 338464714.0, + "step": 13078 + }, + { + "epoch": 1.4363057324840764, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1284940242767334, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7068756818771362, + "num_tokens": 338496171.0, + "step": 13079 + }, + { + "epoch": 1.4364155501866902, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.179616689682007, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7395377159118652, + "num_tokens": 338521962.0, + "step": 13080 + }, + { + "epoch": 1.4365253678893037, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3670272827148438, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.717570960521698, + "num_tokens": 338547278.0, + "step": 13081 + }, + { + "epoch": 1.4366351855919175, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.263944387435913, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7171002626419067, + "num_tokens": 338574124.0, + "step": 13082 + }, + { + "epoch": 1.436745003294531, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3917980194091797, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7264758348464966, + "num_tokens": 338598037.0, + "step": 13083 + }, + { + "epoch": 1.4368548209971448, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.071215867996216, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.7027597427368164, + "num_tokens": 338629611.0, + "step": 13084 + }, + { + "epoch": 1.4369646386997583, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.216053009033203, + "learning_rate": 1e-06, + "loss": 0.7754, + "mean_token_accuracy": 0.7576034069061279, + "num_tokens": 338652735.0, + "step": 13085 + }, + { + "epoch": 1.437074456402372, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.492976427078247, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7264969944953918, + "num_tokens": 338673571.0, + "step": 13086 + }, + { + "epoch": 1.4371842741049856, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.194197654724121, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.707611083984375, + "num_tokens": 338700594.0, + "step": 13087 + }, + { + "epoch": 1.4372940918075994, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.5684523582458496, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7298687696456909, + "num_tokens": 338721880.0, + "step": 13088 + }, + { + "epoch": 1.4374039095102131, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.3633322715759277, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7348304986953735, + "num_tokens": 338745508.0, + "step": 13089 + }, + { + "epoch": 1.4375137272128267, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.4163942337036133, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7209391593933105, + "num_tokens": 338769394.0, + "step": 13090 + }, + { + "epoch": 1.4376235449154404, + "ewc_loss": 0.000118255615234375, + "grad_norm": 2.1162571907043457, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7132197618484497, + "num_tokens": 338799700.0, + "step": 13091 + }, + { + "epoch": 1.437733362618054, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4309804439544678, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7156896591186523, + "num_tokens": 338823255.0, + "step": 13092 + }, + { + "epoch": 1.4378431803206677, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9567079544067383, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7110649347305298, + "num_tokens": 338855495.0, + "step": 13093 + }, + { + "epoch": 1.4379529980232815, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.5668256282806396, + "learning_rate": 1e-06, + "loss": 0.7262, + "mean_token_accuracy": 0.7666343450546265, + "num_tokens": 338873672.0, + "step": 13094 + }, + { + "epoch": 1.438062815725895, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1605145931243896, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6984755396842957, + "num_tokens": 338900914.0, + "step": 13095 + }, + { + "epoch": 1.4381726334285085, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3307156562805176, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7141789197921753, + "num_tokens": 338926093.0, + "step": 13096 + }, + { + "epoch": 1.4382824511311223, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.8914439678192139, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7040271759033203, + "num_tokens": 338959557.0, + "step": 13097 + }, + { + "epoch": 1.438392268833736, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9597519636154175, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.693581223487854, + "num_tokens": 338995283.0, + "step": 13098 + }, + { + "epoch": 1.4385020865363496, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.408756971359253, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.703677773475647, + "num_tokens": 339017549.0, + "step": 13099 + }, + { + "epoch": 1.4386119042389633, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.0606768131256104, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7146600484848022, + "num_tokens": 339045339.0, + "step": 13100 + }, + { + "epoch": 1.4387217219415769, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.0491106510162354, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7108205556869507, + "num_tokens": 339076105.0, + "step": 13101 + }, + { + "epoch": 1.4388315396441906, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.225447177886963, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.7000987529754639, + "num_tokens": 339101917.0, + "step": 13102 + }, + { + "epoch": 1.4389413573468044, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3750228881835938, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7244359254837036, + "num_tokens": 339125277.0, + "step": 13103 + }, + { + "epoch": 1.439051175049418, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.6188766956329346, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7395318746566772, + "num_tokens": 339142988.0, + "step": 13104 + }, + { + "epoch": 1.4391609927520317, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.50079083442688, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7110891342163086, + "num_tokens": 339164220.0, + "step": 13105 + }, + { + "epoch": 1.4392708104546452, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.098111867904663, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7073794603347778, + "num_tokens": 339194424.0, + "step": 13106 + }, + { + "epoch": 1.439380628157259, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.297778367996216, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7057583332061768, + "num_tokens": 339219591.0, + "step": 13107 + }, + { + "epoch": 1.4394904458598727, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.050294876098633, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7077608108520508, + "num_tokens": 339250645.0, + "step": 13108 + }, + { + "epoch": 1.4396002635624863, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2005538940429688, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.708749532699585, + "num_tokens": 339276743.0, + "step": 13109 + }, + { + "epoch": 1.4397100812650998, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.43202543258667, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7333357334136963, + "num_tokens": 339297691.0, + "step": 13110 + }, + { + "epoch": 1.4398198989677136, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3737895488739014, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7206705808639526, + "num_tokens": 339320684.0, + "step": 13111 + }, + { + "epoch": 1.4399297166703273, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3269481658935547, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7227616906166077, + "num_tokens": 339345478.0, + "step": 13112 + }, + { + "epoch": 1.4400395343729409, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.466884136199951, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7217945456504822, + "num_tokens": 339368782.0, + "step": 13113 + }, + { + "epoch": 1.4401493520755546, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2564754486083984, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7040143013000488, + "num_tokens": 339396266.0, + "step": 13114 + }, + { + "epoch": 1.4402591697781681, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1123428344726562, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7288779020309448, + "num_tokens": 339425471.0, + "step": 13115 + }, + { + "epoch": 1.440368987480782, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.310878038406372, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7406473159790039, + "num_tokens": 339448977.0, + "step": 13116 + }, + { + "epoch": 1.4404788051833957, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.338869333267212, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.728751540184021, + "num_tokens": 339471164.0, + "step": 13117 + }, + { + "epoch": 1.4405886228860092, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.131826877593994, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7196069359779358, + "num_tokens": 339497571.0, + "step": 13118 + }, + { + "epoch": 1.440698440588623, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.234452247619629, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7249889373779297, + "num_tokens": 339522326.0, + "step": 13119 + }, + { + "epoch": 1.4408082582912365, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3295302391052246, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7120999097824097, + "num_tokens": 339547243.0, + "step": 13120 + }, + { + "epoch": 1.4409180759938502, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.350001335144043, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7161424160003662, + "num_tokens": 339570558.0, + "step": 13121 + }, + { + "epoch": 1.441027893696464, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4083402156829834, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7387224435806274, + "num_tokens": 339592445.0, + "step": 13122 + }, + { + "epoch": 1.4411377113990775, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.294313669204712, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7092801928520203, + "num_tokens": 339617239.0, + "step": 13123 + }, + { + "epoch": 1.441247529101691, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4817848205566406, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.715571939945221, + "num_tokens": 339639633.0, + "step": 13124 + }, + { + "epoch": 1.4413573468043048, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3238484859466553, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7247269153594971, + "num_tokens": 339663745.0, + "step": 13125 + }, + { + "epoch": 1.4414671645069186, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4987921714782715, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7267575860023499, + "num_tokens": 339683723.0, + "step": 13126 + }, + { + "epoch": 1.4415769822095321, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.6957056522369385, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7441784143447876, + "num_tokens": 339702235.0, + "step": 13127 + }, + { + "epoch": 1.4416867999121459, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1831936836242676, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7429114580154419, + "num_tokens": 339727218.0, + "step": 13128 + }, + { + "epoch": 1.4417966176147594, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1656112670898438, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7245601415634155, + "num_tokens": 339755696.0, + "step": 13129 + }, + { + "epoch": 1.4419064353173732, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1673240661621094, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7131901979446411, + "num_tokens": 339781899.0, + "step": 13130 + }, + { + "epoch": 1.442016253019987, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3324532508850098, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7395941019058228, + "num_tokens": 339805352.0, + "step": 13131 + }, + { + "epoch": 1.4421260707226005, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.266935110092163, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7365959286689758, + "num_tokens": 339829741.0, + "step": 13132 + }, + { + "epoch": 1.4422358884252142, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.431185483932495, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7284878492355347, + "num_tokens": 339852460.0, + "step": 13133 + }, + { + "epoch": 1.4423457061278278, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3371779918670654, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7080823183059692, + "num_tokens": 339877706.0, + "step": 13134 + }, + { + "epoch": 1.4424555238304415, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.281252384185791, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7350319027900696, + "num_tokens": 339901952.0, + "step": 13135 + }, + { + "epoch": 1.442565341533055, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.498718023300171, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7249631881713867, + "num_tokens": 339924641.0, + "step": 13136 + }, + { + "epoch": 1.4426751592356688, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9390244483947754, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7205061316490173, + "num_tokens": 339955650.0, + "step": 13137 + }, + { + "epoch": 1.4427849769382823, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1808958053588867, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.714753270149231, + "num_tokens": 339982083.0, + "step": 13138 + }, + { + "epoch": 1.442894794640896, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3969428539276123, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6908474564552307, + "num_tokens": 340003890.0, + "step": 13139 + }, + { + "epoch": 1.4430046123435099, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.299828290939331, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7149004936218262, + "num_tokens": 340028979.0, + "step": 13140 + }, + { + "epoch": 1.4431144300461234, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4044435024261475, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7213346362113953, + "num_tokens": 340051821.0, + "step": 13141 + }, + { + "epoch": 1.4432242477487371, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2375881671905518, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7284561395645142, + "num_tokens": 340077118.0, + "step": 13142 + }, + { + "epoch": 1.4433340654513507, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.180271863937378, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7082889080047607, + "num_tokens": 340103599.0, + "step": 13143 + }, + { + "epoch": 1.4434438831539644, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.431725025177002, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7389022707939148, + "num_tokens": 340125652.0, + "step": 13144 + }, + { + "epoch": 1.4435537008565782, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2945642471313477, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7283162474632263, + "num_tokens": 340150073.0, + "step": 13145 + }, + { + "epoch": 1.4436635185591917, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.403174638748169, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7365503311157227, + "num_tokens": 340172018.0, + "step": 13146 + }, + { + "epoch": 1.4437733362618055, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2427802085876465, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7102755904197693, + "num_tokens": 340197922.0, + "step": 13147 + }, + { + "epoch": 1.443883153964419, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9447230100631714, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6879067420959473, + "num_tokens": 340232043.0, + "step": 13148 + }, + { + "epoch": 1.4439929716670328, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9751675128936768, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7072169184684753, + "num_tokens": 340263950.0, + "step": 13149 + }, + { + "epoch": 1.4441027893696463, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.266618490219116, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7203380465507507, + "num_tokens": 340290039.0, + "step": 13150 + }, + { + "epoch": 1.44421260707226, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.4560546875, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7057867050170898, + "num_tokens": 340312673.0, + "step": 13151 + }, + { + "epoch": 1.4443224247748736, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.21530818939209, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7351539731025696, + "num_tokens": 340339078.0, + "step": 13152 + }, + { + "epoch": 1.4444322424774874, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2759101390838623, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7146403789520264, + "num_tokens": 340364148.0, + "step": 13153 + }, + { + "epoch": 1.4445420601801011, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.227426290512085, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7275208830833435, + "num_tokens": 340390357.0, + "step": 13154 + }, + { + "epoch": 1.4446518778827147, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.347628116607666, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7289111018180847, + "num_tokens": 340414598.0, + "step": 13155 + }, + { + "epoch": 1.4447616955853284, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3935539722442627, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7136129140853882, + "num_tokens": 340436416.0, + "step": 13156 + }, + { + "epoch": 1.444871513287942, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.0491011142730713, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7125294804573059, + "num_tokens": 340468002.0, + "step": 13157 + }, + { + "epoch": 1.4449813309905557, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.6406614780426025, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7118628025054932, + "num_tokens": 340490276.0, + "step": 13158 + }, + { + "epoch": 1.4450911486931695, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.533806562423706, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7295329570770264, + "num_tokens": 340511926.0, + "step": 13159 + }, + { + "epoch": 1.445200966395783, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2099690437316895, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7276895642280579, + "num_tokens": 340537993.0, + "step": 13160 + }, + { + "epoch": 1.4453107840983965, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2930173873901367, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7111347913742065, + "num_tokens": 340562051.0, + "step": 13161 + }, + { + "epoch": 1.4454206018010103, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.381098508834839, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.709021270275116, + "num_tokens": 340584541.0, + "step": 13162 + }, + { + "epoch": 1.445530419503624, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.464576005935669, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.730252742767334, + "num_tokens": 340607199.0, + "step": 13163 + }, + { + "epoch": 1.4456402372062376, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3539388179779053, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7158119678497314, + "num_tokens": 340631608.0, + "step": 13164 + }, + { + "epoch": 1.4457500549088513, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.12160325050354, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6994224786758423, + "num_tokens": 340659819.0, + "step": 13165 + }, + { + "epoch": 1.4458598726114649, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.6515188217163086, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7374984622001648, + "num_tokens": 340678889.0, + "step": 13166 + }, + { + "epoch": 1.4459696903140786, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.384495973587036, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.719490647315979, + "num_tokens": 340704348.0, + "step": 13167 + }, + { + "epoch": 1.4460795080166924, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2402782440185547, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7280951142311096, + "num_tokens": 340728379.0, + "step": 13168 + }, + { + "epoch": 1.446189325719306, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.290910243988037, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7053557634353638, + "num_tokens": 340753423.0, + "step": 13169 + }, + { + "epoch": 1.4462991434219197, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1612212657928467, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7087250351905823, + "num_tokens": 340779801.0, + "step": 13170 + }, + { + "epoch": 1.4464089611245332, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.386638641357422, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7065520882606506, + "num_tokens": 340805614.0, + "step": 13171 + }, + { + "epoch": 1.446518778827147, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.8931140899658203, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7121609449386597, + "num_tokens": 340840393.0, + "step": 13172 + }, + { + "epoch": 1.4466285965297607, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.049471855163574, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7044387459754944, + "num_tokens": 340871054.0, + "step": 13173 + }, + { + "epoch": 1.4467384142323743, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2092795372009277, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6916527152061462, + "num_tokens": 340898566.0, + "step": 13174 + }, + { + "epoch": 1.4468482319349878, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.290642738342285, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7432571053504944, + "num_tokens": 340921348.0, + "step": 13175 + }, + { + "epoch": 1.4469580496376016, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.18626070022583, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7180392742156982, + "num_tokens": 340948757.0, + "step": 13176 + }, + { + "epoch": 1.4470678673402153, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.360485792160034, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7113937139511108, + "num_tokens": 340975774.0, + "step": 13177 + }, + { + "epoch": 1.4471776850428288, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.516186475753784, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7401311993598938, + "num_tokens": 340996031.0, + "step": 13178 + }, + { + "epoch": 1.4472875027454426, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.4383695125579834, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7136357426643372, + "num_tokens": 341020197.0, + "step": 13179 + }, + { + "epoch": 1.4473973204480561, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.194272518157959, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7217806577682495, + "num_tokens": 341047284.0, + "step": 13180 + }, + { + "epoch": 1.44750713815067, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.339348077774048, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7066627740859985, + "num_tokens": 341074120.0, + "step": 13181 + }, + { + "epoch": 1.4476169558532836, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.0559535026550293, + "learning_rate": 1e-06, + "loss": 1.033, + "mean_token_accuracy": 0.6991010904312134, + "num_tokens": 341106844.0, + "step": 13182 + }, + { + "epoch": 1.4477267735558972, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.1451878547668457, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.6910774111747742, + "num_tokens": 341133552.0, + "step": 13183 + }, + { + "epoch": 1.447836591258511, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.187532901763916, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7413424849510193, + "num_tokens": 341159404.0, + "step": 13184 + }, + { + "epoch": 1.4479464089611245, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.0802536010742188, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7136479020118713, + "num_tokens": 341188828.0, + "step": 13185 + }, + { + "epoch": 1.4480562266637382, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.2618401050567627, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7304539084434509, + "num_tokens": 341212905.0, + "step": 13186 + }, + { + "epoch": 1.448166044366352, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.212470293045044, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7462292313575745, + "num_tokens": 341237002.0, + "step": 13187 + }, + { + "epoch": 1.4482758620689655, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.03574800491333, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7123950719833374, + "num_tokens": 341268766.0, + "step": 13188 + }, + { + "epoch": 1.448385679771579, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.5868008136749268, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.72660231590271, + "num_tokens": 341290040.0, + "step": 13189 + }, + { + "epoch": 1.4484954974741928, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.97443687915802, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6911232471466064, + "num_tokens": 341322524.0, + "step": 13190 + }, + { + "epoch": 1.4486053151768066, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.3800997734069824, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7062418460845947, + "num_tokens": 341345369.0, + "step": 13191 + }, + { + "epoch": 1.44871513287942, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2908971309661865, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7218626737594604, + "num_tokens": 341370538.0, + "step": 13192 + }, + { + "epoch": 1.4488249505820339, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 1.9628396034240723, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7065740823745728, + "num_tokens": 341402489.0, + "step": 13193 + }, + { + "epoch": 1.4489347682846474, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2273786067962646, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7183501720428467, + "num_tokens": 341428296.0, + "step": 13194 + }, + { + "epoch": 1.4490445859872612, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2003231048583984, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.706270158290863, + "num_tokens": 341455598.0, + "step": 13195 + }, + { + "epoch": 1.449154403689875, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.116985559463501, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7154984474182129, + "num_tokens": 341485051.0, + "step": 13196 + }, + { + "epoch": 1.4492642213924884, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2513270378112793, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7172303199768066, + "num_tokens": 341508687.0, + "step": 13197 + }, + { + "epoch": 1.4493740390951022, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1175999641418457, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7479338645935059, + "num_tokens": 341536130.0, + "step": 13198 + }, + { + "epoch": 1.4494838567977157, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.266151189804077, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7222484946250916, + "num_tokens": 341560952.0, + "step": 13199 + }, + { + "epoch": 1.4495936745003295, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.0795955657958984, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7104526162147522, + "num_tokens": 341591777.0, + "step": 13200 + }, + { + "epoch": 1.449703492202943, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2308192253112793, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7123621106147766, + "num_tokens": 341621448.0, + "step": 13201 + }, + { + "epoch": 1.4498133099055568, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.353818416595459, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6990017890930176, + "num_tokens": 341647764.0, + "step": 13202 + }, + { + "epoch": 1.4499231276081703, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.1925318241119385, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6861288547515869, + "num_tokens": 341677722.0, + "step": 13203 + }, + { + "epoch": 1.450032945310784, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.47078537940979, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7076472043991089, + "num_tokens": 341699976.0, + "step": 13204 + }, + { + "epoch": 1.4501427630133978, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.197883367538452, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7317396402359009, + "num_tokens": 341727137.0, + "step": 13205 + }, + { + "epoch": 1.4502525807160114, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.0317764282226562, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7171241641044617, + "num_tokens": 341756721.0, + "step": 13206 + }, + { + "epoch": 1.4503623984186251, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.3766705989837646, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.701662540435791, + "num_tokens": 341781790.0, + "step": 13207 + }, + { + "epoch": 1.4504722161212387, + "ewc_loss": 0.00011920928955078125, + "grad_norm": 2.366241455078125, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7256870269775391, + "num_tokens": 341804509.0, + "step": 13208 + }, + { + "epoch": 1.4505820338238524, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.5444531440734863, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7018229961395264, + "num_tokens": 341828413.0, + "step": 13209 + }, + { + "epoch": 1.4506918515264662, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2598273754119873, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7319585084915161, + "num_tokens": 341851612.0, + "step": 13210 + }, + { + "epoch": 1.4508016692290797, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.438894510269165, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7536334991455078, + "num_tokens": 341873403.0, + "step": 13211 + }, + { + "epoch": 1.4509114869316933, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.0675528049468994, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7204732894897461, + "num_tokens": 341903244.0, + "step": 13212 + }, + { + "epoch": 1.451021304634307, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.803927183151245, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.723109245300293, + "num_tokens": 341921413.0, + "step": 13213 + }, + { + "epoch": 1.4511311223369208, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.284069061279297, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7258136868476868, + "num_tokens": 341946752.0, + "step": 13214 + }, + { + "epoch": 1.4512409400395343, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2051637172698975, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7242938280105591, + "num_tokens": 341974994.0, + "step": 13215 + }, + { + "epoch": 1.451350757742148, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.2504360675811768, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7113468050956726, + "num_tokens": 342001400.0, + "step": 13216 + }, + { + "epoch": 1.4514605754447616, + "ewc_loss": 0.00011873245239257812, + "grad_norm": 2.419045925140381, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7074166536331177, + "num_tokens": 342025161.0, + "step": 13217 + }, + { + "epoch": 1.4515703931473753, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.19565486907959, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7262148261070251, + "num_tokens": 342052036.0, + "step": 13218 + }, + { + "epoch": 1.451680210849989, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0238280296325684, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7098087072372437, + "num_tokens": 342081345.0, + "step": 13219 + }, + { + "epoch": 1.4517900285526026, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4407808780670166, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7181746363639832, + "num_tokens": 342103168.0, + "step": 13220 + }, + { + "epoch": 1.4518998462552164, + "ewc_loss": 0.00011968612670898438, + "grad_norm": 2.2140090465545654, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7171436548233032, + "num_tokens": 342128722.0, + "step": 13221 + }, + { + "epoch": 1.45200966395783, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.239699363708496, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7116567492485046, + "num_tokens": 342153962.0, + "step": 13222 + }, + { + "epoch": 1.4521194816604437, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.445005178451538, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7003374695777893, + "num_tokens": 342177007.0, + "step": 13223 + }, + { + "epoch": 1.4522292993630574, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.240816116333008, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7196072340011597, + "num_tokens": 342204681.0, + "step": 13224 + }, + { + "epoch": 1.452339117065671, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5425093173980713, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7446516752243042, + "num_tokens": 342223924.0, + "step": 13225 + }, + { + "epoch": 1.4524489347682845, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1886277198791504, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7329750657081604, + "num_tokens": 342250807.0, + "step": 13226 + }, + { + "epoch": 1.4525587524708983, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1535441875457764, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7262250185012817, + "num_tokens": 342275965.0, + "step": 13227 + }, + { + "epoch": 1.452668570173512, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0322282314300537, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6861386299133301, + "num_tokens": 342307194.0, + "step": 13228 + }, + { + "epoch": 1.4527783878761256, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.144667625427246, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7152068614959717, + "num_tokens": 342335970.0, + "step": 13229 + }, + { + "epoch": 1.4528882055787393, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.36612868309021, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7203402519226074, + "num_tokens": 342359907.0, + "step": 13230 + }, + { + "epoch": 1.4529980232813529, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2695446014404297, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.726300835609436, + "num_tokens": 342384495.0, + "step": 13231 + }, + { + "epoch": 1.4531078409839666, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9430397748947144, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6999791860580444, + "num_tokens": 342419582.0, + "step": 13232 + }, + { + "epoch": 1.4532176586865804, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.233649253845215, + "learning_rate": 1e-06, + "loss": 0.9919, + "mean_token_accuracy": 0.6950549483299255, + "num_tokens": 342446851.0, + "step": 13233 + }, + { + "epoch": 1.453327476389194, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3488194942474365, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7356221079826355, + "num_tokens": 342470904.0, + "step": 13234 + }, + { + "epoch": 1.4534372940918077, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4925599098205566, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7285321950912476, + "num_tokens": 342491848.0, + "step": 13235 + }, + { + "epoch": 1.4535471117944212, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9998297691345215, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7140417098999023, + "num_tokens": 342521773.0, + "step": 13236 + }, + { + "epoch": 1.453656929497035, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.234234094619751, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6947847604751587, + "num_tokens": 342548782.0, + "step": 13237 + }, + { + "epoch": 1.4537667471996487, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0615170001983643, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6876873970031738, + "num_tokens": 342581713.0, + "step": 13238 + }, + { + "epoch": 1.4538765649022622, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.121605396270752, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7305601835250854, + "num_tokens": 342608817.0, + "step": 13239 + }, + { + "epoch": 1.4539863826048758, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.040442705154419, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7247171401977539, + "num_tokens": 342635871.0, + "step": 13240 + }, + { + "epoch": 1.4540962003074895, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.232564926147461, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.718163251876831, + "num_tokens": 342660502.0, + "step": 13241 + }, + { + "epoch": 1.4542060180101033, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4479053020477295, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7293675541877747, + "num_tokens": 342681603.0, + "step": 13242 + }, + { + "epoch": 1.4543158357127168, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.15602707862854, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7305513620376587, + "num_tokens": 342707872.0, + "step": 13243 + }, + { + "epoch": 1.4544256534153306, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2275054454803467, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6950969696044922, + "num_tokens": 342738552.0, + "step": 13244 + }, + { + "epoch": 1.4545354711179441, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3213653564453125, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7239805459976196, + "num_tokens": 342762795.0, + "step": 13245 + }, + { + "epoch": 1.4546452888205579, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.380082845687866, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7271910905838013, + "num_tokens": 342787005.0, + "step": 13246 + }, + { + "epoch": 1.4547551065231716, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.493821859359741, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7113620042800903, + "num_tokens": 342807908.0, + "step": 13247 + }, + { + "epoch": 1.4548649242257852, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.063593864440918, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7007333636283875, + "num_tokens": 342836125.0, + "step": 13248 + }, + { + "epoch": 1.454974741928399, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9210485219955444, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7341535091400146, + "num_tokens": 342868122.0, + "step": 13249 + }, + { + "epoch": 1.4550845596310125, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.111964225769043, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7263889908790588, + "num_tokens": 342895977.0, + "step": 13250 + }, + { + "epoch": 1.4551943773336262, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2248120307922363, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7061275243759155, + "num_tokens": 342921265.0, + "step": 13251 + }, + { + "epoch": 1.45530419503624, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2280473709106445, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7215080261230469, + "num_tokens": 342948013.0, + "step": 13252 + }, + { + "epoch": 1.4554140127388535, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9784328937530518, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7195480465888977, + "num_tokens": 342979030.0, + "step": 13253 + }, + { + "epoch": 1.455523830441467, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3802051544189453, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7262847423553467, + "num_tokens": 343001334.0, + "step": 13254 + }, + { + "epoch": 1.4556336481440808, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.165505886077881, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7090275287628174, + "num_tokens": 343030971.0, + "step": 13255 + }, + { + "epoch": 1.4557434658466946, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.318981647491455, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7039510607719421, + "num_tokens": 343057286.0, + "step": 13256 + }, + { + "epoch": 1.455853283549308, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.098905563354492, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7084180116653442, + "num_tokens": 343087943.0, + "step": 13257 + }, + { + "epoch": 1.4559631012519219, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4016242027282715, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7305259704589844, + "num_tokens": 343110953.0, + "step": 13258 + }, + { + "epoch": 1.4560729189545354, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.431303024291992, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.718036413192749, + "num_tokens": 343133106.0, + "step": 13259 + }, + { + "epoch": 1.4561827366571491, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3219897747039795, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7263108491897583, + "num_tokens": 343157036.0, + "step": 13260 + }, + { + "epoch": 1.456292554359763, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.143155097961426, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6905814409255981, + "num_tokens": 343186413.0, + "step": 13261 + }, + { + "epoch": 1.4564023720623764, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9793812036514282, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.700068473815918, + "num_tokens": 343218981.0, + "step": 13262 + }, + { + "epoch": 1.4565121897649902, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.143963098526001, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7061281800270081, + "num_tokens": 343245759.0, + "step": 13263 + }, + { + "epoch": 1.4566220074676037, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4756920337677, + "learning_rate": 1e-06, + "loss": 0.7383, + "mean_token_accuracy": 0.76204913854599, + "num_tokens": 343265382.0, + "step": 13264 + }, + { + "epoch": 1.4567318251702175, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5483486652374268, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7049808502197266, + "num_tokens": 343287786.0, + "step": 13265 + }, + { + "epoch": 1.456841642872831, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.231008529663086, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7245808839797974, + "num_tokens": 343312571.0, + "step": 13266 + }, + { + "epoch": 1.4569514605754448, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.177387237548828, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.6997403502464294, + "num_tokens": 343338845.0, + "step": 13267 + }, + { + "epoch": 1.4570612782780583, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1009836196899414, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7104918956756592, + "num_tokens": 343366259.0, + "step": 13268 + }, + { + "epoch": 1.457171095980672, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4221763610839844, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7076135873794556, + "num_tokens": 343389709.0, + "step": 13269 + }, + { + "epoch": 1.4572809136832858, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.526602268218994, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7551900744438171, + "num_tokens": 343409779.0, + "step": 13270 + }, + { + "epoch": 1.4573907313858994, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.417206287384033, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7262259125709534, + "num_tokens": 343433056.0, + "step": 13271 + }, + { + "epoch": 1.4575005490885131, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4311797618865967, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7193634510040283, + "num_tokens": 343455858.0, + "step": 13272 + }, + { + "epoch": 1.4576103667911267, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.652676820755005, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7252376675605774, + "num_tokens": 343474974.0, + "step": 13273 + }, + { + "epoch": 1.4577201844937404, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5434060096740723, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7186917662620544, + "num_tokens": 343494711.0, + "step": 13274 + }, + { + "epoch": 1.4578300021963542, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2285430431365967, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7128989100456238, + "num_tokens": 343523387.0, + "step": 13275 + }, + { + "epoch": 1.4579398198989677, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5292131900787354, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7368324995040894, + "num_tokens": 343544079.0, + "step": 13276 + }, + { + "epoch": 1.4580496376015812, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0567190647125244, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7199705839157104, + "num_tokens": 343571845.0, + "step": 13277 + }, + { + "epoch": 1.458159455304195, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.184849739074707, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7259682416915894, + "num_tokens": 343596591.0, + "step": 13278 + }, + { + "epoch": 1.4582692730068088, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.143277883529663, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7293473482131958, + "num_tokens": 343622319.0, + "step": 13279 + }, + { + "epoch": 1.4583790907094223, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.287914991378784, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7189161777496338, + "num_tokens": 343645477.0, + "step": 13280 + }, + { + "epoch": 1.458488908412036, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.595771074295044, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7157942652702332, + "num_tokens": 343664693.0, + "step": 13281 + }, + { + "epoch": 1.4585987261146496, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5835890769958496, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7469631433486938, + "num_tokens": 343683758.0, + "step": 13282 + }, + { + "epoch": 1.4587085438172633, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3673579692840576, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.708733081817627, + "num_tokens": 343708334.0, + "step": 13283 + }, + { + "epoch": 1.458818361519877, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0893521308898926, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7128086686134338, + "num_tokens": 343740386.0, + "step": 13284 + }, + { + "epoch": 1.4589281792224906, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2852883338928223, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7313212752342224, + "num_tokens": 343765478.0, + "step": 13285 + }, + { + "epoch": 1.4590379969251044, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3707969188690186, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7300172448158264, + "num_tokens": 343790678.0, + "step": 13286 + }, + { + "epoch": 1.459147814627718, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0794601440429688, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7146111130714417, + "num_tokens": 343818608.0, + "step": 13287 + }, + { + "epoch": 1.4592576323303317, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.107865571975708, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7243239879608154, + "num_tokens": 343844894.0, + "step": 13288 + }, + { + "epoch": 1.4593674500329454, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.396543264389038, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7201778888702393, + "num_tokens": 343866536.0, + "step": 13289 + }, + { + "epoch": 1.459477267735559, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.011425256729126, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.716439425945282, + "num_tokens": 343894798.0, + "step": 13290 + }, + { + "epoch": 1.4595870854381725, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2165396213531494, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.6999226808547974, + "num_tokens": 343921430.0, + "step": 13291 + }, + { + "epoch": 1.4596969031407863, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.082059383392334, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7336531281471252, + "num_tokens": 343949715.0, + "step": 13292 + }, + { + "epoch": 1.4598067208434, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2787930965423584, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7345480918884277, + "num_tokens": 343974416.0, + "step": 13293 + }, + { + "epoch": 1.4599165385460136, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.143958330154419, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.6966109275817871, + "num_tokens": 344004519.0, + "step": 13294 + }, + { + "epoch": 1.4600263562486273, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2394704818725586, + "learning_rate": 1e-06, + "loss": 0.7547, + "mean_token_accuracy": 0.7639554738998413, + "num_tokens": 344028304.0, + "step": 13295 + }, + { + "epoch": 1.4601361739512408, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2768325805664062, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7347244024276733, + "num_tokens": 344053256.0, + "step": 13296 + }, + { + "epoch": 1.4602459916538546, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.719970703125, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7232629656791687, + "num_tokens": 344073480.0, + "step": 13297 + }, + { + "epoch": 1.4603558093564684, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0438313484191895, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7010828256607056, + "num_tokens": 344103404.0, + "step": 13298 + }, + { + "epoch": 1.460465627059082, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1576309204101562, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7237366437911987, + "num_tokens": 344129978.0, + "step": 13299 + }, + { + "epoch": 1.4605754447616957, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2719509601593018, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7419730424880981, + "num_tokens": 344153913.0, + "step": 13300 + }, + { + "epoch": 1.4606852624643092, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3747494220733643, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7204155921936035, + "num_tokens": 344179762.0, + "step": 13301 + }, + { + "epoch": 1.460795080166923, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.382384777069092, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7237207889556885, + "num_tokens": 344203479.0, + "step": 13302 + }, + { + "epoch": 1.4609048978695367, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.392727851867676, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7129693031311035, + "num_tokens": 344227702.0, + "step": 13303 + }, + { + "epoch": 1.4610147155721502, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4962093830108643, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7228127717971802, + "num_tokens": 344250615.0, + "step": 13304 + }, + { + "epoch": 1.4611245332747638, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0988221168518066, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6989597678184509, + "num_tokens": 344279535.0, + "step": 13305 + }, + { + "epoch": 1.4612343509773775, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.8843753337860107, + "learning_rate": 1e-06, + "loss": 1.1153, + "mean_token_accuracy": 0.6678730845451355, + "num_tokens": 344315144.0, + "step": 13306 + }, + { + "epoch": 1.4613441686799913, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.30660343170166, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7139990329742432, + "num_tokens": 344338533.0, + "step": 13307 + }, + { + "epoch": 1.4614539863826048, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2567689418792725, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7086702585220337, + "num_tokens": 344364248.0, + "step": 13308 + }, + { + "epoch": 1.4615638040852186, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.5196568965911865, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7290849089622498, + "num_tokens": 344386105.0, + "step": 13309 + }, + { + "epoch": 1.4616736217878321, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4604272842407227, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7222088575363159, + "num_tokens": 344407088.0, + "step": 13310 + }, + { + "epoch": 1.4617834394904459, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.461322546005249, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7148399353027344, + "num_tokens": 344428868.0, + "step": 13311 + }, + { + "epoch": 1.4618932571930596, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2857484817504883, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.706908106803894, + "num_tokens": 344454809.0, + "step": 13312 + }, + { + "epoch": 1.4620030748956732, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1397006511688232, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.69199538230896, + "num_tokens": 344487596.0, + "step": 13313 + }, + { + "epoch": 1.462112892598287, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.254028081893921, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6970401406288147, + "num_tokens": 344515855.0, + "step": 13314 + }, + { + "epoch": 1.4622227103009005, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.274355411529541, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7275192141532898, + "num_tokens": 344540736.0, + "step": 13315 + }, + { + "epoch": 1.4623325280035142, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2170960903167725, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7007545232772827, + "num_tokens": 344569990.0, + "step": 13316 + }, + { + "epoch": 1.462442345706128, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0839295387268066, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7296253442764282, + "num_tokens": 344598249.0, + "step": 13317 + }, + { + "epoch": 1.4625521634087415, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.305413246154785, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.6971592307090759, + "num_tokens": 344623371.0, + "step": 13318 + }, + { + "epoch": 1.462661981111355, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.6085753440856934, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7283275127410889, + "num_tokens": 344645429.0, + "step": 13319 + }, + { + "epoch": 1.4627717988139688, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.278418779373169, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7310138940811157, + "num_tokens": 344668941.0, + "step": 13320 + }, + { + "epoch": 1.4628816165165826, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.612035036087036, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7514983415603638, + "num_tokens": 344687947.0, + "step": 13321 + }, + { + "epoch": 1.462991434219196, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.190545082092285, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7460164427757263, + "num_tokens": 344714531.0, + "step": 13322 + }, + { + "epoch": 1.4631012519218098, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.960940957069397, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7204544544219971, + "num_tokens": 344745952.0, + "step": 13323 + }, + { + "epoch": 1.4632110696244234, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.307680606842041, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7217026948928833, + "num_tokens": 344769489.0, + "step": 13324 + }, + { + "epoch": 1.4633208873270371, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.202068567276001, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6889966726303101, + "num_tokens": 344798325.0, + "step": 13325 + }, + { + "epoch": 1.463430705029651, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.109881639480591, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7151776552200317, + "num_tokens": 344826816.0, + "step": 13326 + }, + { + "epoch": 1.4635405227322644, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0381574630737305, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6997183561325073, + "num_tokens": 344857874.0, + "step": 13327 + }, + { + "epoch": 1.4636503404348782, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2676198482513428, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7051559090614319, + "num_tokens": 344883472.0, + "step": 13328 + }, + { + "epoch": 1.4637601581374917, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.229175567626953, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7163096070289612, + "num_tokens": 344911717.0, + "step": 13329 + }, + { + "epoch": 1.4638699758401055, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.3472096920013428, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7072774171829224, + "num_tokens": 344935917.0, + "step": 13330 + }, + { + "epoch": 1.463979793542719, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.315502166748047, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7258397936820984, + "num_tokens": 344960203.0, + "step": 13331 + }, + { + "epoch": 1.4640896112453328, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.4515085220336914, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7291281223297119, + "num_tokens": 344983619.0, + "step": 13332 + }, + { + "epoch": 1.4641994289479463, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 1.9851793050765991, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.714238166809082, + "num_tokens": 345015122.0, + "step": 13333 + }, + { + "epoch": 1.46430924665056, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.0052096843719482, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.714122474193573, + "num_tokens": 345045631.0, + "step": 13334 + }, + { + "epoch": 1.4644190643531738, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.078986883163452, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7171851396560669, + "num_tokens": 345076630.0, + "step": 13335 + }, + { + "epoch": 1.4645288820557874, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1573052406311035, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7382343411445618, + "num_tokens": 345101823.0, + "step": 13336 + }, + { + "epoch": 1.464638699758401, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.304309606552124, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7192448377609253, + "num_tokens": 345128795.0, + "step": 13337 + }, + { + "epoch": 1.4647485174610146, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1710410118103027, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7165666222572327, + "num_tokens": 345156852.0, + "step": 13338 + }, + { + "epoch": 1.4648583351636284, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.325507402420044, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7278189063072205, + "num_tokens": 345182349.0, + "step": 13339 + }, + { + "epoch": 1.4649681528662422, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1850757598876953, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7324907779693604, + "num_tokens": 345210182.0, + "step": 13340 + }, + { + "epoch": 1.4650779705688557, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.517472743988037, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.7533509731292725, + "num_tokens": 345230677.0, + "step": 13341 + }, + { + "epoch": 1.4651877882714692, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2108535766601562, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7242489457130432, + "num_tokens": 345256864.0, + "step": 13342 + }, + { + "epoch": 1.465297605974083, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1835765838623047, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7381069660186768, + "num_tokens": 345283478.0, + "step": 13343 + }, + { + "epoch": 1.4654074236766967, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.271559238433838, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7098398208618164, + "num_tokens": 345307144.0, + "step": 13344 + }, + { + "epoch": 1.4655172413793103, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.004693031311035, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.701836347579956, + "num_tokens": 345336359.0, + "step": 13345 + }, + { + "epoch": 1.465627059081924, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.2678470611572266, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7330180406570435, + "num_tokens": 345359927.0, + "step": 13346 + }, + { + "epoch": 1.4657368767845376, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.61430025100708, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7340308427810669, + "num_tokens": 345380985.0, + "step": 13347 + }, + { + "epoch": 1.4658466944871513, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1413753032684326, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.692072868347168, + "num_tokens": 345409928.0, + "step": 13348 + }, + { + "epoch": 1.465956512189765, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.1630845069885254, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7205606698989868, + "num_tokens": 345435622.0, + "step": 13349 + }, + { + "epoch": 1.4660663298923786, + "ewc_loss": 0.0001201629638671875, + "grad_norm": 2.412592649459839, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7175655961036682, + "num_tokens": 345457381.0, + "step": 13350 + }, + { + "epoch": 1.4661761475949924, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.357501268386841, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7299751043319702, + "num_tokens": 345480374.0, + "step": 13351 + }, + { + "epoch": 1.466285965297606, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2889318466186523, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7515091896057129, + "num_tokens": 345505238.0, + "step": 13352 + }, + { + "epoch": 1.4663957830002197, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.28242564201355, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6990726590156555, + "num_tokens": 345531950.0, + "step": 13353 + }, + { + "epoch": 1.4665056007028334, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.372130870819092, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6957739591598511, + "num_tokens": 345556257.0, + "step": 13354 + }, + { + "epoch": 1.466615418405447, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2913951873779297, + "learning_rate": 1e-06, + "loss": 1.0602, + "mean_token_accuracy": 0.6927884817123413, + "num_tokens": 345586451.0, + "step": 13355 + }, + { + "epoch": 1.4667252361080605, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.5202126502990723, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7183998823165894, + "num_tokens": 345610752.0, + "step": 13356 + }, + { + "epoch": 1.4668350538106742, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.6500661373138428, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7319077253341675, + "num_tokens": 345631872.0, + "step": 13357 + }, + { + "epoch": 1.466944871513288, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.128892421722412, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7017179131507874, + "num_tokens": 345661342.0, + "step": 13358 + }, + { + "epoch": 1.4670546892159015, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3985016345977783, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7542343139648438, + "num_tokens": 345683908.0, + "step": 13359 + }, + { + "epoch": 1.4671645069185153, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.251948118209839, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7076996564865112, + "num_tokens": 345709061.0, + "step": 13360 + }, + { + "epoch": 1.4672743246211288, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2884435653686523, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7213802337646484, + "num_tokens": 345734093.0, + "step": 13361 + }, + { + "epoch": 1.4673841423237426, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1824021339416504, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7044512629508972, + "num_tokens": 345762128.0, + "step": 13362 + }, + { + "epoch": 1.4674939600263563, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2558093070983887, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7230392098426819, + "num_tokens": 345785034.0, + "step": 13363 + }, + { + "epoch": 1.4676037777289699, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.301661968231201, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7202484011650085, + "num_tokens": 345808752.0, + "step": 13364 + }, + { + "epoch": 1.4677135954315836, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.148669958114624, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7106478810310364, + "num_tokens": 345838267.0, + "step": 13365 + }, + { + "epoch": 1.4678234131341972, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3058111667633057, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7372801303863525, + "num_tokens": 345861573.0, + "step": 13366 + }, + { + "epoch": 1.467933230836811, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1458308696746826, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.717055082321167, + "num_tokens": 345888270.0, + "step": 13367 + }, + { + "epoch": 1.4680430485394247, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.5292468070983887, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.733589768409729, + "num_tokens": 345909122.0, + "step": 13368 + }, + { + "epoch": 1.4681528662420382, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.934325098991394, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7057605981826782, + "num_tokens": 345941407.0, + "step": 13369 + }, + { + "epoch": 1.4682626839446518, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1705398559570312, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7100770473480225, + "num_tokens": 345967749.0, + "step": 13370 + }, + { + "epoch": 1.4683725016472655, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.6304216384887695, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.742487907409668, + "num_tokens": 345986427.0, + "step": 13371 + }, + { + "epoch": 1.4684823193498793, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2819015979766846, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.719340443611145, + "num_tokens": 346014159.0, + "step": 13372 + }, + { + "epoch": 1.4685921370524928, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1017119884490967, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7246329188346863, + "num_tokens": 346043607.0, + "step": 13373 + }, + { + "epoch": 1.4687019547551066, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.363974094390869, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7235491871833801, + "num_tokens": 346068148.0, + "step": 13374 + }, + { + "epoch": 1.46881177245772, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.491223096847534, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7178827524185181, + "num_tokens": 346089553.0, + "step": 13375 + }, + { + "epoch": 1.4689215901603339, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1542251110076904, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7101580500602722, + "num_tokens": 346119179.0, + "step": 13376 + }, + { + "epoch": 1.4690314078629476, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.53330397605896, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7484263777732849, + "num_tokens": 346137783.0, + "step": 13377 + }, + { + "epoch": 1.4691412255655611, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.305420160293579, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7128285765647888, + "num_tokens": 346163726.0, + "step": 13378 + }, + { + "epoch": 1.469251043268175, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.06544828414917, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7098455429077148, + "num_tokens": 346192312.0, + "step": 13379 + }, + { + "epoch": 1.4693608609707884, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.136967897415161, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7048152685165405, + "num_tokens": 346221468.0, + "step": 13380 + }, + { + "epoch": 1.4694706786734022, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.143852472305298, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7166124582290649, + "num_tokens": 346247660.0, + "step": 13381 + }, + { + "epoch": 1.4695804963760157, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1775965690612793, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7144337296485901, + "num_tokens": 346275419.0, + "step": 13382 + }, + { + "epoch": 1.4696903140786295, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.366807222366333, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7303321361541748, + "num_tokens": 346299583.0, + "step": 13383 + }, + { + "epoch": 1.469800131781243, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.679126262664795, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7345521450042725, + "num_tokens": 346319314.0, + "step": 13384 + }, + { + "epoch": 1.4699099494838568, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.229249954223633, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6997648477554321, + "num_tokens": 346344487.0, + "step": 13385 + }, + { + "epoch": 1.4700197671864705, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1342921257019043, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7273861169815063, + "num_tokens": 346371288.0, + "step": 13386 + }, + { + "epoch": 1.470129584889084, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2665576934814453, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7296172976493835, + "num_tokens": 346397949.0, + "step": 13387 + }, + { + "epoch": 1.4702394025916978, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.574768543243408, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7261204719543457, + "num_tokens": 346418085.0, + "step": 13388 + }, + { + "epoch": 1.4703492202943114, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.134993314743042, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7115660905838013, + "num_tokens": 346445774.0, + "step": 13389 + }, + { + "epoch": 1.4704590379969251, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.9276293516159058, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7115944623947144, + "num_tokens": 346479092.0, + "step": 13390 + }, + { + "epoch": 1.4705688556995389, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.6382875442504883, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7082418203353882, + "num_tokens": 346500975.0, + "step": 13391 + }, + { + "epoch": 1.4706786734021524, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3017194271087646, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7196074724197388, + "num_tokens": 346524780.0, + "step": 13392 + }, + { + "epoch": 1.4707884911047662, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2684743404388428, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7004961967468262, + "num_tokens": 346554229.0, + "step": 13393 + }, + { + "epoch": 1.4708983088073797, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1352345943450928, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7346029281616211, + "num_tokens": 346581471.0, + "step": 13394 + }, + { + "epoch": 1.4710081265099935, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.262094497680664, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7281768321990967, + "num_tokens": 346606506.0, + "step": 13395 + }, + { + "epoch": 1.471117944212607, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2545249462127686, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7040982246398926, + "num_tokens": 346632543.0, + "step": 13396 + }, + { + "epoch": 1.4712277619152208, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.524791955947876, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7258259654045105, + "num_tokens": 346654268.0, + "step": 13397 + }, + { + "epoch": 1.4713375796178343, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.057356595993042, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7545886039733887, + "num_tokens": 346681670.0, + "step": 13398 + }, + { + "epoch": 1.471447397320448, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2554569244384766, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7412368059158325, + "num_tokens": 346706343.0, + "step": 13399 + }, + { + "epoch": 1.4715572150230618, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2101328372955322, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7261784076690674, + "num_tokens": 346731063.0, + "step": 13400 + }, + { + "epoch": 1.4716670327256753, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.331632614135742, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.723671555519104, + "num_tokens": 346755843.0, + "step": 13401 + }, + { + "epoch": 1.471776850428289, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.20566463470459, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7105899453163147, + "num_tokens": 346782314.0, + "step": 13402 + }, + { + "epoch": 1.4718866681309026, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.457866668701172, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7360589504241943, + "num_tokens": 346804881.0, + "step": 13403 + }, + { + "epoch": 1.4719964858335164, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2316410541534424, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.6941536068916321, + "num_tokens": 346833038.0, + "step": 13404 + }, + { + "epoch": 1.4721063035361301, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.4440715312957764, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7365375757217407, + "num_tokens": 346854806.0, + "step": 13405 + }, + { + "epoch": 1.4722161212387437, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.0979132652282715, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.700124979019165, + "num_tokens": 346886827.0, + "step": 13406 + }, + { + "epoch": 1.4723259389413572, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2995188236236572, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7191483378410339, + "num_tokens": 346909719.0, + "step": 13407 + }, + { + "epoch": 1.472435756643971, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.113121509552002, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7207298874855042, + "num_tokens": 346936087.0, + "step": 13408 + }, + { + "epoch": 1.4725455743465847, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3253262042999268, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7028233408927917, + "num_tokens": 346962253.0, + "step": 13409 + }, + { + "epoch": 1.4726553920491983, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2578628063201904, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7161718606948853, + "num_tokens": 346987083.0, + "step": 13410 + }, + { + "epoch": 1.472765209751812, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.468181610107422, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7352432012557983, + "num_tokens": 347008954.0, + "step": 13411 + }, + { + "epoch": 1.4728750274544256, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.229518175125122, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7072733044624329, + "num_tokens": 347037526.0, + "step": 13412 + }, + { + "epoch": 1.4729848451570393, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2668285369873047, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7307537794113159, + "num_tokens": 347060196.0, + "step": 13413 + }, + { + "epoch": 1.473094662859653, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.683969259262085, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.736173152923584, + "num_tokens": 347079279.0, + "step": 13414 + }, + { + "epoch": 1.4732044805622666, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3770484924316406, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7213074564933777, + "num_tokens": 347101296.0, + "step": 13415 + }, + { + "epoch": 1.4733142982648804, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.301433563232422, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7273662090301514, + "num_tokens": 347124779.0, + "step": 13416 + }, + { + "epoch": 1.473424115967494, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.361645460128784, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.727324366569519, + "num_tokens": 347148769.0, + "step": 13417 + }, + { + "epoch": 1.4735339336701077, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.088871955871582, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7317423820495605, + "num_tokens": 347176103.0, + "step": 13418 + }, + { + "epoch": 1.4736437513727214, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2677125930786133, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.718755304813385, + "num_tokens": 347201415.0, + "step": 13419 + }, + { + "epoch": 1.473753569075335, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1768386363983154, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7368102669715881, + "num_tokens": 347228122.0, + "step": 13420 + }, + { + "epoch": 1.4738633867779485, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1260125637054443, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7198946475982666, + "num_tokens": 347254149.0, + "step": 13421 + }, + { + "epoch": 1.4739732044805622, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3644654750823975, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.729726254940033, + "num_tokens": 347278195.0, + "step": 13422 + }, + { + "epoch": 1.474083022183176, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3445839881896973, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7473034858703613, + "num_tokens": 347301112.0, + "step": 13423 + }, + { + "epoch": 1.4741928398857895, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2181715965270996, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7166932225227356, + "num_tokens": 347326955.0, + "step": 13424 + }, + { + "epoch": 1.4743026575884033, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.289154291152954, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7177606225013733, + "num_tokens": 347352680.0, + "step": 13425 + }, + { + "epoch": 1.4744124752910168, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.243324041366577, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7155638933181763, + "num_tokens": 347379559.0, + "step": 13426 + }, + { + "epoch": 1.4745222929936306, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3212063312530518, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7352641820907593, + "num_tokens": 347404330.0, + "step": 13427 + }, + { + "epoch": 1.4746321106962443, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1252312660217285, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7165545225143433, + "num_tokens": 347432882.0, + "step": 13428 + }, + { + "epoch": 1.4747419283988579, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1667683124542236, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7235710024833679, + "num_tokens": 347460572.0, + "step": 13429 + }, + { + "epoch": 1.4748517461014716, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.322139263153076, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7091817855834961, + "num_tokens": 347486650.0, + "step": 13430 + }, + { + "epoch": 1.4749615638040852, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1950438022613525, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7279952764511108, + "num_tokens": 347514758.0, + "step": 13431 + }, + { + "epoch": 1.475071381506699, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.271390676498413, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7188447713851929, + "num_tokens": 347539986.0, + "step": 13432 + }, + { + "epoch": 1.4751811992093127, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2385542392730713, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7136995792388916, + "num_tokens": 347563513.0, + "step": 13433 + }, + { + "epoch": 1.4752910169119262, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.092911958694458, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7282335758209229, + "num_tokens": 347591159.0, + "step": 13434 + }, + { + "epoch": 1.4754008346145397, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.162741184234619, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7225888967514038, + "num_tokens": 347618018.0, + "step": 13435 + }, + { + "epoch": 1.4755106523171535, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1571388244628906, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7025057077407837, + "num_tokens": 347645281.0, + "step": 13436 + }, + { + "epoch": 1.4756204700197673, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.306198835372925, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7198334336280823, + "num_tokens": 347669321.0, + "step": 13437 + }, + { + "epoch": 1.4757302877223808, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1749370098114014, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7084454894065857, + "num_tokens": 347696628.0, + "step": 13438 + }, + { + "epoch": 1.4758401054249946, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.318209409713745, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7092269062995911, + "num_tokens": 347720666.0, + "step": 13439 + }, + { + "epoch": 1.475949923127608, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2895305156707764, + "learning_rate": 1e-06, + "loss": 1.0384, + "mean_token_accuracy": 0.6899043917655945, + "num_tokens": 347748398.0, + "step": 13440 + }, + { + "epoch": 1.4760597408302218, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.0643255710601807, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6986225843429565, + "num_tokens": 347779219.0, + "step": 13441 + }, + { + "epoch": 1.4761695585328356, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.4071929454803467, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7193487882614136, + "num_tokens": 347800401.0, + "step": 13442 + }, + { + "epoch": 1.4762793762354491, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.340897798538208, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7350966930389404, + "num_tokens": 347824024.0, + "step": 13443 + }, + { + "epoch": 1.476389193938063, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2373318672180176, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.732785701751709, + "num_tokens": 347850139.0, + "step": 13444 + }, + { + "epoch": 1.4764990116406764, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1575987339019775, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7184920310974121, + "num_tokens": 347875516.0, + "step": 13445 + }, + { + "epoch": 1.4766088293432902, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.007319450378418, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6972929835319519, + "num_tokens": 347906133.0, + "step": 13446 + }, + { + "epoch": 1.4767186470459037, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.065319299697876, + "learning_rate": 1e-06, + "loss": 1.0083, + "mean_token_accuracy": 0.6947706341743469, + "num_tokens": 347938177.0, + "step": 13447 + }, + { + "epoch": 1.4768284647485175, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.203916072845459, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7058316469192505, + "num_tokens": 347964338.0, + "step": 13448 + }, + { + "epoch": 1.476938282451131, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.6426026821136475, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7227572202682495, + "num_tokens": 347983881.0, + "step": 13449 + }, + { + "epoch": 1.4770481001537448, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1051926612854004, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7125089168548584, + "num_tokens": 348011842.0, + "step": 13450 + }, + { + "epoch": 1.4771579178563585, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.331085681915283, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7204733490943909, + "num_tokens": 348037002.0, + "step": 13451 + }, + { + "epoch": 1.477267735558972, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.254152536392212, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7208211421966553, + "num_tokens": 348063768.0, + "step": 13452 + }, + { + "epoch": 1.4773775532615858, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.081212043762207, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7421532273292542, + "num_tokens": 348092857.0, + "step": 13453 + }, + { + "epoch": 1.4774873709641994, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.312291383743286, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7159267663955688, + "num_tokens": 348116796.0, + "step": 13454 + }, + { + "epoch": 1.477597188666813, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2373387813568115, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7284942865371704, + "num_tokens": 348142158.0, + "step": 13455 + }, + { + "epoch": 1.4777070063694269, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1220037937164307, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.710547149181366, + "num_tokens": 348170191.0, + "step": 13456 + }, + { + "epoch": 1.4778168240720404, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1982262134552, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7354840040206909, + "num_tokens": 348194238.0, + "step": 13457 + }, + { + "epoch": 1.477926641774654, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.9310163259506226, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7423454523086548, + "num_tokens": 348226443.0, + "step": 13458 + }, + { + "epoch": 1.4780364594772677, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2143194675445557, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7158629298210144, + "num_tokens": 348253733.0, + "step": 13459 + }, + { + "epoch": 1.4781462771798815, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3177902698516846, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7166631817817688, + "num_tokens": 348278012.0, + "step": 13460 + }, + { + "epoch": 1.478256094882495, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.5728936195373535, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7272486090660095, + "num_tokens": 348298231.0, + "step": 13461 + }, + { + "epoch": 1.4783659125851087, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.9127774238586426, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.725476086139679, + "num_tokens": 348330234.0, + "step": 13462 + }, + { + "epoch": 1.4784757302877223, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.0906896591186523, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7063503265380859, + "num_tokens": 348359195.0, + "step": 13463 + }, + { + "epoch": 1.478585547990336, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.125007390975952, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7190914750099182, + "num_tokens": 348385778.0, + "step": 13464 + }, + { + "epoch": 1.4786953656929498, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3981924057006836, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6895357370376587, + "num_tokens": 348409555.0, + "step": 13465 + }, + { + "epoch": 1.4788051833955633, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.147115468978882, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7258959412574768, + "num_tokens": 348436878.0, + "step": 13466 + }, + { + "epoch": 1.478915001098177, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2295429706573486, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7240623235702515, + "num_tokens": 348460782.0, + "step": 13467 + }, + { + "epoch": 1.4790248188007906, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.9253060817718506, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6889205574989319, + "num_tokens": 348493364.0, + "step": 13468 + }, + { + "epoch": 1.4791346365034044, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 1.9899570941925049, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7275696396827698, + "num_tokens": 348522052.0, + "step": 13469 + }, + { + "epoch": 1.4792444542060181, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.178805112838745, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7123857140541077, + "num_tokens": 348550787.0, + "step": 13470 + }, + { + "epoch": 1.4793542719086317, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.5519795417785645, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.746551513671875, + "num_tokens": 348569893.0, + "step": 13471 + }, + { + "epoch": 1.4794640896112452, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.596579074859619, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7396934628486633, + "num_tokens": 348589231.0, + "step": 13472 + }, + { + "epoch": 1.479573907313859, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3815407752990723, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7441455125808716, + "num_tokens": 348612204.0, + "step": 13473 + }, + { + "epoch": 1.4796837250164727, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.353787422180176, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7073766589164734, + "num_tokens": 348635785.0, + "step": 13474 + }, + { + "epoch": 1.4797935427190863, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.071446180343628, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.6987400054931641, + "num_tokens": 348667896.0, + "step": 13475 + }, + { + "epoch": 1.4799033604217, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.307840347290039, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7268456220626831, + "num_tokens": 348693407.0, + "step": 13476 + }, + { + "epoch": 1.4800131781243135, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.483372688293457, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.736690104007721, + "num_tokens": 348716369.0, + "step": 13477 + }, + { + "epoch": 1.4801229958269273, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1358418464660645, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7057043313980103, + "num_tokens": 348746705.0, + "step": 13478 + }, + { + "epoch": 1.480232813529541, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.535817861557007, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7501417398452759, + "num_tokens": 348766027.0, + "step": 13479 + }, + { + "epoch": 1.4803426312321546, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.0983810424804688, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7204372882843018, + "num_tokens": 348793646.0, + "step": 13480 + }, + { + "epoch": 1.4804524489347684, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2239718437194824, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7236766219139099, + "num_tokens": 348820525.0, + "step": 13481 + }, + { + "epoch": 1.4805622666373819, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2341442108154297, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7109382152557373, + "num_tokens": 348847130.0, + "step": 13482 + }, + { + "epoch": 1.4806720843399956, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3797595500946045, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7258949279785156, + "num_tokens": 348870238.0, + "step": 13483 + }, + { + "epoch": 1.4807819020426094, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.223809003829956, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.696564793586731, + "num_tokens": 348898170.0, + "step": 13484 + }, + { + "epoch": 1.480891719745223, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.2207624912261963, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7008442878723145, + "num_tokens": 348928003.0, + "step": 13485 + }, + { + "epoch": 1.4810015374478365, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.3869199752807617, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7233418226242065, + "num_tokens": 348953028.0, + "step": 13486 + }, + { + "epoch": 1.4811113551504502, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.145397901535034, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7392573356628418, + "num_tokens": 348981766.0, + "step": 13487 + }, + { + "epoch": 1.481221172853064, + "ewc_loss": 0.00012111663818359375, + "grad_norm": 2.1128244400024414, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7114514112472534, + "num_tokens": 349010130.0, + "step": 13488 + }, + { + "epoch": 1.4813309905556775, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2770638465881348, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.692790687084198, + "num_tokens": 349037423.0, + "step": 13489 + }, + { + "epoch": 1.4814408082582913, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2774081230163574, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.7018440365791321, + "num_tokens": 349063582.0, + "step": 13490 + }, + { + "epoch": 1.4815506259609048, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9330974817276, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7048242092132568, + "num_tokens": 349097443.0, + "step": 13491 + }, + { + "epoch": 1.4816604436635186, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.170383930206299, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7025724649429321, + "num_tokens": 349124836.0, + "step": 13492 + }, + { + "epoch": 1.4817702613661323, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.446861743927002, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.7000830173492432, + "num_tokens": 349150018.0, + "step": 13493 + }, + { + "epoch": 1.4818800790687459, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.12876558303833, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7179162502288818, + "num_tokens": 349178029.0, + "step": 13494 + }, + { + "epoch": 1.4819898967713596, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.455301284790039, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7057323455810547, + "num_tokens": 349201132.0, + "step": 13495 + }, + { + "epoch": 1.4820997144739732, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.450683116912842, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7311590909957886, + "num_tokens": 349222121.0, + "step": 13496 + }, + { + "epoch": 1.482209532176587, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.163304567337036, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7375329732894897, + "num_tokens": 349250114.0, + "step": 13497 + }, + { + "epoch": 1.4823193498792007, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0266780853271484, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7076456546783447, + "num_tokens": 349278670.0, + "step": 13498 + }, + { + "epoch": 1.4824291675818142, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.30884051322937, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7066829204559326, + "num_tokens": 349304750.0, + "step": 13499 + }, + { + "epoch": 1.4825389852844277, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.1920628547668457, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7350825667381287, + "num_tokens": 349332424.0, + "step": 13500 + }, + { + "epoch": 1.4826488029870415, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.216733694076538, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7062472105026245, + "num_tokens": 349359251.0, + "step": 13501 + }, + { + "epoch": 1.4827586206896552, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0848093032836914, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7016894221305847, + "num_tokens": 349390802.0, + "step": 13502 + }, + { + "epoch": 1.4828684383922688, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.429147958755493, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7375432848930359, + "num_tokens": 349411462.0, + "step": 13503 + }, + { + "epoch": 1.4829782560948825, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.091137409210205, + "learning_rate": 1e-06, + "loss": 0.7861, + "mean_token_accuracy": 0.7507491707801819, + "num_tokens": 349437804.0, + "step": 13504 + }, + { + "epoch": 1.483088073797496, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.3797953128814697, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7266104817390442, + "num_tokens": 349461433.0, + "step": 13505 + }, + { + "epoch": 1.4831978915001098, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.099292039871216, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6949340105056763, + "num_tokens": 349492785.0, + "step": 13506 + }, + { + "epoch": 1.4833077092027236, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.5303566455841064, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7209420204162598, + "num_tokens": 349514255.0, + "step": 13507 + }, + { + "epoch": 1.4834175269053371, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.250077486038208, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7207475900650024, + "num_tokens": 349540250.0, + "step": 13508 + }, + { + "epoch": 1.4835273446079509, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 3.0267274379730225, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7332957983016968, + "num_tokens": 349555718.0, + "step": 13509 + }, + { + "epoch": 1.4836371623105644, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.280148506164551, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7239514589309692, + "num_tokens": 349581064.0, + "step": 13510 + }, + { + "epoch": 1.4837469800131782, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.409358024597168, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.703522801399231, + "num_tokens": 349610750.0, + "step": 13511 + }, + { + "epoch": 1.4838567977157917, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.4860801696777344, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7223556041717529, + "num_tokens": 349631750.0, + "step": 13512 + }, + { + "epoch": 1.4839666154184055, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2114617824554443, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7053661346435547, + "num_tokens": 349660878.0, + "step": 13513 + }, + { + "epoch": 1.484076433121019, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.315538167953491, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7313846945762634, + "num_tokens": 349684386.0, + "step": 13514 + }, + { + "epoch": 1.4841862508236328, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0352985858917236, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6924476623535156, + "num_tokens": 349713907.0, + "step": 13515 + }, + { + "epoch": 1.4842960685262465, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9998741149902344, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7159479856491089, + "num_tokens": 349744778.0, + "step": 13516 + }, + { + "epoch": 1.48440588622886, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.409174919128418, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7527824640274048, + "num_tokens": 349767661.0, + "step": 13517 + }, + { + "epoch": 1.4845157039314738, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.269026279449463, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7059075832366943, + "num_tokens": 349794478.0, + "step": 13518 + }, + { + "epoch": 1.4846255216340873, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.364544630050659, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7074967622756958, + "num_tokens": 349819844.0, + "step": 13519 + }, + { + "epoch": 1.484735339336701, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.248598575592041, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7215579748153687, + "num_tokens": 349847165.0, + "step": 13520 + }, + { + "epoch": 1.4848451570393149, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.1425013542175293, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.701514720916748, + "num_tokens": 349874515.0, + "step": 13521 + }, + { + "epoch": 1.4849549747419284, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2412805557250977, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7224500775337219, + "num_tokens": 349899695.0, + "step": 13522 + }, + { + "epoch": 1.485064792444542, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9871002435684204, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7051859498023987, + "num_tokens": 349930611.0, + "step": 13523 + }, + { + "epoch": 1.4851746101471557, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.169309139251709, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7055187821388245, + "num_tokens": 349958000.0, + "step": 13524 + }, + { + "epoch": 1.4852844278497694, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.158433437347412, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.714126706123352, + "num_tokens": 349984827.0, + "step": 13525 + }, + { + "epoch": 1.485394245552383, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.3794620037078857, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7208634614944458, + "num_tokens": 350008778.0, + "step": 13526 + }, + { + "epoch": 1.4855040632549967, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.205777883529663, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7191076278686523, + "num_tokens": 350034827.0, + "step": 13527 + }, + { + "epoch": 1.4856138809576103, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.120849609375, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7307862639427185, + "num_tokens": 350062547.0, + "step": 13528 + }, + { + "epoch": 1.485723698660224, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2202718257904053, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7225483655929565, + "num_tokens": 350089143.0, + "step": 13529 + }, + { + "epoch": 1.4858335163628378, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.259763479232788, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7063694000244141, + "num_tokens": 350113252.0, + "step": 13530 + }, + { + "epoch": 1.4859433340654513, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.962068796157837, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7155462503433228, + "num_tokens": 350145085.0, + "step": 13531 + }, + { + "epoch": 1.486053151768065, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.4953486919403076, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.729333758354187, + "num_tokens": 350167183.0, + "step": 13532 + }, + { + "epoch": 1.4861629694706786, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.3373072147369385, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7280055284500122, + "num_tokens": 350189984.0, + "step": 13533 + }, + { + "epoch": 1.4862727871732924, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2453861236572266, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7132731676101685, + "num_tokens": 350216411.0, + "step": 13534 + }, + { + "epoch": 1.4863826048759061, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9687119722366333, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7238540053367615, + "num_tokens": 350247603.0, + "step": 13535 + }, + { + "epoch": 1.4864924225785197, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.28338885307312, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.730873167514801, + "num_tokens": 350273919.0, + "step": 13536 + }, + { + "epoch": 1.4866022402811332, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.178271532058716, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7171127200126648, + "num_tokens": 350300025.0, + "step": 13537 + }, + { + "epoch": 1.486712057983747, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.5180747509002686, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7360399961471558, + "num_tokens": 350323919.0, + "step": 13538 + }, + { + "epoch": 1.4868218756863607, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.398859739303589, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.719302773475647, + "num_tokens": 350349168.0, + "step": 13539 + }, + { + "epoch": 1.4869316933889742, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.5672760009765625, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7275269627571106, + "num_tokens": 350370466.0, + "step": 13540 + }, + { + "epoch": 1.487041511091588, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.4114108085632324, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7301281690597534, + "num_tokens": 350392292.0, + "step": 13541 + }, + { + "epoch": 1.4871513287942015, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.290689706802368, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6923952102661133, + "num_tokens": 350418165.0, + "step": 13542 + }, + { + "epoch": 1.4872611464968153, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.4720187187194824, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7191100120544434, + "num_tokens": 350438677.0, + "step": 13543 + }, + { + "epoch": 1.487370964199429, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.343787431716919, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7177380323410034, + "num_tokens": 350463147.0, + "step": 13544 + }, + { + "epoch": 1.4874807819020426, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0189054012298584, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7021231055259705, + "num_tokens": 350493784.0, + "step": 13545 + }, + { + "epoch": 1.4875905996046563, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.519015073776245, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7143675684928894, + "num_tokens": 350515693.0, + "step": 13546 + }, + { + "epoch": 1.4877004173072699, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9821975231170654, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7225143909454346, + "num_tokens": 350549384.0, + "step": 13547 + }, + { + "epoch": 1.4878102350098836, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.1052515506744385, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6957921981811523, + "num_tokens": 350580496.0, + "step": 13548 + }, + { + "epoch": 1.4879200527124974, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.4902498722076416, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7203503847122192, + "num_tokens": 350602962.0, + "step": 13549 + }, + { + "epoch": 1.488029870415111, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.226402997970581, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7154909372329712, + "num_tokens": 350628895.0, + "step": 13550 + }, + { + "epoch": 1.4881396881177245, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.6566948890686035, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7520893812179565, + "num_tokens": 350646645.0, + "step": 13551 + }, + { + "epoch": 1.4882495058203382, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.378453493118286, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7091562151908875, + "num_tokens": 350671701.0, + "step": 13552 + }, + { + "epoch": 1.488359323522952, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.912575364112854, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7100107669830322, + "num_tokens": 350703873.0, + "step": 13553 + }, + { + "epoch": 1.4884691412255655, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0773708820343018, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7126119136810303, + "num_tokens": 350733475.0, + "step": 13554 + }, + { + "epoch": 1.4885789589281793, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.195716381072998, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7447689771652222, + "num_tokens": 350758833.0, + "step": 13555 + }, + { + "epoch": 1.4886887766307928, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.540229320526123, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7191160917282104, + "num_tokens": 350780400.0, + "step": 13556 + }, + { + "epoch": 1.4887985943334066, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0415897369384766, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7361588478088379, + "num_tokens": 350808141.0, + "step": 13557 + }, + { + "epoch": 1.4889084120360203, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.281287670135498, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.72026526927948, + "num_tokens": 350835002.0, + "step": 13558 + }, + { + "epoch": 1.4890182297386338, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.3967361450195312, + "learning_rate": 1e-06, + "loss": 0.7748, + "mean_token_accuracy": 0.7596228122711182, + "num_tokens": 350856990.0, + "step": 13559 + }, + { + "epoch": 1.4891280474412476, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.1559324264526367, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7084539532661438, + "num_tokens": 350886291.0, + "step": 13560 + }, + { + "epoch": 1.4892378651438611, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.0753896236419678, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7179155349731445, + "num_tokens": 350914005.0, + "step": 13561 + }, + { + "epoch": 1.489347682846475, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2602756023406982, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.6937097311019897, + "num_tokens": 350941507.0, + "step": 13562 + }, + { + "epoch": 1.4894575005490884, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.67063307762146, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7361356616020203, + "num_tokens": 350960208.0, + "step": 13563 + }, + { + "epoch": 1.4895673182517022, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.175884246826172, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6913880109786987, + "num_tokens": 350988155.0, + "step": 13564 + }, + { + "epoch": 1.4896771359543157, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.2435083389282227, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7274539470672607, + "num_tokens": 351013036.0, + "step": 13565 + }, + { + "epoch": 1.4897869536569295, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.3720953464508057, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7280372381210327, + "num_tokens": 351036133.0, + "step": 13566 + }, + { + "epoch": 1.4898967713595432, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 1.9756700992584229, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7215620279312134, + "num_tokens": 351069401.0, + "step": 13567 + }, + { + "epoch": 1.4900065890621568, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.017902135848999, + "learning_rate": 1e-06, + "loss": 1.0315, + "mean_token_accuracy": 0.6968569755554199, + "num_tokens": 351102336.0, + "step": 13568 + }, + { + "epoch": 1.4901164067647705, + "ewc_loss": 0.00012159347534179688, + "grad_norm": 2.7517826557159424, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7329995632171631, + "num_tokens": 351120748.0, + "step": 13569 + }, + { + "epoch": 1.490226224467384, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.5917742252349854, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7190117835998535, + "num_tokens": 351140821.0, + "step": 13570 + }, + { + "epoch": 1.4903360421699978, + "ewc_loss": 0.0001220703125, + "grad_norm": 1.954687237739563, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.6977826952934265, + "num_tokens": 351170688.0, + "step": 13571 + }, + { + "epoch": 1.4904458598726116, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2393226623535156, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7131125926971436, + "num_tokens": 351195904.0, + "step": 13572 + }, + { + "epoch": 1.4905556775752251, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4458813667297363, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7208541631698608, + "num_tokens": 351218093.0, + "step": 13573 + }, + { + "epoch": 1.4906654952778389, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.224498987197876, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7097019553184509, + "num_tokens": 351243442.0, + "step": 13574 + }, + { + "epoch": 1.4907753129804524, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.617810010910034, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6990516185760498, + "num_tokens": 351264376.0, + "step": 13575 + }, + { + "epoch": 1.4908851306830662, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.6845972537994385, + "learning_rate": 1e-06, + "loss": 0.7819, + "mean_token_accuracy": 0.751603364944458, + "num_tokens": 351282166.0, + "step": 13576 + }, + { + "epoch": 1.4909949483856797, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1505215167999268, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7039188742637634, + "num_tokens": 351310966.0, + "step": 13577 + }, + { + "epoch": 1.4911047660882935, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2025537490844727, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7292599081993103, + "num_tokens": 351338849.0, + "step": 13578 + }, + { + "epoch": 1.491214583790907, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.309605836868286, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7193197011947632, + "num_tokens": 351364478.0, + "step": 13579 + }, + { + "epoch": 1.4913244014935207, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.319204330444336, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7169497013092041, + "num_tokens": 351389319.0, + "step": 13580 + }, + { + "epoch": 1.4914342191961345, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.207028865814209, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7160506248474121, + "num_tokens": 351413848.0, + "step": 13581 + }, + { + "epoch": 1.491544036898748, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.032787799835205, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6988195776939392, + "num_tokens": 351443166.0, + "step": 13582 + }, + { + "epoch": 1.4916538546013618, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.7226622104644775, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7215373516082764, + "num_tokens": 351462274.0, + "step": 13583 + }, + { + "epoch": 1.4917636723039753, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4145689010620117, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.726143479347229, + "num_tokens": 351487010.0, + "step": 13584 + }, + { + "epoch": 1.491873490006589, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.146404266357422, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7421334981918335, + "num_tokens": 351514148.0, + "step": 13585 + }, + { + "epoch": 1.4919833077092028, + "ewc_loss": 0.0001220703125, + "grad_norm": 1.9776378870010376, + "learning_rate": 1e-06, + "loss": 1.0264, + "mean_token_accuracy": 0.6888492703437805, + "num_tokens": 351548859.0, + "step": 13586 + }, + { + "epoch": 1.4920931254118164, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.5398409366607666, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7475219368934631, + "num_tokens": 351568450.0, + "step": 13587 + }, + { + "epoch": 1.49220294311443, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.42183256149292, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7216856479644775, + "num_tokens": 351590179.0, + "step": 13588 + }, + { + "epoch": 1.4923127608170437, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0590322017669678, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.735380232334137, + "num_tokens": 351617740.0, + "step": 13589 + }, + { + "epoch": 1.4924225785196574, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1213631629943848, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6886340379714966, + "num_tokens": 351649879.0, + "step": 13590 + }, + { + "epoch": 1.492532396222271, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.5371079444885254, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7141044735908508, + "num_tokens": 351671227.0, + "step": 13591 + }, + { + "epoch": 1.4926422139248847, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.486037015914917, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7179219722747803, + "num_tokens": 351693296.0, + "step": 13592 + }, + { + "epoch": 1.4927520316274983, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.6271674633026123, + "learning_rate": 1e-06, + "loss": 0.677, + "mean_token_accuracy": 0.7750267386436462, + "num_tokens": 351711924.0, + "step": 13593 + }, + { + "epoch": 1.492861849330112, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.270660877227783, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7311661243438721, + "num_tokens": 351737262.0, + "step": 13594 + }, + { + "epoch": 1.4929716670327258, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2950897216796875, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7311728596687317, + "num_tokens": 351760464.0, + "step": 13595 + }, + { + "epoch": 1.4930814847353393, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.302189588546753, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.6973403692245483, + "num_tokens": 351784671.0, + "step": 13596 + }, + { + "epoch": 1.493191302437953, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.233022928237915, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7060098648071289, + "num_tokens": 351813130.0, + "step": 13597 + }, + { + "epoch": 1.4933011201405666, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.519754648208618, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7099829912185669, + "num_tokens": 351836354.0, + "step": 13598 + }, + { + "epoch": 1.4934109378431804, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1684110164642334, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7326139807701111, + "num_tokens": 351864756.0, + "step": 13599 + }, + { + "epoch": 1.493520755545794, + "ewc_loss": 0.0001220703125, + "grad_norm": 1.8887492418289185, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7011502385139465, + "num_tokens": 351900574.0, + "step": 13600 + }, + { + "epoch": 1.4936305732484076, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3927221298217773, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7319900393486023, + "num_tokens": 351922710.0, + "step": 13601 + }, + { + "epoch": 1.4937403909510212, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.20682954788208, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7239438891410828, + "num_tokens": 351949225.0, + "step": 13602 + }, + { + "epoch": 1.493850208653635, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.258432626724243, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7130710482597351, + "num_tokens": 351974234.0, + "step": 13603 + }, + { + "epoch": 1.4939600263562487, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.353304624557495, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7120347023010254, + "num_tokens": 351996602.0, + "step": 13604 + }, + { + "epoch": 1.4940698440588622, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3476486206054688, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7117776274681091, + "num_tokens": 352020711.0, + "step": 13605 + }, + { + "epoch": 1.494179661761476, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.086080551147461, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6932641267776489, + "num_tokens": 352052927.0, + "step": 13606 + }, + { + "epoch": 1.4942894794640895, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3754897117614746, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7166152000427246, + "num_tokens": 352077195.0, + "step": 13607 + }, + { + "epoch": 1.4943992971667033, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0278897285461426, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7063064575195312, + "num_tokens": 352106228.0, + "step": 13608 + }, + { + "epoch": 1.494509114869317, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0585203170776367, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7141081094741821, + "num_tokens": 352133965.0, + "step": 13609 + }, + { + "epoch": 1.4946189325719306, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.37092924118042, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7109495997428894, + "num_tokens": 352158760.0, + "step": 13610 + }, + { + "epoch": 1.4947287502745443, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.8578248023986816, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7273865938186646, + "num_tokens": 352177506.0, + "step": 13611 + }, + { + "epoch": 1.4948385679771579, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.6189939975738525, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7222113013267517, + "num_tokens": 352198833.0, + "step": 13612 + }, + { + "epoch": 1.4949483856797716, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3187811374664307, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7399472594261169, + "num_tokens": 352220545.0, + "step": 13613 + }, + { + "epoch": 1.4950582033823854, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0549890995025635, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7097663879394531, + "num_tokens": 352251267.0, + "step": 13614 + }, + { + "epoch": 1.495168021084999, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3026816844940186, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7024909257888794, + "num_tokens": 352277812.0, + "step": 13615 + }, + { + "epoch": 1.4952778387876124, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4069294929504395, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.711426317691803, + "num_tokens": 352302730.0, + "step": 13616 + }, + { + "epoch": 1.4953876564902262, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.5625288486480713, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.708612322807312, + "num_tokens": 352323936.0, + "step": 13617 + }, + { + "epoch": 1.49549747419284, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.170628070831299, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7102189660072327, + "num_tokens": 352353122.0, + "step": 13618 + }, + { + "epoch": 1.4956072918954535, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.355700731277466, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7135821580886841, + "num_tokens": 352379397.0, + "step": 13619 + }, + { + "epoch": 1.4957171095980673, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2147068977355957, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.727292537689209, + "num_tokens": 352405394.0, + "step": 13620 + }, + { + "epoch": 1.4958269273006808, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.376392364501953, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.733945369720459, + "num_tokens": 352427832.0, + "step": 13621 + }, + { + "epoch": 1.4959367450032945, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2546539306640625, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7313041687011719, + "num_tokens": 352451288.0, + "step": 13622 + }, + { + "epoch": 1.4960465627059083, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0485668182373047, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7196959257125854, + "num_tokens": 352480422.0, + "step": 13623 + }, + { + "epoch": 1.4961563804085218, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4891157150268555, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7180163860321045, + "num_tokens": 352502417.0, + "step": 13624 + }, + { + "epoch": 1.4962661981111356, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.037264823913574, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6903336048126221, + "num_tokens": 352534885.0, + "step": 13625 + }, + { + "epoch": 1.4963760158137491, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0329549312591553, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.727118194103241, + "num_tokens": 352564471.0, + "step": 13626 + }, + { + "epoch": 1.4964858335163629, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3793652057647705, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6972494721412659, + "num_tokens": 352586736.0, + "step": 13627 + }, + { + "epoch": 1.4965956512189764, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0715744495391846, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7222449779510498, + "num_tokens": 352615469.0, + "step": 13628 + }, + { + "epoch": 1.4967054689215902, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1127073764801025, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7212128639221191, + "num_tokens": 352643181.0, + "step": 13629 + }, + { + "epoch": 1.4968152866242037, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.115196466445923, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6862915754318237, + "num_tokens": 352673504.0, + "step": 13630 + }, + { + "epoch": 1.4969251043268175, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.0599210262298584, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.728103756904602, + "num_tokens": 352701778.0, + "step": 13631 + }, + { + "epoch": 1.4970349220294312, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2655084133148193, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7176975607872009, + "num_tokens": 352728066.0, + "step": 13632 + }, + { + "epoch": 1.4971447397320448, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2923099994659424, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7108873128890991, + "num_tokens": 352752048.0, + "step": 13633 + }, + { + "epoch": 1.4972545574346585, + "ewc_loss": 0.0001220703125, + "grad_norm": 1.9110329151153564, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6958707571029663, + "num_tokens": 352786096.0, + "step": 13634 + }, + { + "epoch": 1.497364375137272, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2160301208496094, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7365323305130005, + "num_tokens": 352812176.0, + "step": 13635 + }, + { + "epoch": 1.4974741928398858, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3522961139678955, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7196623086929321, + "num_tokens": 352835227.0, + "step": 13636 + }, + { + "epoch": 1.4975840105424996, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.209158182144165, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7219774723052979, + "num_tokens": 352861196.0, + "step": 13637 + }, + { + "epoch": 1.497693828245113, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4089102745056152, + "learning_rate": 1e-06, + "loss": 0.732, + "mean_token_accuracy": 0.763562798500061, + "num_tokens": 352882435.0, + "step": 13638 + }, + { + "epoch": 1.4978036459477266, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.369004249572754, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.6991348266601562, + "num_tokens": 352907753.0, + "step": 13639 + }, + { + "epoch": 1.4979134636503404, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.480463981628418, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.7479133009910583, + "num_tokens": 352929689.0, + "step": 13640 + }, + { + "epoch": 1.4980232813529542, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4135468006134033, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7259931564331055, + "num_tokens": 352950689.0, + "step": 13641 + }, + { + "epoch": 1.4981330990555677, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.135688304901123, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7214472889900208, + "num_tokens": 352978013.0, + "step": 13642 + }, + { + "epoch": 1.4982429167581814, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1183345317840576, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7205548882484436, + "num_tokens": 353007336.0, + "step": 13643 + }, + { + "epoch": 1.498352734460795, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2780027389526367, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7266977429389954, + "num_tokens": 353032226.0, + "step": 13644 + }, + { + "epoch": 1.4984625521634087, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4008798599243164, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7370689511299133, + "num_tokens": 353054202.0, + "step": 13645 + }, + { + "epoch": 1.4985723698660225, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.393366813659668, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7252271771430969, + "num_tokens": 353077624.0, + "step": 13646 + }, + { + "epoch": 1.498682187568636, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.151555299758911, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7218455076217651, + "num_tokens": 353105436.0, + "step": 13647 + }, + { + "epoch": 1.4987920052712498, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1655123233795166, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.7049088478088379, + "num_tokens": 353133795.0, + "step": 13648 + }, + { + "epoch": 1.4989018229738633, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.555997610092163, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7416162490844727, + "num_tokens": 353153980.0, + "step": 13649 + }, + { + "epoch": 1.499011640676477, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2453560829162598, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7248601913452148, + "num_tokens": 353180072.0, + "step": 13650 + }, + { + "epoch": 1.4991214583790908, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2369225025177, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7052861452102661, + "num_tokens": 353209854.0, + "step": 13651 + }, + { + "epoch": 1.4992312760817044, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.8559131622314453, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7285993695259094, + "num_tokens": 353227772.0, + "step": 13652 + }, + { + "epoch": 1.499341093784318, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.486222743988037, + "learning_rate": 1e-06, + "loss": 0.8019, + "mean_token_accuracy": 0.7499216794967651, + "num_tokens": 353248130.0, + "step": 13653 + }, + { + "epoch": 1.4994509114869317, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.7003896236419678, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.716589629650116, + "num_tokens": 353268005.0, + "step": 13654 + }, + { + "epoch": 1.4995607291895454, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.2983322143554688, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7354140877723694, + "num_tokens": 353293531.0, + "step": 13655 + }, + { + "epoch": 1.499670546892159, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.1039822101593018, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.6983045935630798, + "num_tokens": 353323734.0, + "step": 13656 + }, + { + "epoch": 1.4997803645947727, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.4635818004608154, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.7077288627624512, + "num_tokens": 353345463.0, + "step": 13657 + }, + { + "epoch": 1.4998901822973862, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.3566861152648926, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7159260511398315, + "num_tokens": 353369808.0, + "step": 13658 + }, + { + "epoch": 1.5, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.30560302734375, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7473299503326416, + "num_tokens": 353393328.0, + "step": 13659 + }, + { + "epoch": 1.5001098177026138, + "ewc_loss": 0.0001220703125, + "grad_norm": 2.456942558288574, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7208645343780518, + "num_tokens": 353417908.0, + "step": 13660 + }, + { + "epoch": 1.5002196354052273, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2568047046661377, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7181380987167358, + "num_tokens": 353443856.0, + "step": 13661 + }, + { + "epoch": 1.5003294531078408, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1597397327423096, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7232624292373657, + "num_tokens": 353469890.0, + "step": 13662 + }, + { + "epoch": 1.5004392708104546, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2353427410125732, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7017812728881836, + "num_tokens": 353496955.0, + "step": 13663 + }, + { + "epoch": 1.5005490885130683, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0416245460510254, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7079028487205505, + "num_tokens": 353527810.0, + "step": 13664 + }, + { + "epoch": 1.500658906215682, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1992244720458984, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7068453431129456, + "num_tokens": 353554731.0, + "step": 13665 + }, + { + "epoch": 1.5007687239182956, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0281286239624023, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7048807144165039, + "num_tokens": 353586332.0, + "step": 13666 + }, + { + "epoch": 1.5008785416209092, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1618995666503906, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.6997891664505005, + "num_tokens": 353612790.0, + "step": 13667 + }, + { + "epoch": 1.500988359323523, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.08968448638916, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.715699315071106, + "num_tokens": 353640385.0, + "step": 13668 + }, + { + "epoch": 1.5010981770261367, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.349548578262329, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7034127712249756, + "num_tokens": 353665468.0, + "step": 13669 + }, + { + "epoch": 1.5012079947287504, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2210793495178223, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7332830429077148, + "num_tokens": 353690606.0, + "step": 13670 + }, + { + "epoch": 1.501317812431364, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.382939338684082, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7274088859558105, + "num_tokens": 353713806.0, + "step": 13671 + }, + { + "epoch": 1.5014276301339775, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.270372152328491, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.726636528968811, + "num_tokens": 353738924.0, + "step": 13672 + }, + { + "epoch": 1.5015374478365913, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2570297718048096, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7129454612731934, + "num_tokens": 353763454.0, + "step": 13673 + }, + { + "epoch": 1.501647265539205, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.421226978302002, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7398518323898315, + "num_tokens": 353787635.0, + "step": 13674 + }, + { + "epoch": 1.5017570832418186, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4597933292388916, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7166891098022461, + "num_tokens": 353815796.0, + "step": 13675 + }, + { + "epoch": 1.501866900944432, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.238713026046753, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7453466653823853, + "num_tokens": 353840383.0, + "step": 13676 + }, + { + "epoch": 1.5019767186470458, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1919450759887695, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7191561460494995, + "num_tokens": 353866233.0, + "step": 13677 + }, + { + "epoch": 1.5020865363496596, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.083895683288574, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7135995626449585, + "num_tokens": 353894026.0, + "step": 13678 + }, + { + "epoch": 1.5021963540522734, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.048095703125, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6938672065734863, + "num_tokens": 353923991.0, + "step": 13679 + }, + { + "epoch": 1.502306171754887, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.121438980102539, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.728141725063324, + "num_tokens": 353954798.0, + "step": 13680 + }, + { + "epoch": 1.5024159894575004, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.681556224822998, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7250345945358276, + "num_tokens": 353975069.0, + "step": 13681 + }, + { + "epoch": 1.5025258071601142, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.6387851238250732, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7448968291282654, + "num_tokens": 353994106.0, + "step": 13682 + }, + { + "epoch": 1.502635624862728, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.353438377380371, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.6999592781066895, + "num_tokens": 354018331.0, + "step": 13683 + }, + { + "epoch": 1.5027454425653415, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.5095021724700928, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7364825010299683, + "num_tokens": 354040243.0, + "step": 13684 + }, + { + "epoch": 1.5028552602679552, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3521993160247803, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7258985042572021, + "num_tokens": 354064205.0, + "step": 13685 + }, + { + "epoch": 1.5029650779705688, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.396463632583618, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7114559412002563, + "num_tokens": 354089638.0, + "step": 13686 + }, + { + "epoch": 1.5030748956731825, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.53178334236145, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7304043769836426, + "num_tokens": 354113131.0, + "step": 13687 + }, + { + "epoch": 1.5031847133757963, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3661386966705322, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7301324605941772, + "num_tokens": 354137117.0, + "step": 13688 + }, + { + "epoch": 1.5032945310784098, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.341747999191284, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7376570701599121, + "num_tokens": 354161436.0, + "step": 13689 + }, + { + "epoch": 1.5034043487810234, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.295758008956909, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7116678357124329, + "num_tokens": 354186678.0, + "step": 13690 + }, + { + "epoch": 1.5035141664836371, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.042762279510498, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7369856238365173, + "num_tokens": 354214855.0, + "step": 13691 + }, + { + "epoch": 1.5036239841862509, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.079563856124878, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.68116295337677, + "num_tokens": 354247092.0, + "step": 13692 + }, + { + "epoch": 1.5037338018888646, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.491337537765503, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.737683892250061, + "num_tokens": 354268746.0, + "step": 13693 + }, + { + "epoch": 1.5038436195914782, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.135129690170288, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.721552312374115, + "num_tokens": 354296795.0, + "step": 13694 + }, + { + "epoch": 1.5039534372940917, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.070631504058838, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7275720238685608, + "num_tokens": 354325884.0, + "step": 13695 + }, + { + "epoch": 1.5040632549967055, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3623735904693604, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7288707494735718, + "num_tokens": 354348878.0, + "step": 13696 + }, + { + "epoch": 1.5041730726993192, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3026559352874756, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7477423548698425, + "num_tokens": 354373386.0, + "step": 13697 + }, + { + "epoch": 1.5042828904019327, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1512694358825684, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.72129225730896, + "num_tokens": 354399582.0, + "step": 13698 + }, + { + "epoch": 1.5043927081045465, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1823294162750244, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.683566689491272, + "num_tokens": 354427759.0, + "step": 13699 + }, + { + "epoch": 1.50450252580716, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.468755006790161, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6999176740646362, + "num_tokens": 354450286.0, + "step": 13700 + }, + { + "epoch": 1.5046123435097738, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.280054807662964, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.706785261631012, + "num_tokens": 354476107.0, + "step": 13701 + }, + { + "epoch": 1.5047221612123876, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.129467010498047, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6949667930603027, + "num_tokens": 354505040.0, + "step": 13702 + }, + { + "epoch": 1.504831978915001, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1571245193481445, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7146662473678589, + "num_tokens": 354534336.0, + "step": 13703 + }, + { + "epoch": 1.5049417966176146, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.284209966659546, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7172853946685791, + "num_tokens": 354559588.0, + "step": 13704 + }, + { + "epoch": 1.5050516143202284, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3192481994628906, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.71513432264328, + "num_tokens": 354583966.0, + "step": 13705 + }, + { + "epoch": 1.5051614320228421, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.336003065109253, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.721074104309082, + "num_tokens": 354608592.0, + "step": 13706 + }, + { + "epoch": 1.505271249725456, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.435696840286255, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7327286005020142, + "num_tokens": 354631381.0, + "step": 13707 + }, + { + "epoch": 1.5053810674280694, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.29683256149292, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.720191478729248, + "num_tokens": 354655687.0, + "step": 13708 + }, + { + "epoch": 1.505490885130683, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.293123483657837, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7104791402816772, + "num_tokens": 354680390.0, + "step": 13709 + }, + { + "epoch": 1.5056007028332967, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2516770362854004, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.6998977661132812, + "num_tokens": 354706633.0, + "step": 13710 + }, + { + "epoch": 1.5057105205359105, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.411985158920288, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7207332253456116, + "num_tokens": 354730625.0, + "step": 13711 + }, + { + "epoch": 1.505820338238524, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.76530385017395, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7415107488632202, + "num_tokens": 354747197.0, + "step": 13712 + }, + { + "epoch": 1.5059301559411375, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3533294200897217, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.724976658821106, + "num_tokens": 354771713.0, + "step": 13713 + }, + { + "epoch": 1.5060399736437513, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2320568561553955, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7239251732826233, + "num_tokens": 354799653.0, + "step": 13714 + }, + { + "epoch": 1.506149791346365, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1938164234161377, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7077511548995972, + "num_tokens": 354827490.0, + "step": 13715 + }, + { + "epoch": 1.5062596090489788, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3710920810699463, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7102463841438293, + "num_tokens": 354854128.0, + "step": 13716 + }, + { + "epoch": 1.5063694267515924, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3135154247283936, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7144041657447815, + "num_tokens": 354879134.0, + "step": 13717 + }, + { + "epoch": 1.506479244454206, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.5196638107299805, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7133628129959106, + "num_tokens": 354902971.0, + "step": 13718 + }, + { + "epoch": 1.5065890621568196, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.363605260848999, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7297415137290955, + "num_tokens": 354926622.0, + "step": 13719 + }, + { + "epoch": 1.5066988798594334, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.234905242919922, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7218682765960693, + "num_tokens": 354950837.0, + "step": 13720 + }, + { + "epoch": 1.5068086975620472, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2123639583587646, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.6999388337135315, + "num_tokens": 354978412.0, + "step": 13721 + }, + { + "epoch": 1.5069185152646607, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.362283945083618, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7257031202316284, + "num_tokens": 355003622.0, + "step": 13722 + }, + { + "epoch": 1.5070283329672742, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.5985360145568848, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7182047367095947, + "num_tokens": 355023549.0, + "step": 13723 + }, + { + "epoch": 1.507138150669888, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.258073091506958, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7086136341094971, + "num_tokens": 355048551.0, + "step": 13724 + }, + { + "epoch": 1.5072479683725017, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.289980411529541, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7165877819061279, + "num_tokens": 355075336.0, + "step": 13725 + }, + { + "epoch": 1.5073577860751153, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.079271078109741, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.711665689945221, + "num_tokens": 355105548.0, + "step": 13726 + }, + { + "epoch": 1.5074676037777288, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.125791549682617, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7294746041297913, + "num_tokens": 355133482.0, + "step": 13727 + }, + { + "epoch": 1.5075774214803426, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.295100688934326, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7116843461990356, + "num_tokens": 355158234.0, + "step": 13728 + }, + { + "epoch": 1.5076872391829563, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1607251167297363, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7155039310455322, + "num_tokens": 355184906.0, + "step": 13729 + }, + { + "epoch": 1.50779705688557, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.407294273376465, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7003238201141357, + "num_tokens": 355206627.0, + "step": 13730 + }, + { + "epoch": 1.5079068745881836, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9536710977554321, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7113142609596252, + "num_tokens": 355236100.0, + "step": 13731 + }, + { + "epoch": 1.5080166922907972, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9342552423477173, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6914691925048828, + "num_tokens": 355270428.0, + "step": 13732 + }, + { + "epoch": 1.508126509993411, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9975601434707642, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7121625542640686, + "num_tokens": 355303500.0, + "step": 13733 + }, + { + "epoch": 1.5082363276960247, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1336796283721924, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7084428071975708, + "num_tokens": 355333276.0, + "step": 13734 + }, + { + "epoch": 1.5083461453986382, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3575479984283447, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7341986298561096, + "num_tokens": 355356832.0, + "step": 13735 + }, + { + "epoch": 1.508455963101252, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.454387664794922, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7220304012298584, + "num_tokens": 355379912.0, + "step": 13736 + }, + { + "epoch": 1.5085657808038655, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2171552181243896, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7025002241134644, + "num_tokens": 355408113.0, + "step": 13737 + }, + { + "epoch": 1.5086755985064793, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.348846435546875, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7320764064788818, + "num_tokens": 355432286.0, + "step": 13738 + }, + { + "epoch": 1.508785416209093, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.365358829498291, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7009707689285278, + "num_tokens": 355457847.0, + "step": 13739 + }, + { + "epoch": 1.5088952339117065, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3160672187805176, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7275356650352478, + "num_tokens": 355483148.0, + "step": 13740 + }, + { + "epoch": 1.50900505161432, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.123267412185669, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7291547060012817, + "num_tokens": 355510583.0, + "step": 13741 + }, + { + "epoch": 1.5091148693169338, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.332582950592041, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7440392374992371, + "num_tokens": 355535520.0, + "step": 13742 + }, + { + "epoch": 1.5092246870195476, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1295583248138428, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7212828993797302, + "num_tokens": 355562645.0, + "step": 13743 + }, + { + "epoch": 1.5093345047221614, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1381583213806152, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7075297236442566, + "num_tokens": 355590352.0, + "step": 13744 + }, + { + "epoch": 1.5094443224247749, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3282735347747803, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7036482095718384, + "num_tokens": 355616447.0, + "step": 13745 + }, + { + "epoch": 1.5095541401273884, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.200350522994995, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7072995901107788, + "num_tokens": 355643921.0, + "step": 13746 + }, + { + "epoch": 1.5096639578300022, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.5262928009033203, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7362689971923828, + "num_tokens": 355663688.0, + "step": 13747 + }, + { + "epoch": 1.509773775532616, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.7296981811523438, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7181999087333679, + "num_tokens": 355684287.0, + "step": 13748 + }, + { + "epoch": 1.5098835932352295, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0810904502868652, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7058078646659851, + "num_tokens": 355714256.0, + "step": 13749 + }, + { + "epoch": 1.5099934109378432, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1624906063079834, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7138482332229614, + "num_tokens": 355741462.0, + "step": 13750 + }, + { + "epoch": 1.5101032286404568, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3653321266174316, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7435548305511475, + "num_tokens": 355762296.0, + "step": 13751 + }, + { + "epoch": 1.5102130463430705, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0805864334106445, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.699104905128479, + "num_tokens": 355792887.0, + "step": 13752 + }, + { + "epoch": 1.5103228640456843, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.168956995010376, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7126328945159912, + "num_tokens": 355821379.0, + "step": 13753 + }, + { + "epoch": 1.5104326817482978, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.430947780609131, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7260087132453918, + "num_tokens": 355843447.0, + "step": 13754 + }, + { + "epoch": 1.5105424994509113, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.392589569091797, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7124138474464417, + "num_tokens": 355867998.0, + "step": 13755 + }, + { + "epoch": 1.510652317153525, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.398630380630493, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7374387979507446, + "num_tokens": 355888455.0, + "step": 13756 + }, + { + "epoch": 1.5107621348561389, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1814515590667725, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.739560604095459, + "num_tokens": 355913597.0, + "step": 13757 + }, + { + "epoch": 1.5108719525587526, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.20648193359375, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7145462036132812, + "num_tokens": 355939107.0, + "step": 13758 + }, + { + "epoch": 1.5109817702613662, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9671716690063477, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7183141708374023, + "num_tokens": 355969652.0, + "step": 13759 + }, + { + "epoch": 1.5110915879639797, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.297790288925171, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7389395833015442, + "num_tokens": 355995923.0, + "step": 13760 + }, + { + "epoch": 1.5112014056665934, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2738277912139893, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7171806693077087, + "num_tokens": 356020220.0, + "step": 13761 + }, + { + "epoch": 1.5113112233692072, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1364927291870117, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7084234952926636, + "num_tokens": 356047911.0, + "step": 13762 + }, + { + "epoch": 1.5114210410718207, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1500191688537598, + "learning_rate": 1e-06, + "loss": 1.0433, + "mean_token_accuracy": 0.6886389255523682, + "num_tokens": 356074462.0, + "step": 13763 + }, + { + "epoch": 1.5115308587744345, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1134774684906006, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7309149503707886, + "num_tokens": 356101105.0, + "step": 13764 + }, + { + "epoch": 1.511640676477048, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2480218410491943, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7004655003547668, + "num_tokens": 356128051.0, + "step": 13765 + }, + { + "epoch": 1.5117504941796618, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4085118770599365, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7274610996246338, + "num_tokens": 356151505.0, + "step": 13766 + }, + { + "epoch": 1.5118603118822755, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.356203079223633, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7218899726867676, + "num_tokens": 356177332.0, + "step": 13767 + }, + { + "epoch": 1.511970129584889, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0284829139709473, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7474396228790283, + "num_tokens": 356205591.0, + "step": 13768 + }, + { + "epoch": 1.5120799472875026, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4166057109832764, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7458377480506897, + "num_tokens": 356226094.0, + "step": 13769 + }, + { + "epoch": 1.5121897649901164, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2484896183013916, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6947146058082581, + "num_tokens": 356252556.0, + "step": 13770 + }, + { + "epoch": 1.5122995826927301, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2516844272613525, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7391142845153809, + "num_tokens": 356276592.0, + "step": 13771 + }, + { + "epoch": 1.5124094003953439, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4425816535949707, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7228114604949951, + "num_tokens": 356298457.0, + "step": 13772 + }, + { + "epoch": 1.5125192180979574, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2311923503875732, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7122716903686523, + "num_tokens": 356324337.0, + "step": 13773 + }, + { + "epoch": 1.512629035800571, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4687726497650146, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7259543538093567, + "num_tokens": 356345853.0, + "step": 13774 + }, + { + "epoch": 1.5127388535031847, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0827860832214355, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.684192955493927, + "num_tokens": 356379117.0, + "step": 13775 + }, + { + "epoch": 1.5128486712057985, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3124606609344482, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7173376083374023, + "num_tokens": 356405071.0, + "step": 13776 + }, + { + "epoch": 1.512958488908412, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.48710560798645, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7231982946395874, + "num_tokens": 356425984.0, + "step": 13777 + }, + { + "epoch": 1.5130683066110255, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3010752201080322, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7253957986831665, + "num_tokens": 356448982.0, + "step": 13778 + }, + { + "epoch": 1.5131781243136393, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.554152011871338, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7312495112419128, + "num_tokens": 356468840.0, + "step": 13779 + }, + { + "epoch": 1.513287942016253, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1661384105682373, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6959711909294128, + "num_tokens": 356496916.0, + "step": 13780 + }, + { + "epoch": 1.5133977597188668, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.377061605453491, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7397701740264893, + "num_tokens": 356519639.0, + "step": 13781 + }, + { + "epoch": 1.5135075774214803, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4077491760253906, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7116489410400391, + "num_tokens": 356543945.0, + "step": 13782 + }, + { + "epoch": 1.5136173951240939, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.177931785583496, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7118222117424011, + "num_tokens": 356571028.0, + "step": 13783 + }, + { + "epoch": 1.5137272128267076, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.133164405822754, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.721113920211792, + "num_tokens": 356600256.0, + "step": 13784 + }, + { + "epoch": 1.5138370305293214, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.344175338745117, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.7058243751525879, + "num_tokens": 356625344.0, + "step": 13785 + }, + { + "epoch": 1.5139468482319351, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1769182682037354, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7273889183998108, + "num_tokens": 356650057.0, + "step": 13786 + }, + { + "epoch": 1.5140566659345487, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.277078151702881, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7208813428878784, + "num_tokens": 356675194.0, + "step": 13787 + }, + { + "epoch": 1.5141664836371622, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4065520763397217, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7334415912628174, + "num_tokens": 356698600.0, + "step": 13788 + }, + { + "epoch": 1.514276301339776, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0670387744903564, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7047032117843628, + "num_tokens": 356729849.0, + "step": 13789 + }, + { + "epoch": 1.5143861190423897, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.317894458770752, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7259370684623718, + "num_tokens": 356755212.0, + "step": 13790 + }, + { + "epoch": 1.5144959367450033, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9886410236358643, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.7021254301071167, + "num_tokens": 356786954.0, + "step": 13791 + }, + { + "epoch": 1.5146057544476168, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2875702381134033, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7047373056411743, + "num_tokens": 356811653.0, + "step": 13792 + }, + { + "epoch": 1.5147155721502306, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.9961401224136353, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7150070667266846, + "num_tokens": 356841369.0, + "step": 13793 + }, + { + "epoch": 1.5148253898528443, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.224052667617798, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7081133127212524, + "num_tokens": 356870485.0, + "step": 13794 + }, + { + "epoch": 1.514935207555458, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 1.936877727508545, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7033857107162476, + "num_tokens": 356904064.0, + "step": 13795 + }, + { + "epoch": 1.5150450252580716, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.467379331588745, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7360026836395264, + "num_tokens": 356924167.0, + "step": 13796 + }, + { + "epoch": 1.5151548429606851, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.568511486053467, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7457308769226074, + "num_tokens": 356944442.0, + "step": 13797 + }, + { + "epoch": 1.515264660663299, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.669910430908203, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7151144742965698, + "num_tokens": 356963483.0, + "step": 13798 + }, + { + "epoch": 1.5153744783659127, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.41275691986084, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7049219012260437, + "num_tokens": 356986322.0, + "step": 13799 + }, + { + "epoch": 1.5154842960685262, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.059326648712158, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6899241209030151, + "num_tokens": 357019143.0, + "step": 13800 + }, + { + "epoch": 1.51559411377114, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4345362186431885, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7257341742515564, + "num_tokens": 357041884.0, + "step": 13801 + }, + { + "epoch": 1.5157039314737535, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.092905282974243, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7041804790496826, + "num_tokens": 357072800.0, + "step": 13802 + }, + { + "epoch": 1.5158137491763672, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2165136337280273, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.722868800163269, + "num_tokens": 357096559.0, + "step": 13803 + }, + { + "epoch": 1.515923566878981, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.432143211364746, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.699564516544342, + "num_tokens": 357118220.0, + "step": 13804 + }, + { + "epoch": 1.5160333845815945, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1012539863586426, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6832351684570312, + "num_tokens": 357150262.0, + "step": 13805 + }, + { + "epoch": 1.516143202284208, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.216958999633789, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7277429103851318, + "num_tokens": 357176008.0, + "step": 13806 + }, + { + "epoch": 1.5162530199868218, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.257145881652832, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7077275514602661, + "num_tokens": 357202813.0, + "step": 13807 + }, + { + "epoch": 1.5163628376894356, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2018096446990967, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7230198383331299, + "num_tokens": 357229265.0, + "step": 13808 + }, + { + "epoch": 1.5164726553920493, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1414661407470703, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7340645790100098, + "num_tokens": 357256720.0, + "step": 13809 + }, + { + "epoch": 1.5165824730946629, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.235402822494507, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7181358933448792, + "num_tokens": 357281438.0, + "step": 13810 + }, + { + "epoch": 1.5166922907972764, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0359714031219482, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7202389240264893, + "num_tokens": 357309993.0, + "step": 13811 + }, + { + "epoch": 1.5168021084998902, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.156304121017456, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7047178745269775, + "num_tokens": 357338261.0, + "step": 13812 + }, + { + "epoch": 1.516911926202504, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.235852003097534, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7218051552772522, + "num_tokens": 357364164.0, + "step": 13813 + }, + { + "epoch": 1.5170217439051175, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1831419467926025, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7325936555862427, + "num_tokens": 357390889.0, + "step": 13814 + }, + { + "epoch": 1.5171315616077312, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2084696292877197, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7116338014602661, + "num_tokens": 357417326.0, + "step": 13815 + }, + { + "epoch": 1.5172413793103448, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.6466798782348633, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7307562828063965, + "num_tokens": 357437130.0, + "step": 13816 + }, + { + "epoch": 1.5173511970129585, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2777132987976074, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7227200269699097, + "num_tokens": 357462575.0, + "step": 13817 + }, + { + "epoch": 1.5174610147155723, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3197176456451416, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7353401184082031, + "num_tokens": 357485898.0, + "step": 13818 + }, + { + "epoch": 1.5175708324181858, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.477187156677246, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7205346822738647, + "num_tokens": 357508517.0, + "step": 13819 + }, + { + "epoch": 1.5176806501207993, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.192561626434326, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7299622297286987, + "num_tokens": 357535443.0, + "step": 13820 + }, + { + "epoch": 1.517790467823413, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0806164741516113, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.714890718460083, + "num_tokens": 357565210.0, + "step": 13821 + }, + { + "epoch": 1.5179002855260268, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3082025051116943, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7104200720787048, + "num_tokens": 357589676.0, + "step": 13822 + }, + { + "epoch": 1.5180101032286406, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.076524496078491, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7376682758331299, + "num_tokens": 357616481.0, + "step": 13823 + }, + { + "epoch": 1.5181199209312541, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.579166889190674, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7580642700195312, + "num_tokens": 357637844.0, + "step": 13824 + }, + { + "epoch": 1.5182297386338677, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.085411548614502, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7168241739273071, + "num_tokens": 357665617.0, + "step": 13825 + }, + { + "epoch": 1.5183395563364814, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.0816028118133545, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7176580429077148, + "num_tokens": 357693346.0, + "step": 13826 + }, + { + "epoch": 1.5184493740390952, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3470590114593506, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7184127569198608, + "num_tokens": 357716822.0, + "step": 13827 + }, + { + "epoch": 1.5185591917417087, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.1728720664978027, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7222268581390381, + "num_tokens": 357743016.0, + "step": 13828 + }, + { + "epoch": 1.5186690094443223, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.3587679862976074, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7118481993675232, + "num_tokens": 357767385.0, + "step": 13829 + }, + { + "epoch": 1.518778827146936, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.2812352180480957, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7084118723869324, + "num_tokens": 357794899.0, + "step": 13830 + }, + { + "epoch": 1.5188886448495498, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.360657215118408, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7343854308128357, + "num_tokens": 357817416.0, + "step": 13831 + }, + { + "epoch": 1.5189984625521635, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.297710418701172, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7079316973686218, + "num_tokens": 357843749.0, + "step": 13832 + }, + { + "epoch": 1.519108280254777, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4053003787994385, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7273828983306885, + "num_tokens": 357867032.0, + "step": 13833 + }, + { + "epoch": 1.5192180979573906, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.4095191955566406, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.7487690448760986, + "num_tokens": 357888379.0, + "step": 13834 + }, + { + "epoch": 1.5193279156600044, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.317403793334961, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7154836058616638, + "num_tokens": 357911911.0, + "step": 13835 + }, + { + "epoch": 1.5194377333626181, + "ewc_loss": 0.00012302398681640625, + "grad_norm": 2.702897548675537, + "learning_rate": 1e-06, + "loss": 0.7529, + "mean_token_accuracy": 0.7585389614105225, + "num_tokens": 357929136.0, + "step": 13836 + }, + { + "epoch": 1.5195475510652319, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.170933246612549, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.721477746963501, + "num_tokens": 357955890.0, + "step": 13837 + }, + { + "epoch": 1.5196573687678454, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4183201789855957, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7254782915115356, + "num_tokens": 357978799.0, + "step": 13838 + }, + { + "epoch": 1.519767186470459, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.775082588195801, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7209717035293579, + "num_tokens": 357997097.0, + "step": 13839 + }, + { + "epoch": 1.5198770041730727, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1344075202941895, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7146610021591187, + "num_tokens": 358023895.0, + "step": 13840 + }, + { + "epoch": 1.5199868218756865, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0448122024536133, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7069358825683594, + "num_tokens": 358055221.0, + "step": 13841 + }, + { + "epoch": 1.5200966395783, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9581260681152344, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7253687381744385, + "num_tokens": 358086249.0, + "step": 13842 + }, + { + "epoch": 1.5202064572809135, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.376216173171997, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7328715324401855, + "num_tokens": 358108163.0, + "step": 13843 + }, + { + "epoch": 1.5203162749835273, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.179447889328003, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7098802328109741, + "num_tokens": 358136746.0, + "step": 13844 + }, + { + "epoch": 1.520426092686141, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2813868522644043, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7165918350219727, + "num_tokens": 358161247.0, + "step": 13845 + }, + { + "epoch": 1.5205359103887548, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3482766151428223, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7078822255134583, + "num_tokens": 358185895.0, + "step": 13846 + }, + { + "epoch": 1.5206457280913683, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.375016689300537, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7355724573135376, + "num_tokens": 358207700.0, + "step": 13847 + }, + { + "epoch": 1.5207555457939819, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.44323992729187, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7290993928909302, + "num_tokens": 358229003.0, + "step": 13848 + }, + { + "epoch": 1.5208653634965956, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1575164794921875, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7254102230072021, + "num_tokens": 358254281.0, + "step": 13849 + }, + { + "epoch": 1.5209751811992094, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3433756828308105, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6997121572494507, + "num_tokens": 358280670.0, + "step": 13850 + }, + { + "epoch": 1.5210849989018231, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0297017097473145, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6990983486175537, + "num_tokens": 358310030.0, + "step": 13851 + }, + { + "epoch": 1.5211948166044367, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3770346641540527, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7056866884231567, + "num_tokens": 358335096.0, + "step": 13852 + }, + { + "epoch": 1.5213046343070502, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.359260320663452, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7278245091438293, + "num_tokens": 358356406.0, + "step": 13853 + }, + { + "epoch": 1.521414452009664, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5359067916870117, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.732611894607544, + "num_tokens": 358377855.0, + "step": 13854 + }, + { + "epoch": 1.5215242697122777, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.211472988128662, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7403603196144104, + "num_tokens": 358405613.0, + "step": 13855 + }, + { + "epoch": 1.5216340874148913, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0521163940429688, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6861467361450195, + "num_tokens": 358434728.0, + "step": 13856 + }, + { + "epoch": 1.5217439051175048, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.343034267425537, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7254526615142822, + "num_tokens": 358458558.0, + "step": 13857 + }, + { + "epoch": 1.5218537228201185, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.676146984100342, + "learning_rate": 1e-06, + "loss": 0.7646, + "mean_token_accuracy": 0.7575079798698425, + "num_tokens": 358475837.0, + "step": 13858 + }, + { + "epoch": 1.5219635405227323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.366209030151367, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7328908443450928, + "num_tokens": 358500531.0, + "step": 13859 + }, + { + "epoch": 1.522073358225346, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7061245441436768, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7365508675575256, + "num_tokens": 358519057.0, + "step": 13860 + }, + { + "epoch": 1.5221831759279596, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2950737476348877, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7204005122184753, + "num_tokens": 358543699.0, + "step": 13861 + }, + { + "epoch": 1.5222929936305731, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5592360496520996, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7276567220687866, + "num_tokens": 358564328.0, + "step": 13862 + }, + { + "epoch": 1.522402811333187, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.274394989013672, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7395534515380859, + "num_tokens": 358587141.0, + "step": 13863 + }, + { + "epoch": 1.5225126290358006, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0394256114959717, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7258708477020264, + "num_tokens": 358616127.0, + "step": 13864 + }, + { + "epoch": 1.5226224467384142, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4212863445281982, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7136998772621155, + "num_tokens": 358639200.0, + "step": 13865 + }, + { + "epoch": 1.522732264441028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.408003568649292, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7134488224983215, + "num_tokens": 358666300.0, + "step": 13866 + }, + { + "epoch": 1.5228420821436415, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0703725814819336, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7052838206291199, + "num_tokens": 358696624.0, + "step": 13867 + }, + { + "epoch": 1.5229518998462552, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0740256309509277, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.709520697593689, + "num_tokens": 358725844.0, + "step": 13868 + }, + { + "epoch": 1.523061717548869, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.48791766166687, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7136346101760864, + "num_tokens": 358746996.0, + "step": 13869 + }, + { + "epoch": 1.5231715352514825, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2213902473449707, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7275229692459106, + "num_tokens": 358772315.0, + "step": 13870 + }, + { + "epoch": 1.523281352954096, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1809728145599365, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.719752848148346, + "num_tokens": 358798304.0, + "step": 13871 + }, + { + "epoch": 1.5233911706567098, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.204836130142212, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7211154699325562, + "num_tokens": 358823430.0, + "step": 13872 + }, + { + "epoch": 1.5235009883593236, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 3.166086196899414, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7080029249191284, + "num_tokens": 358845496.0, + "step": 13873 + }, + { + "epoch": 1.5236108060619373, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.067249059677124, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7263126373291016, + "num_tokens": 358874770.0, + "step": 13874 + }, + { + "epoch": 1.5237206237645509, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.251251459121704, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7260470390319824, + "num_tokens": 358898948.0, + "step": 13875 + }, + { + "epoch": 1.5238304414671644, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1685421466827393, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7395400404930115, + "num_tokens": 358925509.0, + "step": 13876 + }, + { + "epoch": 1.5239402591697782, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.020878791809082, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7143278121948242, + "num_tokens": 358956547.0, + "step": 13877 + }, + { + "epoch": 1.524050076872392, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.055817127227783, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.722863495349884, + "num_tokens": 358985992.0, + "step": 13878 + }, + { + "epoch": 1.5241598945750054, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2382726669311523, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.717367947101593, + "num_tokens": 359011171.0, + "step": 13879 + }, + { + "epoch": 1.5242697122776192, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.749203681945801, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.7488195896148682, + "num_tokens": 359029728.0, + "step": 13880 + }, + { + "epoch": 1.5243795299802327, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.491086959838867, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7083348631858826, + "num_tokens": 359051188.0, + "step": 13881 + }, + { + "epoch": 1.5244893476828465, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.230175018310547, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7211447954177856, + "num_tokens": 359075009.0, + "step": 13882 + }, + { + "epoch": 1.5245991653854603, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.506998300552368, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7423363924026489, + "num_tokens": 359096328.0, + "step": 13883 + }, + { + "epoch": 1.5247089830880738, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4565699100494385, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7321771383285522, + "num_tokens": 359120262.0, + "step": 13884 + }, + { + "epoch": 1.5248188007906873, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.480045795440674, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7258263826370239, + "num_tokens": 359140696.0, + "step": 13885 + }, + { + "epoch": 1.524928618493301, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3409669399261475, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7103564739227295, + "num_tokens": 359164468.0, + "step": 13886 + }, + { + "epoch": 1.5250384361959148, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.040166139602661, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7031921744346619, + "num_tokens": 359195179.0, + "step": 13887 + }, + { + "epoch": 1.5251482538985286, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.253715753555298, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6958332061767578, + "num_tokens": 359222376.0, + "step": 13888 + }, + { + "epoch": 1.5252580716011421, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1940758228302, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6960626840591431, + "num_tokens": 359251589.0, + "step": 13889 + }, + { + "epoch": 1.5253678893037557, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0051441192626953, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7048995494842529, + "num_tokens": 359283194.0, + "step": 13890 + }, + { + "epoch": 1.5254777070063694, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.323923110961914, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7325165271759033, + "num_tokens": 359306658.0, + "step": 13891 + }, + { + "epoch": 1.5255875247089832, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.04567813873291, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6976137757301331, + "num_tokens": 359338558.0, + "step": 13892 + }, + { + "epoch": 1.5256973424115967, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1277053356170654, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.72069251537323, + "num_tokens": 359365423.0, + "step": 13893 + }, + { + "epoch": 1.5258071601142102, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5860238075256348, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.703850269317627, + "num_tokens": 359387069.0, + "step": 13894 + }, + { + "epoch": 1.525916977816824, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.273188829421997, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7145034074783325, + "num_tokens": 359412170.0, + "step": 13895 + }, + { + "epoch": 1.5260267955194378, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.103724956512451, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.6964911818504333, + "num_tokens": 359441555.0, + "step": 13896 + }, + { + "epoch": 1.5261366132220515, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.482642412185669, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.746548056602478, + "num_tokens": 359463418.0, + "step": 13897 + }, + { + "epoch": 1.526246430924665, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0771961212158203, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7257017493247986, + "num_tokens": 359493384.0, + "step": 13898 + }, + { + "epoch": 1.5263562486272786, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9913150072097778, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7243342399597168, + "num_tokens": 359525207.0, + "step": 13899 + }, + { + "epoch": 1.5264660663298923, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.129185676574707, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.718489408493042, + "num_tokens": 359554399.0, + "step": 13900 + }, + { + "epoch": 1.526575884032506, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5052759647369385, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7276319265365601, + "num_tokens": 359574869.0, + "step": 13901 + }, + { + "epoch": 1.5266857017351199, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.432271957397461, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7173879742622375, + "num_tokens": 359598923.0, + "step": 13902 + }, + { + "epoch": 1.5267955194377334, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0531179904937744, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7051438093185425, + "num_tokens": 359633756.0, + "step": 13903 + }, + { + "epoch": 1.526905337140347, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.494873046875, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7231582403182983, + "num_tokens": 359655317.0, + "step": 13904 + }, + { + "epoch": 1.5270151548429607, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.378119945526123, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6936955451965332, + "num_tokens": 359679187.0, + "step": 13905 + }, + { + "epoch": 1.5271249725455744, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4359359741210938, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.723787784576416, + "num_tokens": 359701225.0, + "step": 13906 + }, + { + "epoch": 1.527234790248188, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1168744564056396, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7251416444778442, + "num_tokens": 359731034.0, + "step": 13907 + }, + { + "epoch": 1.5273446079508015, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.151099443435669, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7044166326522827, + "num_tokens": 359759267.0, + "step": 13908 + }, + { + "epoch": 1.5274544256534153, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2959113121032715, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7233763933181763, + "num_tokens": 359783366.0, + "step": 13909 + }, + { + "epoch": 1.527564243356029, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.083296060562134, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6992969512939453, + "num_tokens": 359813623.0, + "step": 13910 + }, + { + "epoch": 1.5276740610586428, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.128342628479004, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.736782968044281, + "num_tokens": 359839622.0, + "step": 13911 + }, + { + "epoch": 1.5277838787612563, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5634167194366455, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7193591594696045, + "num_tokens": 359861511.0, + "step": 13912 + }, + { + "epoch": 1.5278936964638699, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.581467390060425, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7097028493881226, + "num_tokens": 359882319.0, + "step": 13913 + }, + { + "epoch": 1.5280035141664836, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2175700664520264, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7146254181861877, + "num_tokens": 359907346.0, + "step": 13914 + }, + { + "epoch": 1.5281133318690974, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.08754301071167, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.726418137550354, + "num_tokens": 359937308.0, + "step": 13915 + }, + { + "epoch": 1.5282231495717111, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9460402727127075, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7215548753738403, + "num_tokens": 359969915.0, + "step": 13916 + }, + { + "epoch": 1.5283329672743247, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.373516082763672, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.722655713558197, + "num_tokens": 359992266.0, + "step": 13917 + }, + { + "epoch": 1.5284427849769382, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4390296936035156, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7310330867767334, + "num_tokens": 360014096.0, + "step": 13918 + }, + { + "epoch": 1.528552602679552, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0496294498443604, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.6984249949455261, + "num_tokens": 360044076.0, + "step": 13919 + }, + { + "epoch": 1.5286624203821657, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.017928123474121, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7196210622787476, + "num_tokens": 360073970.0, + "step": 13920 + }, + { + "epoch": 1.5287722380847792, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1820287704467773, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7236813902854919, + "num_tokens": 360099803.0, + "step": 13921 + }, + { + "epoch": 1.5288820557873928, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2192513942718506, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7170591354370117, + "num_tokens": 360124665.0, + "step": 13922 + }, + { + "epoch": 1.5289918734900065, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0533487796783447, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6977694034576416, + "num_tokens": 360154624.0, + "step": 13923 + }, + { + "epoch": 1.5291016911926203, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.164395570755005, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7258865833282471, + "num_tokens": 360181063.0, + "step": 13924 + }, + { + "epoch": 1.529211508895234, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.023585796356201, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.71095871925354, + "num_tokens": 360209406.0, + "step": 13925 + }, + { + "epoch": 1.5293213265978476, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.432614326477051, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7364215850830078, + "num_tokens": 360231664.0, + "step": 13926 + }, + { + "epoch": 1.5294311443004611, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1276471614837646, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.723829984664917, + "num_tokens": 360259971.0, + "step": 13927 + }, + { + "epoch": 1.5295409620030749, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3531742095947266, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7317968606948853, + "num_tokens": 360283538.0, + "step": 13928 + }, + { + "epoch": 1.5296507797056886, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1103639602661133, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7013444900512695, + "num_tokens": 360311061.0, + "step": 13929 + }, + { + "epoch": 1.5297605974083022, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.8625717163085938, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7393258213996887, + "num_tokens": 360327539.0, + "step": 13930 + }, + { + "epoch": 1.529870415110916, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.276510238647461, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.714607834815979, + "num_tokens": 360352858.0, + "step": 13931 + }, + { + "epoch": 1.5299802328135295, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1246109008789062, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7114728689193726, + "num_tokens": 360383213.0, + "step": 13932 + }, + { + "epoch": 1.5300900505161432, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.213838815689087, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7093456387519836, + "num_tokens": 360411710.0, + "step": 13933 + }, + { + "epoch": 1.530199868218757, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4117236137390137, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7094914317131042, + "num_tokens": 360434777.0, + "step": 13934 + }, + { + "epoch": 1.5303096859213705, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9911233186721802, + "learning_rate": 1e-06, + "loss": 1.0151, + "mean_token_accuracy": 0.6899064779281616, + "num_tokens": 360467550.0, + "step": 13935 + }, + { + "epoch": 1.530419503623984, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2392685413360596, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7035107016563416, + "num_tokens": 360494011.0, + "step": 13936 + }, + { + "epoch": 1.5305293213265978, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1808078289031982, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7191692590713501, + "num_tokens": 360521569.0, + "step": 13937 + }, + { + "epoch": 1.5306391390292116, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.277170181274414, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7211310863494873, + "num_tokens": 360548911.0, + "step": 13938 + }, + { + "epoch": 1.5307489567318253, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2758309841156006, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7091348171234131, + "num_tokens": 360574672.0, + "step": 13939 + }, + { + "epoch": 1.5308587744344389, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.30155873298645, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7222340106964111, + "num_tokens": 360599959.0, + "step": 13940 + }, + { + "epoch": 1.5309685921370524, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.272686004638672, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7189158201217651, + "num_tokens": 360626176.0, + "step": 13941 + }, + { + "epoch": 1.5310784098396661, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.234968900680542, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7268757224082947, + "num_tokens": 360650766.0, + "step": 13942 + }, + { + "epoch": 1.53118822754228, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3809382915496826, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7081539630889893, + "num_tokens": 360674516.0, + "step": 13943 + }, + { + "epoch": 1.5312980452448934, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.022406578063965, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7080153822898865, + "num_tokens": 360702725.0, + "step": 13944 + }, + { + "epoch": 1.5314078629475072, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3029096126556396, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7069677114486694, + "num_tokens": 360727098.0, + "step": 13945 + }, + { + "epoch": 1.5315176806501207, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2249181270599365, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7194864749908447, + "num_tokens": 360755180.0, + "step": 13946 + }, + { + "epoch": 1.5316274983527345, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.314075469970703, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7341619729995728, + "num_tokens": 360779517.0, + "step": 13947 + }, + { + "epoch": 1.5317373160553482, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4465901851654053, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.718863844871521, + "num_tokens": 360799708.0, + "step": 13948 + }, + { + "epoch": 1.5318471337579618, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.284508466720581, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7034915089607239, + "num_tokens": 360825254.0, + "step": 13949 + }, + { + "epoch": 1.5319569514605753, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.575335741043091, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7161763906478882, + "num_tokens": 360844500.0, + "step": 13950 + }, + { + "epoch": 1.532066769163189, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6226117610931396, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7143154740333557, + "num_tokens": 360864445.0, + "step": 13951 + }, + { + "epoch": 1.5321765868658028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3640830516815186, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7420382499694824, + "num_tokens": 360887670.0, + "step": 13952 + }, + { + "epoch": 1.5322864045684166, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.309242010116577, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7379545569419861, + "num_tokens": 360911609.0, + "step": 13953 + }, + { + "epoch": 1.5323962222710301, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.333345413208008, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7184874415397644, + "num_tokens": 360935516.0, + "step": 13954 + }, + { + "epoch": 1.5325060399736437, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.539839267730713, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7133536338806152, + "num_tokens": 360957172.0, + "step": 13955 + }, + { + "epoch": 1.5326158576762574, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.131169557571411, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7419623732566833, + "num_tokens": 360984712.0, + "step": 13956 + }, + { + "epoch": 1.5327256753788712, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2056190967559814, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.699063241481781, + "num_tokens": 361013040.0, + "step": 13957 + }, + { + "epoch": 1.5328354930814847, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0603437423706055, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7013689279556274, + "num_tokens": 361045049.0, + "step": 13958 + }, + { + "epoch": 1.5329453107840982, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.216123104095459, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6939084529876709, + "num_tokens": 361074053.0, + "step": 13959 + }, + { + "epoch": 1.533055128486712, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1852691173553467, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6880874633789062, + "num_tokens": 361104096.0, + "step": 13960 + }, + { + "epoch": 1.5331649461893258, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3247478008270264, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7238544225692749, + "num_tokens": 361127413.0, + "step": 13961 + }, + { + "epoch": 1.5332747638919395, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.041045665740967, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7103819847106934, + "num_tokens": 361156264.0, + "step": 13962 + }, + { + "epoch": 1.533384581594553, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4400463104248047, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7165815830230713, + "num_tokens": 361179506.0, + "step": 13963 + }, + { + "epoch": 1.5334943992971666, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0151851177215576, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7111002206802368, + "num_tokens": 361208503.0, + "step": 13964 + }, + { + "epoch": 1.5336042169997803, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.094334363937378, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7088756561279297, + "num_tokens": 361236807.0, + "step": 13965 + }, + { + "epoch": 1.533714034702394, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.162513017654419, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7431601285934448, + "num_tokens": 361260891.0, + "step": 13966 + }, + { + "epoch": 1.5338238524050078, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5399420261383057, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7503077387809753, + "num_tokens": 361280470.0, + "step": 13967 + }, + { + "epoch": 1.5339336701076214, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0508017539978027, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7169845104217529, + "num_tokens": 361308667.0, + "step": 13968 + }, + { + "epoch": 1.534043487810235, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2524049282073975, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7282937169075012, + "num_tokens": 361333878.0, + "step": 13969 + }, + { + "epoch": 1.5341533055128487, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2236075401306152, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.700972318649292, + "num_tokens": 361360833.0, + "step": 13970 + }, + { + "epoch": 1.5342631232154624, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2539000511169434, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7111894488334656, + "num_tokens": 361388075.0, + "step": 13971 + }, + { + "epoch": 1.534372940918076, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3076438903808594, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7252940535545349, + "num_tokens": 361413148.0, + "step": 13972 + }, + { + "epoch": 1.5344827586206895, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2740721702575684, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7120398879051208, + "num_tokens": 361441051.0, + "step": 13973 + }, + { + "epoch": 1.5345925763233033, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.38091778755188, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7270342111587524, + "num_tokens": 361464033.0, + "step": 13974 + }, + { + "epoch": 1.534702394025917, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.189246654510498, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7228987216949463, + "num_tokens": 361490130.0, + "step": 13975 + }, + { + "epoch": 1.5348122117285308, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1046905517578125, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7335787415504456, + "num_tokens": 361516786.0, + "step": 13976 + }, + { + "epoch": 1.5349220294311443, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3004910945892334, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7352849841117859, + "num_tokens": 361541102.0, + "step": 13977 + }, + { + "epoch": 1.5350318471337578, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5303990840911865, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7055433988571167, + "num_tokens": 361564405.0, + "step": 13978 + }, + { + "epoch": 1.5351416648363716, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.352849245071411, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.737406313419342, + "num_tokens": 361587221.0, + "step": 13979 + }, + { + "epoch": 1.5352514825389854, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.151594400405884, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6899706125259399, + "num_tokens": 361617470.0, + "step": 13980 + }, + { + "epoch": 1.535361300241599, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.411172866821289, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7074185609817505, + "num_tokens": 361641579.0, + "step": 13981 + }, + { + "epoch": 1.5354711179442126, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.65159273147583, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7213816046714783, + "num_tokens": 361660463.0, + "step": 13982 + }, + { + "epoch": 1.5355809356468262, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1842470169067383, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7152968645095825, + "num_tokens": 361687467.0, + "step": 13983 + }, + { + "epoch": 1.53569075334944, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3201324939727783, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7331380844116211, + "num_tokens": 361711970.0, + "step": 13984 + }, + { + "epoch": 1.5358005710520537, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5082848072052, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7224631309509277, + "num_tokens": 361734584.0, + "step": 13985 + }, + { + "epoch": 1.5359103887546672, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0746426582336426, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7145876288414001, + "num_tokens": 361764865.0, + "step": 13986 + }, + { + "epoch": 1.5360202064572808, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.085472583770752, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.712368369102478, + "num_tokens": 361794556.0, + "step": 13987 + }, + { + "epoch": 1.5361300241598945, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.341545581817627, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7224377393722534, + "num_tokens": 361820100.0, + "step": 13988 + }, + { + "epoch": 1.5362398418625083, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1829991340637207, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7105884552001953, + "num_tokens": 361847555.0, + "step": 13989 + }, + { + "epoch": 1.536349659565122, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2621712684631348, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7247462272644043, + "num_tokens": 361873202.0, + "step": 13990 + }, + { + "epoch": 1.5364594772677356, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0955159664154053, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7284386157989502, + "num_tokens": 361900641.0, + "step": 13991 + }, + { + "epoch": 1.536569294970349, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.555882453918457, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7204732894897461, + "num_tokens": 361920081.0, + "step": 13992 + }, + { + "epoch": 1.5366791126729629, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9827799797058105, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7094299793243408, + "num_tokens": 361951476.0, + "step": 13993 + }, + { + "epoch": 1.5367889303755766, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2416515350341797, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7153952121734619, + "num_tokens": 361978693.0, + "step": 13994 + }, + { + "epoch": 1.5368987480781902, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.492138147354126, + "learning_rate": 1e-06, + "loss": 1.0707, + "mean_token_accuracy": 0.6828336715698242, + "num_tokens": 362001740.0, + "step": 13995 + }, + { + "epoch": 1.537008565780804, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3734397888183594, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7171398401260376, + "num_tokens": 362024253.0, + "step": 13996 + }, + { + "epoch": 1.5371183834834174, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2904889583587646, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.712505578994751, + "num_tokens": 362052889.0, + "step": 13997 + }, + { + "epoch": 1.5372282011860312, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0816192626953125, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7420414090156555, + "num_tokens": 362081419.0, + "step": 13998 + }, + { + "epoch": 1.537338018888645, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7364771366119385, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7340872883796692, + "num_tokens": 362101127.0, + "step": 13999 + }, + { + "epoch": 1.5374478365912585, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1846923828125, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7474729418754578, + "num_tokens": 362129145.0, + "step": 14000 + }, + { + "epoch": 1.537557654293872, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6235733032226562, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7349333763122559, + "num_tokens": 362148434.0, + "step": 14001 + }, + { + "epoch": 1.5376674719964858, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.080155372619629, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.707086443901062, + "num_tokens": 362177949.0, + "step": 14002 + }, + { + "epoch": 1.5377772896990995, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.305147647857666, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7014704942703247, + "num_tokens": 362202940.0, + "step": 14003 + }, + { + "epoch": 1.5378871074017133, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.339521884918213, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.718164324760437, + "num_tokens": 362227009.0, + "step": 14004 + }, + { + "epoch": 1.5379969251043268, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.301445484161377, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.732269823551178, + "num_tokens": 362251623.0, + "step": 14005 + }, + { + "epoch": 1.5381067428069404, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0926249027252197, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6876922845840454, + "num_tokens": 362282130.0, + "step": 14006 + }, + { + "epoch": 1.5382165605095541, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2198667526245117, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.696770429611206, + "num_tokens": 362309947.0, + "step": 14007 + }, + { + "epoch": 1.5383263782121679, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 4.911561489105225, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7242931127548218, + "num_tokens": 362331758.0, + "step": 14008 + }, + { + "epoch": 1.5384361959147814, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.30126953125, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.714993953704834, + "num_tokens": 362355589.0, + "step": 14009 + }, + { + "epoch": 1.538546013617395, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3683385848999023, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7333394289016724, + "num_tokens": 362378441.0, + "step": 14010 + }, + { + "epoch": 1.5386558313200087, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.125169515609741, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7185603380203247, + "num_tokens": 362406264.0, + "step": 14011 + }, + { + "epoch": 1.5387656490226225, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3770134449005127, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7272508144378662, + "num_tokens": 362429698.0, + "step": 14012 + }, + { + "epoch": 1.5388754667252362, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3914034366607666, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7354585528373718, + "num_tokens": 362451815.0, + "step": 14013 + }, + { + "epoch": 1.5389852844278498, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.418564796447754, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7076386213302612, + "num_tokens": 362475655.0, + "step": 14014 + }, + { + "epoch": 1.5390951021304633, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9872575998306274, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6972196102142334, + "num_tokens": 362509816.0, + "step": 14015 + }, + { + "epoch": 1.539204919833077, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.170684814453125, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7266784906387329, + "num_tokens": 362536931.0, + "step": 14016 + }, + { + "epoch": 1.5393147375356908, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4089014530181885, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7278450727462769, + "num_tokens": 362558784.0, + "step": 14017 + }, + { + "epoch": 1.5394245552383046, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.455857276916504, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7521535754203796, + "num_tokens": 362579862.0, + "step": 14018 + }, + { + "epoch": 1.539534372940918, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.235508918762207, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7407243251800537, + "num_tokens": 362604486.0, + "step": 14019 + }, + { + "epoch": 1.5396441906435316, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0240445137023926, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7294963598251343, + "num_tokens": 362632912.0, + "step": 14020 + }, + { + "epoch": 1.5397540083461454, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.169738292694092, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7346155643463135, + "num_tokens": 362658011.0, + "step": 14021 + }, + { + "epoch": 1.5398638260487592, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.530592679977417, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7081093788146973, + "num_tokens": 362678574.0, + "step": 14022 + }, + { + "epoch": 1.5399736437513727, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.13154673576355, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7032064199447632, + "num_tokens": 362707936.0, + "step": 14023 + }, + { + "epoch": 1.5400834614539862, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.228916645050049, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6995471715927124, + "num_tokens": 362736487.0, + "step": 14024 + }, + { + "epoch": 1.5401932791566, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.403301239013672, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.712028443813324, + "num_tokens": 362759878.0, + "step": 14025 + }, + { + "epoch": 1.5403030968592137, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.05307936668396, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7012927532196045, + "num_tokens": 362788417.0, + "step": 14026 + }, + { + "epoch": 1.5404129145618275, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0436534881591797, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7221757173538208, + "num_tokens": 362817928.0, + "step": 14027 + }, + { + "epoch": 1.540522732264441, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1392626762390137, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6890238523483276, + "num_tokens": 362846569.0, + "step": 14028 + }, + { + "epoch": 1.5406325499670546, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.25563907623291, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7361587285995483, + "num_tokens": 362872228.0, + "step": 14029 + }, + { + "epoch": 1.5407423676696683, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.332620143890381, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7330149412155151, + "num_tokens": 362894173.0, + "step": 14030 + }, + { + "epoch": 1.540852185372282, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.218050479888916, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7172236442565918, + "num_tokens": 362920259.0, + "step": 14031 + }, + { + "epoch": 1.5409620030748958, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.620927333831787, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7439388632774353, + "num_tokens": 362938237.0, + "step": 14032 + }, + { + "epoch": 1.5410718207775094, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3252413272857666, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7211915850639343, + "num_tokens": 362961984.0, + "step": 14033 + }, + { + "epoch": 1.541181638480123, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.52067232131958, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6952059268951416, + "num_tokens": 362985064.0, + "step": 14034 + }, + { + "epoch": 1.5412914561827367, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.326176881790161, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7124028205871582, + "num_tokens": 363010611.0, + "step": 14035 + }, + { + "epoch": 1.5414012738853504, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.456435203552246, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7307183742523193, + "num_tokens": 363032997.0, + "step": 14036 + }, + { + "epoch": 1.541511091587964, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1554925441741943, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7008236050605774, + "num_tokens": 363060709.0, + "step": 14037 + }, + { + "epoch": 1.5416209092905775, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2132670879364014, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7131149172782898, + "num_tokens": 363087751.0, + "step": 14038 + }, + { + "epoch": 1.5417307269931912, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.499444007873535, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7163023948669434, + "num_tokens": 363109821.0, + "step": 14039 + }, + { + "epoch": 1.541840544695805, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.101525068283081, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7039520740509033, + "num_tokens": 363140215.0, + "step": 14040 + }, + { + "epoch": 1.5419503623984188, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2693886756896973, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7125282883644104, + "num_tokens": 363164361.0, + "step": 14041 + }, + { + "epoch": 1.5420601801010323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9721192121505737, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.703838586807251, + "num_tokens": 363199268.0, + "step": 14042 + }, + { + "epoch": 1.5421699978036458, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1184046268463135, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7013493776321411, + "num_tokens": 363229026.0, + "step": 14043 + }, + { + "epoch": 1.5422798155062596, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1591837406158447, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7281771302223206, + "num_tokens": 363257039.0, + "step": 14044 + }, + { + "epoch": 1.5423896332088733, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2293801307678223, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7045925855636597, + "num_tokens": 363282409.0, + "step": 14045 + }, + { + "epoch": 1.5424994509114869, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.182033061981201, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7102488279342651, + "num_tokens": 363311899.0, + "step": 14046 + }, + { + "epoch": 1.5426092686141006, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.173860549926758, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7237759828567505, + "num_tokens": 363337782.0, + "step": 14047 + }, + { + "epoch": 1.5427190863167142, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3755180835723877, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7301124334335327, + "num_tokens": 363361218.0, + "step": 14048 + }, + { + "epoch": 1.542828904019328, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.986702799797058, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7291255593299866, + "num_tokens": 363390987.0, + "step": 14049 + }, + { + "epoch": 1.5429387217219417, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.818919062614441, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7376217246055603, + "num_tokens": 363423598.0, + "step": 14050 + }, + { + "epoch": 1.5430485394245552, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3644003868103027, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7289280891418457, + "num_tokens": 363445154.0, + "step": 14051 + }, + { + "epoch": 1.5431583571271688, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0305685997009277, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7032230496406555, + "num_tokens": 363475840.0, + "step": 14052 + }, + { + "epoch": 1.5432681748297825, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3422698974609375, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7386614680290222, + "num_tokens": 363497933.0, + "step": 14053 + }, + { + "epoch": 1.5433779925323963, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1076481342315674, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.716957688331604, + "num_tokens": 363524024.0, + "step": 14054 + }, + { + "epoch": 1.54348781023501, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2769412994384766, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7182687520980835, + "num_tokens": 363549265.0, + "step": 14055 + }, + { + "epoch": 1.5435976279376236, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.512399196624756, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7228741645812988, + "num_tokens": 363571016.0, + "step": 14056 + }, + { + "epoch": 1.543707445640237, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0909032821655273, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6972389221191406, + "num_tokens": 363600326.0, + "step": 14057 + }, + { + "epoch": 1.5438172633428509, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1910316944122314, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7157823443412781, + "num_tokens": 363626848.0, + "step": 14058 + }, + { + "epoch": 1.5439270810454646, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0723581314086914, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7197060585021973, + "num_tokens": 363656209.0, + "step": 14059 + }, + { + "epoch": 1.5440368987480781, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1610381603240967, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7177740335464478, + "num_tokens": 363683720.0, + "step": 14060 + }, + { + "epoch": 1.544146716450692, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.390432357788086, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7159867286682129, + "num_tokens": 363705783.0, + "step": 14061 + }, + { + "epoch": 1.5442565341533054, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3346941471099854, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7198644876480103, + "num_tokens": 363729198.0, + "step": 14062 + }, + { + "epoch": 1.5443663518559192, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.043217658996582, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7379305362701416, + "num_tokens": 363757236.0, + "step": 14063 + }, + { + "epoch": 1.544476169558533, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2064764499664307, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6890321373939514, + "num_tokens": 363784799.0, + "step": 14064 + }, + { + "epoch": 1.5445859872611465, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.462951898574829, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7259677052497864, + "num_tokens": 363806497.0, + "step": 14065 + }, + { + "epoch": 1.54469580496376, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.233107805252075, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7006319165229797, + "num_tokens": 363833430.0, + "step": 14066 + }, + { + "epoch": 1.5448056226663738, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.41432523727417, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7092859148979187, + "num_tokens": 363856361.0, + "step": 14067 + }, + { + "epoch": 1.5449154403689875, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2561216354370117, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.726621687412262, + "num_tokens": 363881557.0, + "step": 14068 + }, + { + "epoch": 1.5450252580716013, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6023333072662354, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7162017822265625, + "num_tokens": 363901078.0, + "step": 14069 + }, + { + "epoch": 1.5451350757742148, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4101195335388184, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7410322427749634, + "num_tokens": 363923541.0, + "step": 14070 + }, + { + "epoch": 1.5452448934768284, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3232295513153076, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7165625095367432, + "num_tokens": 363947757.0, + "step": 14071 + }, + { + "epoch": 1.5453547111794421, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.511620044708252, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7093828320503235, + "num_tokens": 363969680.0, + "step": 14072 + }, + { + "epoch": 1.5454645288820559, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.254122495651245, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6985310316085815, + "num_tokens": 363995485.0, + "step": 14073 + }, + { + "epoch": 1.5455743465846694, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.486690044403076, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7093954086303711, + "num_tokens": 364017312.0, + "step": 14074 + }, + { + "epoch": 1.545684164287283, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.162177085876465, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7138949036598206, + "num_tokens": 364043938.0, + "step": 14075 + }, + { + "epoch": 1.5457939819898967, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.334578514099121, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7384762167930603, + "num_tokens": 364066453.0, + "step": 14076 + }, + { + "epoch": 1.5459037996925105, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3605055809020996, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7004340291023254, + "num_tokens": 364089828.0, + "step": 14077 + }, + { + "epoch": 1.5460136173951242, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2331948280334473, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7208852767944336, + "num_tokens": 364117049.0, + "step": 14078 + }, + { + "epoch": 1.5461234350977378, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3146419525146484, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7224968671798706, + "num_tokens": 364141653.0, + "step": 14079 + }, + { + "epoch": 1.5462332528003513, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2237162590026855, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7462211847305298, + "num_tokens": 364167485.0, + "step": 14080 + }, + { + "epoch": 1.546343070502965, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.837855577468872, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7517697811126709, + "num_tokens": 364183903.0, + "step": 14081 + }, + { + "epoch": 1.5464528882055788, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1897013187408447, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7060855627059937, + "num_tokens": 364210196.0, + "step": 14082 + }, + { + "epoch": 1.5465627059081926, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1607284545898438, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7307830452919006, + "num_tokens": 364235182.0, + "step": 14083 + }, + { + "epoch": 1.546672523610806, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.196549654006958, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7173236608505249, + "num_tokens": 364260900.0, + "step": 14084 + }, + { + "epoch": 1.5467823413134196, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2384653091430664, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6984343528747559, + "num_tokens": 364288023.0, + "step": 14085 + }, + { + "epoch": 1.5468921590160334, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9554001092910767, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7155414819717407, + "num_tokens": 364318840.0, + "step": 14086 + }, + { + "epoch": 1.5470019767186471, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.145371198654175, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7177108526229858, + "num_tokens": 364346546.0, + "step": 14087 + }, + { + "epoch": 1.5471117944212607, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.69108510017395, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7309104204177856, + "num_tokens": 364367853.0, + "step": 14088 + }, + { + "epoch": 1.5472216121238742, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0498573780059814, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7103921175003052, + "num_tokens": 364395855.0, + "step": 14089 + }, + { + "epoch": 1.547331429826488, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3212742805480957, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.6979473829269409, + "num_tokens": 364421719.0, + "step": 14090 + }, + { + "epoch": 1.5474412475291017, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5807089805603027, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7340219020843506, + "num_tokens": 364444366.0, + "step": 14091 + }, + { + "epoch": 1.5475510652317155, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4095394611358643, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7106735110282898, + "num_tokens": 364467061.0, + "step": 14092 + }, + { + "epoch": 1.547660882934329, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4521548748016357, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7312444448471069, + "num_tokens": 364488964.0, + "step": 14093 + }, + { + "epoch": 1.5477707006369426, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.434810161590576, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7142326235771179, + "num_tokens": 364511490.0, + "step": 14094 + }, + { + "epoch": 1.5478805183395563, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.374711513519287, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7231279015541077, + "num_tokens": 364537702.0, + "step": 14095 + }, + { + "epoch": 1.54799033604217, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.035849094390869, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7231402397155762, + "num_tokens": 364566065.0, + "step": 14096 + }, + { + "epoch": 1.5481001537447838, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.275118589401245, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7232614755630493, + "num_tokens": 364591670.0, + "step": 14097 + }, + { + "epoch": 1.5482099714473974, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9756009578704834, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7293779253959656, + "num_tokens": 364622290.0, + "step": 14098 + }, + { + "epoch": 1.548319789150011, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5016796588897705, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7233280539512634, + "num_tokens": 364643694.0, + "step": 14099 + }, + { + "epoch": 1.5484296068526247, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.22489857673645, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7130476236343384, + "num_tokens": 364671334.0, + "step": 14100 + }, + { + "epoch": 1.5485394245552384, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3756673336029053, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.715797483921051, + "num_tokens": 364694191.0, + "step": 14101 + }, + { + "epoch": 1.548649242257852, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3805832862854004, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.712023138999939, + "num_tokens": 364718287.0, + "step": 14102 + }, + { + "epoch": 1.5487590599604655, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3057236671447754, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7454686760902405, + "num_tokens": 364742007.0, + "step": 14103 + }, + { + "epoch": 1.5488688776630792, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0553579330444336, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7125078439712524, + "num_tokens": 364772320.0, + "step": 14104 + }, + { + "epoch": 1.548978695365693, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4153497219085693, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7236645221710205, + "num_tokens": 364794693.0, + "step": 14105 + }, + { + "epoch": 1.5490885130683067, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2579493522644043, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7132587432861328, + "num_tokens": 364821681.0, + "step": 14106 + }, + { + "epoch": 1.5491983307709203, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9637037515640259, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6908571720123291, + "num_tokens": 364854392.0, + "step": 14107 + }, + { + "epoch": 1.5493081484735338, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2053987979888916, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.723158061504364, + "num_tokens": 364882341.0, + "step": 14108 + }, + { + "epoch": 1.5494179661761476, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3411335945129395, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7116391658782959, + "num_tokens": 364905926.0, + "step": 14109 + }, + { + "epoch": 1.5495277838787613, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0899248123168945, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7030813097953796, + "num_tokens": 364934990.0, + "step": 14110 + }, + { + "epoch": 1.5496376015813749, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.355750322341919, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7086837291717529, + "num_tokens": 364960789.0, + "step": 14111 + }, + { + "epoch": 1.5497474192839886, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1047751903533936, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7096864581108093, + "num_tokens": 364986802.0, + "step": 14112 + }, + { + "epoch": 1.5498572369866022, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2440404891967773, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7137174606323242, + "num_tokens": 365012297.0, + "step": 14113 + }, + { + "epoch": 1.549967054689216, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.208329916000366, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7332156896591187, + "num_tokens": 365037997.0, + "step": 14114 + }, + { + "epoch": 1.5500768723918297, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.64129638671875, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7486640214920044, + "num_tokens": 365058239.0, + "step": 14115 + }, + { + "epoch": 1.5501866900944432, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.236501932144165, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7067621946334839, + "num_tokens": 365086407.0, + "step": 14116 + }, + { + "epoch": 1.5502965077970567, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4408631324768066, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.737334132194519, + "num_tokens": 365109176.0, + "step": 14117 + }, + { + "epoch": 1.5504063254996705, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.445194721221924, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7217891216278076, + "num_tokens": 365131615.0, + "step": 14118 + }, + { + "epoch": 1.5505161432022843, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3335535526275635, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.7039947509765625, + "num_tokens": 365158257.0, + "step": 14119 + }, + { + "epoch": 1.550625960904898, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7818052768707275, + "learning_rate": 1e-06, + "loss": 0.7643, + "mean_token_accuracy": 0.7544772028923035, + "num_tokens": 365175178.0, + "step": 14120 + }, + { + "epoch": 1.5507357786075116, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4485509395599365, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7264181971549988, + "num_tokens": 365197933.0, + "step": 14121 + }, + { + "epoch": 1.550845596310125, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0494725704193115, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7053308486938477, + "num_tokens": 365233179.0, + "step": 14122 + }, + { + "epoch": 1.5509554140127388, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1584489345550537, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7096996307373047, + "num_tokens": 365259249.0, + "step": 14123 + }, + { + "epoch": 1.5510652317153526, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.340568780899048, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7219245433807373, + "num_tokens": 365282260.0, + "step": 14124 + }, + { + "epoch": 1.5511750494179661, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1879255771636963, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7165840864181519, + "num_tokens": 365308282.0, + "step": 14125 + }, + { + "epoch": 1.55128486712058, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1571104526519775, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7228503823280334, + "num_tokens": 365335380.0, + "step": 14126 + }, + { + "epoch": 1.5513946848231934, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.202181339263916, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7153902053833008, + "num_tokens": 365360887.0, + "step": 14127 + }, + { + "epoch": 1.5515045025258072, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5457804203033447, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7412651181221008, + "num_tokens": 365380668.0, + "step": 14128 + }, + { + "epoch": 1.551614320228421, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.115043878555298, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7286828756332397, + "num_tokens": 365407323.0, + "step": 14129 + }, + { + "epoch": 1.5517241379310345, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7158255577087402, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7411807775497437, + "num_tokens": 365426765.0, + "step": 14130 + }, + { + "epoch": 1.551833955633648, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.132823944091797, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7228219509124756, + "num_tokens": 365452730.0, + "step": 14131 + }, + { + "epoch": 1.5519437733362618, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2179477214813232, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7252976894378662, + "num_tokens": 365478368.0, + "step": 14132 + }, + { + "epoch": 1.5520535910388755, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3416686058044434, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7157285213470459, + "num_tokens": 365502634.0, + "step": 14133 + }, + { + "epoch": 1.5521634087414893, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5464894771575928, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7289613485336304, + "num_tokens": 365523409.0, + "step": 14134 + }, + { + "epoch": 1.5522732264441028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.152784585952759, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7228056192398071, + "num_tokens": 365550306.0, + "step": 14135 + }, + { + "epoch": 1.5523830441467164, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.169182062149048, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7451385259628296, + "num_tokens": 365575175.0, + "step": 14136 + }, + { + "epoch": 1.55249286184933, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7086410522460938, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7507081627845764, + "num_tokens": 365592589.0, + "step": 14137 + }, + { + "epoch": 1.5526026795519439, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4281680583953857, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7279441356658936, + "num_tokens": 365615450.0, + "step": 14138 + }, + { + "epoch": 1.5527124972545574, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.200575828552246, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7221956253051758, + "num_tokens": 365643137.0, + "step": 14139 + }, + { + "epoch": 1.552822314957171, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.084230661392212, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7238043546676636, + "num_tokens": 365670653.0, + "step": 14140 + }, + { + "epoch": 1.5529321326597847, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1159920692443848, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7008734345436096, + "num_tokens": 365699304.0, + "step": 14141 + }, + { + "epoch": 1.5530419503623984, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1686594486236572, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7022007703781128, + "num_tokens": 365727050.0, + "step": 14142 + }, + { + "epoch": 1.5531517680650122, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3843119144439697, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7196996212005615, + "num_tokens": 365752438.0, + "step": 14143 + }, + { + "epoch": 1.5532615857676257, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.282761812210083, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7163029909133911, + "num_tokens": 365775100.0, + "step": 14144 + }, + { + "epoch": 1.5533714034702393, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.159926176071167, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6886459589004517, + "num_tokens": 365804574.0, + "step": 14145 + }, + { + "epoch": 1.553481221172853, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4851765632629395, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7043780088424683, + "num_tokens": 365827418.0, + "step": 14146 + }, + { + "epoch": 1.5535910388754668, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.46197247505188, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.752549946308136, + "num_tokens": 365847419.0, + "step": 14147 + }, + { + "epoch": 1.5537008565780805, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1514835357666016, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7218174934387207, + "num_tokens": 365874118.0, + "step": 14148 + }, + { + "epoch": 1.553810674280694, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.109910249710083, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7120286226272583, + "num_tokens": 365902006.0, + "step": 14149 + }, + { + "epoch": 1.5539204919833076, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.479353427886963, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7210608720779419, + "num_tokens": 365922732.0, + "step": 14150 + }, + { + "epoch": 1.5540303096859214, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.183800458908081, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6952619552612305, + "num_tokens": 365951237.0, + "step": 14151 + }, + { + "epoch": 1.5541401273885351, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.237074851989746, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7230351567268372, + "num_tokens": 365976298.0, + "step": 14152 + }, + { + "epoch": 1.5542499450911487, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2832159996032715, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7017125487327576, + "num_tokens": 366003615.0, + "step": 14153 + }, + { + "epoch": 1.5543597627937622, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3701560497283936, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7149418592453003, + "num_tokens": 366028463.0, + "step": 14154 + }, + { + "epoch": 1.554469580496376, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.257824420928955, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7132872939109802, + "num_tokens": 366054708.0, + "step": 14155 + }, + { + "epoch": 1.5545793981989897, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0636518001556396, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7201181650161743, + "num_tokens": 366084130.0, + "step": 14156 + }, + { + "epoch": 1.5546892159016035, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5480685234069824, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7411174178123474, + "num_tokens": 366105798.0, + "step": 14157 + }, + { + "epoch": 1.554799033604217, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9041589498519897, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7055532932281494, + "num_tokens": 366138830.0, + "step": 14158 + }, + { + "epoch": 1.5549088513068305, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.341024875640869, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7305034399032593, + "num_tokens": 366164338.0, + "step": 14159 + }, + { + "epoch": 1.5550186690094443, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3968865871429443, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7183415293693542, + "num_tokens": 366188310.0, + "step": 14160 + }, + { + "epoch": 1.555128486712058, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0908915996551514, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7294114232063293, + "num_tokens": 366218560.0, + "step": 14161 + }, + { + "epoch": 1.5552383044146716, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4926419258117676, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7265750765800476, + "num_tokens": 366239808.0, + "step": 14162 + }, + { + "epoch": 1.5553481221172853, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2536675930023193, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7324491143226624, + "num_tokens": 366265356.0, + "step": 14163 + }, + { + "epoch": 1.5554579398198989, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.128185510635376, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7177952527999878, + "num_tokens": 366294312.0, + "step": 14164 + }, + { + "epoch": 1.5555677575225126, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5782039165496826, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7128797173500061, + "num_tokens": 366315345.0, + "step": 14165 + }, + { + "epoch": 1.5556775752251264, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2950425148010254, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7077639698982239, + "num_tokens": 366338596.0, + "step": 14166 + }, + { + "epoch": 1.55578739292774, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.42468523979187, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7461955547332764, + "num_tokens": 366360721.0, + "step": 14167 + }, + { + "epoch": 1.5558972106303535, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.451965570449829, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7337113618850708, + "num_tokens": 366381748.0, + "step": 14168 + }, + { + "epoch": 1.5560070283329672, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1675808429718018, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7239381670951843, + "num_tokens": 366408008.0, + "step": 14169 + }, + { + "epoch": 1.556116846035581, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3559796810150146, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7213995456695557, + "num_tokens": 366431701.0, + "step": 14170 + }, + { + "epoch": 1.5562266637381947, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1459169387817383, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7276889085769653, + "num_tokens": 366457522.0, + "step": 14171 + }, + { + "epoch": 1.5563364814408083, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.218961238861084, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7376998662948608, + "num_tokens": 366481980.0, + "step": 14172 + }, + { + "epoch": 1.5564462991434218, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4508204460144043, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6992770433425903, + "num_tokens": 366504694.0, + "step": 14173 + }, + { + "epoch": 1.5565561168460356, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4267923831939697, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7194010019302368, + "num_tokens": 366526182.0, + "step": 14174 + }, + { + "epoch": 1.5566659345486493, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.076704502105713, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.729282021522522, + "num_tokens": 366555008.0, + "step": 14175 + }, + { + "epoch": 1.5567757522512629, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.237961769104004, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.724493682384491, + "num_tokens": 366580796.0, + "step": 14176 + }, + { + "epoch": 1.5568855699538766, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2953433990478516, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7126786112785339, + "num_tokens": 366606322.0, + "step": 14177 + }, + { + "epoch": 1.5569953876564901, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.337606191635132, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7211290597915649, + "num_tokens": 366629526.0, + "step": 14178 + }, + { + "epoch": 1.557105205359104, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.441786527633667, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7347593307495117, + "num_tokens": 366650121.0, + "step": 14179 + }, + { + "epoch": 1.5572150230617177, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4691810607910156, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7186114192008972, + "num_tokens": 366673256.0, + "step": 14180 + }, + { + "epoch": 1.5573248407643312, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.095127820968628, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7148516178131104, + "num_tokens": 366703456.0, + "step": 14181 + }, + { + "epoch": 1.5574346584669447, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.083111524581909, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7044134140014648, + "num_tokens": 366734352.0, + "step": 14182 + }, + { + "epoch": 1.5575444761695585, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2308149337768555, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7236447334289551, + "num_tokens": 366761205.0, + "step": 14183 + }, + { + "epoch": 1.5576542938721722, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.259444236755371, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7249295711517334, + "num_tokens": 366785165.0, + "step": 14184 + }, + { + "epoch": 1.557764111574786, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2167258262634277, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7240725755691528, + "num_tokens": 366809717.0, + "step": 14185 + }, + { + "epoch": 1.5578739292773995, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2505288124084473, + "learning_rate": 1e-06, + "loss": 0.7822, + "mean_token_accuracy": 0.7462959289550781, + "num_tokens": 366832449.0, + "step": 14186 + }, + { + "epoch": 1.557983746980013, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.386342763900757, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7255680561065674, + "num_tokens": 366857295.0, + "step": 14187 + }, + { + "epoch": 1.5580935646826268, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1870276927948, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6920411586761475, + "num_tokens": 366885833.0, + "step": 14188 + }, + { + "epoch": 1.5582033823852406, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.124640941619873, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7125105857849121, + "num_tokens": 366912767.0, + "step": 14189 + }, + { + "epoch": 1.5583132000878541, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4399349689483643, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.716238260269165, + "num_tokens": 366935518.0, + "step": 14190 + }, + { + "epoch": 1.5584230177904677, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.249150514602661, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7227644920349121, + "num_tokens": 366960821.0, + "step": 14191 + }, + { + "epoch": 1.5585328354930814, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.8963080644607544, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.6955060958862305, + "num_tokens": 366996506.0, + "step": 14192 + }, + { + "epoch": 1.5586426531956952, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3881733417510986, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7113898992538452, + "num_tokens": 367021721.0, + "step": 14193 + }, + { + "epoch": 1.558752470898309, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2778899669647217, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7286474108695984, + "num_tokens": 367047808.0, + "step": 14194 + }, + { + "epoch": 1.5588622886009225, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3163788318634033, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7025145888328552, + "num_tokens": 367073388.0, + "step": 14195 + }, + { + "epoch": 1.558972106303536, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.224562644958496, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7048860788345337, + "num_tokens": 367101125.0, + "step": 14196 + }, + { + "epoch": 1.5590819240061498, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.391355276107788, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7056688666343689, + "num_tokens": 367126849.0, + "step": 14197 + }, + { + "epoch": 1.5591917417087635, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3403520584106445, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7141041159629822, + "num_tokens": 367150035.0, + "step": 14198 + }, + { + "epoch": 1.5593015594113773, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.375405788421631, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.739224910736084, + "num_tokens": 367171590.0, + "step": 14199 + }, + { + "epoch": 1.5594113771139908, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.987870216369629, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7308642268180847, + "num_tokens": 367199356.0, + "step": 14200 + }, + { + "epoch": 1.5595211948166043, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.057893991470337, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.701285719871521, + "num_tokens": 367228959.0, + "step": 14201 + }, + { + "epoch": 1.559631012519218, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1217241287231445, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7200997471809387, + "num_tokens": 367256230.0, + "step": 14202 + }, + { + "epoch": 1.5597408302218319, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0201282501220703, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7075174450874329, + "num_tokens": 367287741.0, + "step": 14203 + }, + { + "epoch": 1.5598506479244454, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4278926849365234, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.73628830909729, + "num_tokens": 367310656.0, + "step": 14204 + }, + { + "epoch": 1.559960465627059, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2057251930236816, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7221790552139282, + "num_tokens": 367337462.0, + "step": 14205 + }, + { + "epoch": 1.5600702833296727, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1999077796936035, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7280829548835754, + "num_tokens": 367363685.0, + "step": 14206 + }, + { + "epoch": 1.5601801010322864, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5355758666992188, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7194573879241943, + "num_tokens": 367386120.0, + "step": 14207 + }, + { + "epoch": 1.5602899187349002, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.102123737335205, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6891269683837891, + "num_tokens": 367417056.0, + "step": 14208 + }, + { + "epoch": 1.5603997364375137, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2067978382110596, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7219104766845703, + "num_tokens": 367442342.0, + "step": 14209 + }, + { + "epoch": 1.5605095541401273, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.960403323173523, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7204872369766235, + "num_tokens": 367472690.0, + "step": 14210 + }, + { + "epoch": 1.560619371842741, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2908620834350586, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7111353278160095, + "num_tokens": 367498412.0, + "step": 14211 + }, + { + "epoch": 1.5607291895453548, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.123568296432495, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7335004806518555, + "num_tokens": 367525955.0, + "step": 14212 + }, + { + "epoch": 1.5608390072479685, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0261712074279785, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7349400520324707, + "num_tokens": 367554794.0, + "step": 14213 + }, + { + "epoch": 1.560948824950582, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.025252103805542, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7101346254348755, + "num_tokens": 367586005.0, + "step": 14214 + }, + { + "epoch": 1.5610586426531956, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.155442237854004, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7307414412498474, + "num_tokens": 367612133.0, + "step": 14215 + }, + { + "epoch": 1.5611684603558094, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4512643814086914, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7253684997558594, + "num_tokens": 367633180.0, + "step": 14216 + }, + { + "epoch": 1.5612782780584231, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.14815354347229, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7076966166496277, + "num_tokens": 367663191.0, + "step": 14217 + }, + { + "epoch": 1.5613880957610367, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.275508403778076, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7093481421470642, + "num_tokens": 367689478.0, + "step": 14218 + }, + { + "epoch": 1.5614979134636502, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2674877643585205, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.739921510219574, + "num_tokens": 367713709.0, + "step": 14219 + }, + { + "epoch": 1.561607731166264, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0262198448181152, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7014856338500977, + "num_tokens": 367744479.0, + "step": 14220 + }, + { + "epoch": 1.5617175488688777, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6231648921966553, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7364656329154968, + "num_tokens": 367765403.0, + "step": 14221 + }, + { + "epoch": 1.5618273665714915, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1801915168762207, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7269887924194336, + "num_tokens": 367792333.0, + "step": 14222 + }, + { + "epoch": 1.561937184274105, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1721720695495605, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7256208658218384, + "num_tokens": 367817925.0, + "step": 14223 + }, + { + "epoch": 1.5620470019767185, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0840070247650146, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7201398611068726, + "num_tokens": 367846064.0, + "step": 14224 + }, + { + "epoch": 1.5621568196793323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4309089183807373, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7328066825866699, + "num_tokens": 367871089.0, + "step": 14225 + }, + { + "epoch": 1.562266637381946, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.76395845413208, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7286975383758545, + "num_tokens": 367891372.0, + "step": 14226 + }, + { + "epoch": 1.5623764550845596, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3637802600860596, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7290282249450684, + "num_tokens": 367914436.0, + "step": 14227 + }, + { + "epoch": 1.5624862727871733, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.737320899963379, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7483715415000916, + "num_tokens": 367931973.0, + "step": 14228 + }, + { + "epoch": 1.5625960904897869, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.173410654067993, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.724951446056366, + "num_tokens": 367957829.0, + "step": 14229 + }, + { + "epoch": 1.5627059081924006, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3734662532806396, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.6992671489715576, + "num_tokens": 367981594.0, + "step": 14230 + }, + { + "epoch": 1.5628157258950144, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0369157791137695, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.720909595489502, + "num_tokens": 368012875.0, + "step": 14231 + }, + { + "epoch": 1.562925543597628, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.238154888153076, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7246001958847046, + "num_tokens": 368038115.0, + "step": 14232 + }, + { + "epoch": 1.5630353613002415, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2601540088653564, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6973001956939697, + "num_tokens": 368064082.0, + "step": 14233 + }, + { + "epoch": 1.5631451790028552, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.486339569091797, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7225978374481201, + "num_tokens": 368085281.0, + "step": 14234 + }, + { + "epoch": 1.563254996705469, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3213703632354736, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7375718355178833, + "num_tokens": 368108808.0, + "step": 14235 + }, + { + "epoch": 1.5633648144080827, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.519862651824951, + "learning_rate": 1e-06, + "loss": 0.7597, + "mean_token_accuracy": 0.7571102380752563, + "num_tokens": 368127261.0, + "step": 14236 + }, + { + "epoch": 1.5634746321106963, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.108039140701294, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7454872131347656, + "num_tokens": 368152865.0, + "step": 14237 + }, + { + "epoch": 1.5635844498133098, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.962646245956421, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7063021659851074, + "num_tokens": 368183795.0, + "step": 14238 + }, + { + "epoch": 1.5636942675159236, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2605156898498535, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.706411600112915, + "num_tokens": 368209576.0, + "step": 14239 + }, + { + "epoch": 1.5638040852185373, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.009587526321411, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7246276140213013, + "num_tokens": 368240376.0, + "step": 14240 + }, + { + "epoch": 1.5639139029211508, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.989929437637329, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.721953809261322, + "num_tokens": 368269971.0, + "step": 14241 + }, + { + "epoch": 1.5640237206237646, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.8447160720825195, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7295951843261719, + "num_tokens": 368288264.0, + "step": 14242 + }, + { + "epoch": 1.5641335383263781, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.292616844177246, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7299978137016296, + "num_tokens": 368313174.0, + "step": 14243 + }, + { + "epoch": 1.564243356028992, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2511355876922607, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7230033278465271, + "num_tokens": 368338117.0, + "step": 14244 + }, + { + "epoch": 1.5643531737316057, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1781363487243652, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7292531728744507, + "num_tokens": 368363618.0, + "step": 14245 + }, + { + "epoch": 1.5644629914342192, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.57810115814209, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7006826400756836, + "num_tokens": 368384049.0, + "step": 14246 + }, + { + "epoch": 1.5645728091368327, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.568354368209839, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7409432530403137, + "num_tokens": 368404667.0, + "step": 14247 + }, + { + "epoch": 1.5646826268394465, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.31534743309021, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7235018014907837, + "num_tokens": 368429066.0, + "step": 14248 + }, + { + "epoch": 1.5647924445420602, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2892954349517822, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7075090408325195, + "num_tokens": 368455380.0, + "step": 14249 + }, + { + "epoch": 1.564902262244674, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6130387783050537, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7330057621002197, + "num_tokens": 368474136.0, + "step": 14250 + }, + { + "epoch": 1.5650120799472875, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.943411946296692, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7073122262954712, + "num_tokens": 368505615.0, + "step": 14251 + }, + { + "epoch": 1.565121897649901, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.309931993484497, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7056186199188232, + "num_tokens": 368530575.0, + "step": 14252 + }, + { + "epoch": 1.5652317153525148, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.590320110321045, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7291252613067627, + "num_tokens": 368550127.0, + "step": 14253 + }, + { + "epoch": 1.5653415330551286, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.545240879058838, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7340630888938904, + "num_tokens": 368570110.0, + "step": 14254 + }, + { + "epoch": 1.565451350757742, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2492311000823975, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7553472518920898, + "num_tokens": 368593773.0, + "step": 14255 + }, + { + "epoch": 1.5655611684603556, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.264267683029175, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7391233444213867, + "num_tokens": 368620602.0, + "step": 14256 + }, + { + "epoch": 1.5656709861629694, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3813560009002686, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7282395362854004, + "num_tokens": 368645461.0, + "step": 14257 + }, + { + "epoch": 1.5657808038655832, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.20176362991333, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7057677507400513, + "num_tokens": 368671850.0, + "step": 14258 + }, + { + "epoch": 1.565890621568197, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.229891061782837, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7275564670562744, + "num_tokens": 368697958.0, + "step": 14259 + }, + { + "epoch": 1.5660004392708105, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.153602361679077, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7129383087158203, + "num_tokens": 368725058.0, + "step": 14260 + }, + { + "epoch": 1.566110256973424, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.211653709411621, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7046211361885071, + "num_tokens": 368750045.0, + "step": 14261 + }, + { + "epoch": 1.5662200746760377, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1833155155181885, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7079685926437378, + "num_tokens": 368773970.0, + "step": 14262 + }, + { + "epoch": 1.5663298923786515, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.322470188140869, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7278060913085938, + "num_tokens": 368797353.0, + "step": 14263 + }, + { + "epoch": 1.5664397100812653, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2277536392211914, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7191003561019897, + "num_tokens": 368822956.0, + "step": 14264 + }, + { + "epoch": 1.5665495277838788, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6148440837860107, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7000163793563843, + "num_tokens": 368844391.0, + "step": 14265 + }, + { + "epoch": 1.5666593454864923, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1268625259399414, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7281910181045532, + "num_tokens": 368871579.0, + "step": 14266 + }, + { + "epoch": 1.566769163189106, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4438772201538086, + "learning_rate": 1e-06, + "loss": 0.7218, + "mean_token_accuracy": 0.7662354707717896, + "num_tokens": 368892953.0, + "step": 14267 + }, + { + "epoch": 1.5668789808917198, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3501205444335938, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7352278232574463, + "num_tokens": 368915633.0, + "step": 14268 + }, + { + "epoch": 1.5669887985943334, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4193460941314697, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.749343752861023, + "num_tokens": 368937851.0, + "step": 14269 + }, + { + "epoch": 1.567098616296947, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3792717456817627, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7292951345443726, + "num_tokens": 368959935.0, + "step": 14270 + }, + { + "epoch": 1.5672084339995607, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3308334350585938, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6965943574905396, + "num_tokens": 368983887.0, + "step": 14271 + }, + { + "epoch": 1.5673182517021744, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4026987552642822, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7084298133850098, + "num_tokens": 369007119.0, + "step": 14272 + }, + { + "epoch": 1.5674280694047882, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.009129524230957, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7159297466278076, + "num_tokens": 369038244.0, + "step": 14273 + }, + { + "epoch": 1.5675378871074017, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.253511667251587, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.721139669418335, + "num_tokens": 369064040.0, + "step": 14274 + }, + { + "epoch": 1.5676477048100153, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.8569238185882568, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7245047688484192, + "num_tokens": 369099983.0, + "step": 14275 + }, + { + "epoch": 1.567757522512629, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.34472918510437, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7181606292724609, + "num_tokens": 369124905.0, + "step": 14276 + }, + { + "epoch": 1.5678673402152428, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1659581661224365, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7065231204032898, + "num_tokens": 369152553.0, + "step": 14277 + }, + { + "epoch": 1.5679771579178565, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.512047290802002, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7363326549530029, + "num_tokens": 369173755.0, + "step": 14278 + }, + { + "epoch": 1.56808697562047, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9247390031814575, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7282336950302124, + "num_tokens": 369204835.0, + "step": 14279 + }, + { + "epoch": 1.5681967933230836, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.332798719406128, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7026103138923645, + "num_tokens": 369229517.0, + "step": 14280 + }, + { + "epoch": 1.5683066110256974, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3407952785491943, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7079521417617798, + "num_tokens": 369251994.0, + "step": 14281 + }, + { + "epoch": 1.568416428728311, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2020010948181152, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7078717947006226, + "num_tokens": 369278656.0, + "step": 14282 + }, + { + "epoch": 1.5685262464309246, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.277987003326416, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7113959789276123, + "num_tokens": 369307060.0, + "step": 14283 + }, + { + "epoch": 1.5686360641335382, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.085437536239624, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7084660530090332, + "num_tokens": 369334937.0, + "step": 14284 + }, + { + "epoch": 1.568745881836152, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4400060176849365, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7194929122924805, + "num_tokens": 369359214.0, + "step": 14285 + }, + { + "epoch": 1.5688556995387657, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.123290777206421, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7293797731399536, + "num_tokens": 369387553.0, + "step": 14286 + }, + { + "epoch": 1.5689655172413794, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2845616340637207, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7241605520248413, + "num_tokens": 369411021.0, + "step": 14287 + }, + { + "epoch": 1.569075334943993, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4584178924560547, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7514019012451172, + "num_tokens": 369433815.0, + "step": 14288 + }, + { + "epoch": 1.5691851526466065, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.397542715072632, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7223719358444214, + "num_tokens": 369456333.0, + "step": 14289 + }, + { + "epoch": 1.5692949703492203, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3082265853881836, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7199300527572632, + "num_tokens": 369479927.0, + "step": 14290 + }, + { + "epoch": 1.569404788051834, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.103830337524414, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7116951942443848, + "num_tokens": 369509725.0, + "step": 14291 + }, + { + "epoch": 1.5695146057544476, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2870450019836426, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.726588785648346, + "num_tokens": 369533129.0, + "step": 14292 + }, + { + "epoch": 1.5696244234570613, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.914360523223877, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7068482637405396, + "num_tokens": 369566005.0, + "step": 14293 + }, + { + "epoch": 1.5697342411596749, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1972317695617676, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7065701484680176, + "num_tokens": 369594485.0, + "step": 14294 + }, + { + "epoch": 1.5698440588622886, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5203356742858887, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.72669917345047, + "num_tokens": 369616813.0, + "step": 14295 + }, + { + "epoch": 1.5699538765649024, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3015542030334473, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7105470895767212, + "num_tokens": 369644035.0, + "step": 14296 + }, + { + "epoch": 1.570063694267516, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2531018257141113, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.715269923210144, + "num_tokens": 369671675.0, + "step": 14297 + }, + { + "epoch": 1.5701735119701294, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2289371490478516, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7376092672348022, + "num_tokens": 369697141.0, + "step": 14298 + }, + { + "epoch": 1.5702833296727432, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.206608772277832, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7218780517578125, + "num_tokens": 369723377.0, + "step": 14299 + }, + { + "epoch": 1.570393147375357, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5039801597595215, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7158070206642151, + "num_tokens": 369746629.0, + "step": 14300 + }, + { + "epoch": 1.5705029650779707, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.270751714706421, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7249706387519836, + "num_tokens": 369772301.0, + "step": 14301 + }, + { + "epoch": 1.5706127827805842, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4587044715881348, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7215186357498169, + "num_tokens": 369794383.0, + "step": 14302 + }, + { + "epoch": 1.5707226004831978, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.309833526611328, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7225458025932312, + "num_tokens": 369820624.0, + "step": 14303 + }, + { + "epoch": 1.5708324181858115, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2348811626434326, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7103965878486633, + "num_tokens": 369847660.0, + "step": 14304 + }, + { + "epoch": 1.5709422358884253, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4389660358428955, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7254848480224609, + "num_tokens": 369869638.0, + "step": 14305 + }, + { + "epoch": 1.5710520535910388, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1166672706604004, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7463521957397461, + "num_tokens": 369894309.0, + "step": 14306 + }, + { + "epoch": 1.5711618712936526, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2495839595794678, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7225003838539124, + "num_tokens": 369921955.0, + "step": 14307 + }, + { + "epoch": 1.5712716889962661, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2275898456573486, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7172173857688904, + "num_tokens": 369947219.0, + "step": 14308 + }, + { + "epoch": 1.5713815066988799, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2761447429656982, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7456908822059631, + "num_tokens": 369972667.0, + "step": 14309 + }, + { + "epoch": 1.5714913244014936, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2788174152374268, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.726847231388092, + "num_tokens": 369998788.0, + "step": 14310 + }, + { + "epoch": 1.5716011421041072, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9448986053466797, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7558965086936951, + "num_tokens": 370029169.0, + "step": 14311 + }, + { + "epoch": 1.5717109598067207, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3203251361846924, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.719059407711029, + "num_tokens": 370053808.0, + "step": 14312 + }, + { + "epoch": 1.5718207775093345, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3171534538269043, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.720232367515564, + "num_tokens": 370078890.0, + "step": 14313 + }, + { + "epoch": 1.5719305952119482, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3605918884277344, + "learning_rate": 1e-06, + "loss": 1.0477, + "mean_token_accuracy": 0.6840823888778687, + "num_tokens": 370104070.0, + "step": 14314 + }, + { + "epoch": 1.572040412914562, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2780420780181885, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7281476259231567, + "num_tokens": 370128170.0, + "step": 14315 + }, + { + "epoch": 1.5721502306171755, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0044467449188232, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7201050519943237, + "num_tokens": 370159853.0, + "step": 14316 + }, + { + "epoch": 1.572260048319789, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.059304714202881, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7262585163116455, + "num_tokens": 370188294.0, + "step": 14317 + }, + { + "epoch": 1.5723698660224028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2907207012176514, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6969828009605408, + "num_tokens": 370213186.0, + "step": 14318 + }, + { + "epoch": 1.5724796837250166, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0459144115448, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7253844141960144, + "num_tokens": 370242167.0, + "step": 14319 + }, + { + "epoch": 1.57258950142763, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.905735969543457, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7222880721092224, + "num_tokens": 370275462.0, + "step": 14320 + }, + { + "epoch": 1.5726993191302436, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.17984676361084, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7209278345108032, + "num_tokens": 370301238.0, + "step": 14321 + }, + { + "epoch": 1.5728091368328574, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3918912410736084, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7250387668609619, + "num_tokens": 370322420.0, + "step": 14322 + }, + { + "epoch": 1.5729189545354711, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2498116493225098, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.690902829170227, + "num_tokens": 370350201.0, + "step": 14323 + }, + { + "epoch": 1.573028772238085, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1734633445739746, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6984438896179199, + "num_tokens": 370379132.0, + "step": 14324 + }, + { + "epoch": 1.5731385899406984, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.318993330001831, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.6983934640884399, + "num_tokens": 370403156.0, + "step": 14325 + }, + { + "epoch": 1.573248407643312, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3852570056915283, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.726372241973877, + "num_tokens": 370425779.0, + "step": 14326 + }, + { + "epoch": 1.5733582253459257, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1100807189941406, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7212654948234558, + "num_tokens": 370456410.0, + "step": 14327 + }, + { + "epoch": 1.5734680430485395, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1674644947052, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7349157333374023, + "num_tokens": 370482807.0, + "step": 14328 + }, + { + "epoch": 1.5735778607511532, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.629168748855591, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.722934365272522, + "num_tokens": 370504454.0, + "step": 14329 + }, + { + "epoch": 1.5736876784537668, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1334280967712402, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7071255445480347, + "num_tokens": 370531949.0, + "step": 14330 + }, + { + "epoch": 1.5737974961563803, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.18526029586792, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7320724725723267, + "num_tokens": 370557677.0, + "step": 14331 + }, + { + "epoch": 1.573907313858994, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.300996780395508, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6984840035438538, + "num_tokens": 370584060.0, + "step": 14332 + }, + { + "epoch": 1.5740171315616078, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2037928104400635, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7276326417922974, + "num_tokens": 370608927.0, + "step": 14333 + }, + { + "epoch": 1.5741269492642214, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.222311496734619, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7224686741828918, + "num_tokens": 370634430.0, + "step": 14334 + }, + { + "epoch": 1.574236766966835, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4272563457489014, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7306229472160339, + "num_tokens": 370655568.0, + "step": 14335 + }, + { + "epoch": 1.5743465846694487, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2333009243011475, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7133137583732605, + "num_tokens": 370681602.0, + "step": 14336 + }, + { + "epoch": 1.5744564023720624, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9482277631759644, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7183899879455566, + "num_tokens": 370713455.0, + "step": 14337 + }, + { + "epoch": 1.5745662200746762, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2816553115844727, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7161896824836731, + "num_tokens": 370739938.0, + "step": 14338 + }, + { + "epoch": 1.5746760377772897, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3692286014556885, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7163819074630737, + "num_tokens": 370763405.0, + "step": 14339 + }, + { + "epoch": 1.5747858554799032, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.037672281265259, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6887508630752563, + "num_tokens": 370795013.0, + "step": 14340 + }, + { + "epoch": 1.574895673182517, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.05629825592041, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7163536548614502, + "num_tokens": 370821450.0, + "step": 14341 + }, + { + "epoch": 1.5750054908851308, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.308976173400879, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7204399108886719, + "num_tokens": 370846368.0, + "step": 14342 + }, + { + "epoch": 1.5751153085877443, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2364253997802734, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7274014949798584, + "num_tokens": 370872383.0, + "step": 14343 + }, + { + "epoch": 1.575225126290358, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2267160415649414, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7037854194641113, + "num_tokens": 370901133.0, + "step": 14344 + }, + { + "epoch": 1.5753349439929716, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.357055902481079, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.700249195098877, + "num_tokens": 370926893.0, + "step": 14345 + }, + { + "epoch": 1.5754447616955853, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2723474502563477, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7255001068115234, + "num_tokens": 370953501.0, + "step": 14346 + }, + { + "epoch": 1.575554579398199, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0291943550109863, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7297526597976685, + "num_tokens": 370983747.0, + "step": 14347 + }, + { + "epoch": 1.5756643971008126, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0480966567993164, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7174186706542969, + "num_tokens": 371012900.0, + "step": 14348 + }, + { + "epoch": 1.5757742148034262, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3494629859924316, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7172137498855591, + "num_tokens": 371037220.0, + "step": 14349 + }, + { + "epoch": 1.57588403250604, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4760446548461914, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7133796215057373, + "num_tokens": 371059267.0, + "step": 14350 + }, + { + "epoch": 1.5759938502086537, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0658161640167236, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6868075132369995, + "num_tokens": 371090398.0, + "step": 14351 + }, + { + "epoch": 1.5761036679112674, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1568996906280518, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.7016294002532959, + "num_tokens": 371119403.0, + "step": 14352 + }, + { + "epoch": 1.576213485613881, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9113909006118774, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7080036401748657, + "num_tokens": 371154075.0, + "step": 14353 + }, + { + "epoch": 1.5763233033164945, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4456982612609863, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7221428751945496, + "num_tokens": 371176912.0, + "step": 14354 + }, + { + "epoch": 1.5764331210191083, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1697323322296143, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7126416563987732, + "num_tokens": 371202272.0, + "step": 14355 + }, + { + "epoch": 1.576542938721722, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2702112197875977, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7203686833381653, + "num_tokens": 371227268.0, + "step": 14356 + }, + { + "epoch": 1.5766527564243356, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.395246744155884, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7315565347671509, + "num_tokens": 371252192.0, + "step": 14357 + }, + { + "epoch": 1.5767625741269493, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.384159803390503, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7232488393783569, + "num_tokens": 371274291.0, + "step": 14358 + }, + { + "epoch": 1.5768723918295628, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0559589862823486, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6989437341690063, + "num_tokens": 371304575.0, + "step": 14359 + }, + { + "epoch": 1.5769822095321766, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0894556045532227, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7295712828636169, + "num_tokens": 371333681.0, + "step": 14360 + }, + { + "epoch": 1.5770920272347904, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.022211790084839, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7124451398849487, + "num_tokens": 371367450.0, + "step": 14361 + }, + { + "epoch": 1.577201844937404, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1808011531829834, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7137788534164429, + "num_tokens": 371393930.0, + "step": 14362 + }, + { + "epoch": 1.5773116626400174, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4274508953094482, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.720455527305603, + "num_tokens": 371415346.0, + "step": 14363 + }, + { + "epoch": 1.5774214803426312, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.445277452468872, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7070843577384949, + "num_tokens": 371436946.0, + "step": 14364 + }, + { + "epoch": 1.577531298045245, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.138025999069214, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7123013734817505, + "num_tokens": 371466547.0, + "step": 14365 + }, + { + "epoch": 1.5776411157478587, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9627141952514648, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7268081903457642, + "num_tokens": 371498664.0, + "step": 14366 + }, + { + "epoch": 1.5777509334504722, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.115633964538574, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7253617644309998, + "num_tokens": 371526832.0, + "step": 14367 + }, + { + "epoch": 1.5778607511530858, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.023505926132202, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7224593162536621, + "num_tokens": 371556697.0, + "step": 14368 + }, + { + "epoch": 1.5779705688556995, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9277020692825317, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.703450083732605, + "num_tokens": 371587747.0, + "step": 14369 + }, + { + "epoch": 1.5780803865583133, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3351714611053467, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7160488367080688, + "num_tokens": 371611011.0, + "step": 14370 + }, + { + "epoch": 1.5781902042609268, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.208740472793579, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7142043113708496, + "num_tokens": 371636310.0, + "step": 14371 + }, + { + "epoch": 1.5783000219635406, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0865745544433594, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7118417024612427, + "num_tokens": 371664235.0, + "step": 14372 + }, + { + "epoch": 1.5784098396661541, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.312260150909424, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.714009165763855, + "num_tokens": 371689834.0, + "step": 14373 + }, + { + "epoch": 1.5785196573687679, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.8918761014938354, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7067004442214966, + "num_tokens": 371725226.0, + "step": 14374 + }, + { + "epoch": 1.5786294750713816, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.523730993270874, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.740553617477417, + "num_tokens": 371748342.0, + "step": 14375 + }, + { + "epoch": 1.5787392927739952, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2492661476135254, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7207398414611816, + "num_tokens": 371771183.0, + "step": 14376 + }, + { + "epoch": 1.5788491104766087, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.302786111831665, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7162756323814392, + "num_tokens": 371797106.0, + "step": 14377 + }, + { + "epoch": 1.5789589281792225, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2283554077148438, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.70381760597229, + "num_tokens": 371824608.0, + "step": 14378 + }, + { + "epoch": 1.5790687458818362, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.324153184890747, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7143077850341797, + "num_tokens": 371849493.0, + "step": 14379 + }, + { + "epoch": 1.57917856358445, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0701236724853516, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7264046669006348, + "num_tokens": 371877672.0, + "step": 14380 + }, + { + "epoch": 1.5792883812870635, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1165213584899902, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6833744645118713, + "num_tokens": 371909281.0, + "step": 14381 + }, + { + "epoch": 1.579398198989677, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3961710929870605, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7119298577308655, + "num_tokens": 371932126.0, + "step": 14382 + }, + { + "epoch": 1.5795080166922908, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5599958896636963, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7417094707489014, + "num_tokens": 371952465.0, + "step": 14383 + }, + { + "epoch": 1.5796178343949046, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.802999258041382, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7392786741256714, + "num_tokens": 371971853.0, + "step": 14384 + }, + { + "epoch": 1.579727652097518, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2186100482940674, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.6958239078521729, + "num_tokens": 371999889.0, + "step": 14385 + }, + { + "epoch": 1.5798374698001316, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0036778450012207, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7023119926452637, + "num_tokens": 372031413.0, + "step": 14386 + }, + { + "epoch": 1.5799472875027454, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.261458158493042, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7290343046188354, + "num_tokens": 372055163.0, + "step": 14387 + }, + { + "epoch": 1.5800571052053591, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.214242458343506, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7073544263839722, + "num_tokens": 372081104.0, + "step": 14388 + }, + { + "epoch": 1.580166922907973, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.393378973007202, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7288520932197571, + "num_tokens": 372105152.0, + "step": 14389 + }, + { + "epoch": 1.5802767406105864, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.166565179824829, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.714271068572998, + "num_tokens": 372134673.0, + "step": 14390 + }, + { + "epoch": 1.5803865583132, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4220056533813477, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7213125824928284, + "num_tokens": 372157655.0, + "step": 14391 + }, + { + "epoch": 1.5804963760158137, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.143404483795166, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7334622144699097, + "num_tokens": 372181746.0, + "step": 14392 + }, + { + "epoch": 1.5806061937184275, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.229097843170166, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7058107852935791, + "num_tokens": 372209356.0, + "step": 14393 + }, + { + "epoch": 1.5807160114210412, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.365159034729004, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6970924139022827, + "num_tokens": 372233574.0, + "step": 14394 + }, + { + "epoch": 1.5808258291236548, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.240920066833496, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7036579847335815, + "num_tokens": 372261343.0, + "step": 14395 + }, + { + "epoch": 1.5809356468262683, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3757030963897705, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7168276309967041, + "num_tokens": 372284609.0, + "step": 14396 + }, + { + "epoch": 1.581045464528882, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2449872493743896, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7238370180130005, + "num_tokens": 372309407.0, + "step": 14397 + }, + { + "epoch": 1.5811552822314958, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3821449279785156, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7316190004348755, + "num_tokens": 372330438.0, + "step": 14398 + }, + { + "epoch": 1.5812650999341094, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3161041736602783, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.724439263343811, + "num_tokens": 372358343.0, + "step": 14399 + }, + { + "epoch": 1.5813749176367229, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5528457164764404, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7446990013122559, + "num_tokens": 372377730.0, + "step": 14400 + }, + { + "epoch": 1.5814847353393366, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9995390176773071, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7066744565963745, + "num_tokens": 372410573.0, + "step": 14401 + }, + { + "epoch": 1.5815945530419504, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.127890110015869, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.7010561227798462, + "num_tokens": 372438940.0, + "step": 14402 + }, + { + "epoch": 1.5817043707445642, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3087522983551025, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7095475792884827, + "num_tokens": 372462985.0, + "step": 14403 + }, + { + "epoch": 1.5818141884471777, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.323460340499878, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7162168025970459, + "num_tokens": 372488502.0, + "step": 14404 + }, + { + "epoch": 1.5819240061497912, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.288010835647583, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7110083699226379, + "num_tokens": 372514893.0, + "step": 14405 + }, + { + "epoch": 1.582033823852405, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4009642601013184, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7134858965873718, + "num_tokens": 372538957.0, + "step": 14406 + }, + { + "epoch": 1.5821436415550187, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.066561698913574, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7043803930282593, + "num_tokens": 372568819.0, + "step": 14407 + }, + { + "epoch": 1.5822534592576323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.16951847076416, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6957809925079346, + "num_tokens": 372599091.0, + "step": 14408 + }, + { + "epoch": 1.582363276960246, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1690614223480225, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7399290800094604, + "num_tokens": 372627719.0, + "step": 14409 + }, + { + "epoch": 1.5824730946628596, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4841105937957764, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7053973078727722, + "num_tokens": 372650903.0, + "step": 14410 + }, + { + "epoch": 1.5825829123654733, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2595627307891846, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7155004739761353, + "num_tokens": 372678165.0, + "step": 14411 + }, + { + "epoch": 1.582692730068087, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9304783344268799, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7033297419548035, + "num_tokens": 372713150.0, + "step": 14412 + }, + { + "epoch": 1.5828025477707006, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2085509300231934, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7108755111694336, + "num_tokens": 372739846.0, + "step": 14413 + }, + { + "epoch": 1.5829123654733142, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2691609859466553, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7101857662200928, + "num_tokens": 372765815.0, + "step": 14414 + }, + { + "epoch": 1.583022183175928, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0171496868133545, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.6991982460021973, + "num_tokens": 372797191.0, + "step": 14415 + }, + { + "epoch": 1.5831320008785417, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1453497409820557, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7269142866134644, + "num_tokens": 372824195.0, + "step": 14416 + }, + { + "epoch": 1.5832418185811554, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4046976566314697, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7159581780433655, + "num_tokens": 372847841.0, + "step": 14417 + }, + { + "epoch": 1.583351636283769, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5198206901550293, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7293025851249695, + "num_tokens": 372867002.0, + "step": 14418 + }, + { + "epoch": 1.5834614539863825, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.290343761444092, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7139178514480591, + "num_tokens": 372892138.0, + "step": 14419 + }, + { + "epoch": 1.5835712716889963, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1411092281341553, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7153322696685791, + "num_tokens": 372917333.0, + "step": 14420 + }, + { + "epoch": 1.58368108939161, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.24528169631958, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7276825904846191, + "num_tokens": 372941779.0, + "step": 14421 + }, + { + "epoch": 1.5837909070942235, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4457802772521973, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7197122573852539, + "num_tokens": 372962844.0, + "step": 14422 + }, + { + "epoch": 1.5839007247968373, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.14920973777771, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7074609994888306, + "num_tokens": 372989413.0, + "step": 14423 + }, + { + "epoch": 1.5840105424994508, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3634471893310547, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7229324579238892, + "num_tokens": 373014865.0, + "step": 14424 + }, + { + "epoch": 1.5841203602020646, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1826066970825195, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7041676044464111, + "num_tokens": 373043360.0, + "step": 14425 + }, + { + "epoch": 1.5842301779046783, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0878212451934814, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7133060693740845, + "num_tokens": 373071402.0, + "step": 14426 + }, + { + "epoch": 1.5843399956072919, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.531153917312622, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.71646648645401, + "num_tokens": 373092339.0, + "step": 14427 + }, + { + "epoch": 1.5844498133099054, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.448336601257324, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7131907939910889, + "num_tokens": 373113931.0, + "step": 14428 + }, + { + "epoch": 1.5845596310125192, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3271796703338623, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7335516214370728, + "num_tokens": 373137454.0, + "step": 14429 + }, + { + "epoch": 1.584669448715133, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.151984691619873, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.701168417930603, + "num_tokens": 373163741.0, + "step": 14430 + }, + { + "epoch": 1.5847792664177467, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.177788734436035, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.712693452835083, + "num_tokens": 373191778.0, + "step": 14431 + }, + { + "epoch": 1.5848890841203602, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3618500232696533, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7106807231903076, + "num_tokens": 373215503.0, + "step": 14432 + }, + { + "epoch": 1.5849989018229738, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3566081523895264, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7203079462051392, + "num_tokens": 373239727.0, + "step": 14433 + }, + { + "epoch": 1.5851087195255875, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2554125785827637, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6907522678375244, + "num_tokens": 373268032.0, + "step": 14434 + }, + { + "epoch": 1.5852185372282013, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.160923719406128, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7150943279266357, + "num_tokens": 373295587.0, + "step": 14435 + }, + { + "epoch": 1.5853283549308148, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5461525917053223, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7129640579223633, + "num_tokens": 373316872.0, + "step": 14436 + }, + { + "epoch": 1.5854381726334283, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.427525520324707, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7379597425460815, + "num_tokens": 373337786.0, + "step": 14437 + }, + { + "epoch": 1.585547990336042, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.491464614868164, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7171773314476013, + "num_tokens": 373359915.0, + "step": 14438 + }, + { + "epoch": 1.5856578080386559, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3456056118011475, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7215442061424255, + "num_tokens": 373384787.0, + "step": 14439 + }, + { + "epoch": 1.5857676257412696, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1331088542938232, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7236657738685608, + "num_tokens": 373414612.0, + "step": 14440 + }, + { + "epoch": 1.5858774434438832, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3462772369384766, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7208551168441772, + "num_tokens": 373439049.0, + "step": 14441 + }, + { + "epoch": 1.5859872611464967, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9966650009155273, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6892557144165039, + "num_tokens": 373474120.0, + "step": 14442 + }, + { + "epoch": 1.5860970788491104, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3266587257385254, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7405150532722473, + "num_tokens": 373497659.0, + "step": 14443 + }, + { + "epoch": 1.5862068965517242, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.217090606689453, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7380446791648865, + "num_tokens": 373522404.0, + "step": 14444 + }, + { + "epoch": 1.586316714254338, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.160606861114502, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7350921630859375, + "num_tokens": 373549575.0, + "step": 14445 + }, + { + "epoch": 1.5864265319569515, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.423567533493042, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7404896020889282, + "num_tokens": 373572238.0, + "step": 14446 + }, + { + "epoch": 1.586536349659565, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3727049827575684, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7186011075973511, + "num_tokens": 373597048.0, + "step": 14447 + }, + { + "epoch": 1.5866461673621788, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.333911895751953, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7208518981933594, + "num_tokens": 373620945.0, + "step": 14448 + }, + { + "epoch": 1.5867559850647925, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4344775676727295, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7138638496398926, + "num_tokens": 373644170.0, + "step": 14449 + }, + { + "epoch": 1.586865802767406, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5015976428985596, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7150156497955322, + "num_tokens": 373665553.0, + "step": 14450 + }, + { + "epoch": 1.5869756204700196, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.253404378890991, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7124086618423462, + "num_tokens": 373693133.0, + "step": 14451 + }, + { + "epoch": 1.5870854381726334, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.326834201812744, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7176122665405273, + "num_tokens": 373716781.0, + "step": 14452 + }, + { + "epoch": 1.5871952558752471, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4099981784820557, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7221333980560303, + "num_tokens": 373742232.0, + "step": 14453 + }, + { + "epoch": 1.5873050735778609, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2343409061431885, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7168969511985779, + "num_tokens": 373768072.0, + "step": 14454 + }, + { + "epoch": 1.5874148912804744, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2202303409576416, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.723711371421814, + "num_tokens": 373794979.0, + "step": 14455 + }, + { + "epoch": 1.587524708983088, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.373251438140869, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7264233827590942, + "num_tokens": 373817546.0, + "step": 14456 + }, + { + "epoch": 1.5876345266857017, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1313939094543457, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7196259498596191, + "num_tokens": 373845958.0, + "step": 14457 + }, + { + "epoch": 1.5877443443883155, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.42814302444458, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7294942140579224, + "num_tokens": 373869453.0, + "step": 14458 + }, + { + "epoch": 1.5878541620909292, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2536637783050537, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7166427373886108, + "num_tokens": 373894136.0, + "step": 14459 + }, + { + "epoch": 1.5879639797935428, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5577173233032227, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7138252258300781, + "num_tokens": 373914939.0, + "step": 14460 + }, + { + "epoch": 1.5880737974961563, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.140197515487671, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6929701566696167, + "num_tokens": 373942048.0, + "step": 14461 + }, + { + "epoch": 1.58818361519877, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1494338512420654, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7181958556175232, + "num_tokens": 373971014.0, + "step": 14462 + }, + { + "epoch": 1.5882934329013838, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.254875421524048, + "learning_rate": 1e-06, + "loss": 0.766, + "mean_token_accuracy": 0.7601144313812256, + "num_tokens": 373994789.0, + "step": 14463 + }, + { + "epoch": 1.5884032506039973, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3854494094848633, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7278690934181213, + "num_tokens": 374016747.0, + "step": 14464 + }, + { + "epoch": 1.5885130683066109, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.484987497329712, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7298084497451782, + "num_tokens": 374039477.0, + "step": 14465 + }, + { + "epoch": 1.5886228860092246, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1625783443450928, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7085906863212585, + "num_tokens": 374068549.0, + "step": 14466 + }, + { + "epoch": 1.5887327037118384, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4901490211486816, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7324704527854919, + "num_tokens": 374090115.0, + "step": 14467 + }, + { + "epoch": 1.5888425214144521, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3073649406433105, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7085572481155396, + "num_tokens": 374114433.0, + "step": 14468 + }, + { + "epoch": 1.5889523391170657, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1506078243255615, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7131284475326538, + "num_tokens": 374144386.0, + "step": 14469 + }, + { + "epoch": 1.5890621568196792, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3989081382751465, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7330028414726257, + "num_tokens": 374166726.0, + "step": 14470 + }, + { + "epoch": 1.589171974522293, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.105046033859253, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7134543061256409, + "num_tokens": 374197369.0, + "step": 14471 + }, + { + "epoch": 1.5892817922249067, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.374565601348877, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7141619920730591, + "num_tokens": 374220021.0, + "step": 14472 + }, + { + "epoch": 1.5893916099275203, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0210886001586914, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7044434547424316, + "num_tokens": 374252238.0, + "step": 14473 + }, + { + "epoch": 1.589501427630134, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.190990686416626, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7154117226600647, + "num_tokens": 374279155.0, + "step": 14474 + }, + { + "epoch": 1.5896112453327476, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1255130767822266, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.719452977180481, + "num_tokens": 374308634.0, + "step": 14475 + }, + { + "epoch": 1.5897210630353613, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4272260665893555, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7189230918884277, + "num_tokens": 374331359.0, + "step": 14476 + }, + { + "epoch": 1.589830880737975, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.546633005142212, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.6960839629173279, + "num_tokens": 374355995.0, + "step": 14477 + }, + { + "epoch": 1.5899406984405886, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.440497875213623, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.721571683883667, + "num_tokens": 374377534.0, + "step": 14478 + }, + { + "epoch": 1.5900505161432021, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.226155996322632, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.723846435546875, + "num_tokens": 374403849.0, + "step": 14479 + }, + { + "epoch": 1.590160333845816, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2332217693328857, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7361078262329102, + "num_tokens": 374429684.0, + "step": 14480 + }, + { + "epoch": 1.5902701515484297, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1962943077087402, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7274978756904602, + "num_tokens": 374455969.0, + "step": 14481 + }, + { + "epoch": 1.5903799692510434, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.299417495727539, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7211242914199829, + "num_tokens": 374481316.0, + "step": 14482 + }, + { + "epoch": 1.590489786953657, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.225215196609497, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7045122385025024, + "num_tokens": 374505819.0, + "step": 14483 + }, + { + "epoch": 1.5905996046562705, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1424758434295654, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.755810022354126, + "num_tokens": 374529943.0, + "step": 14484 + }, + { + "epoch": 1.5907094223588842, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5656542778015137, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7274192571640015, + "num_tokens": 374551720.0, + "step": 14485 + }, + { + "epoch": 1.590819240061498, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.10990571975708, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.710639476776123, + "num_tokens": 374580820.0, + "step": 14486 + }, + { + "epoch": 1.5909290577641115, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4477481842041016, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7271644473075867, + "num_tokens": 374603548.0, + "step": 14487 + }, + { + "epoch": 1.5910388754667253, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2407562732696533, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7106795310974121, + "num_tokens": 374629400.0, + "step": 14488 + }, + { + "epoch": 1.5911486931693388, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3792295455932617, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7179019451141357, + "num_tokens": 374651432.0, + "step": 14489 + }, + { + "epoch": 1.5912585108719526, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9295097589492798, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7122896909713745, + "num_tokens": 374684310.0, + "step": 14490 + }, + { + "epoch": 1.5913683285745663, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.263190746307373, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.706645131111145, + "num_tokens": 374709123.0, + "step": 14491 + }, + { + "epoch": 1.5914781462771799, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.226746082305908, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.719467043876648, + "num_tokens": 374737090.0, + "step": 14492 + }, + { + "epoch": 1.5915879639797934, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2290217876434326, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7188475131988525, + "num_tokens": 374762696.0, + "step": 14493 + }, + { + "epoch": 1.5916977816824072, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3700149059295654, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7255906462669373, + "num_tokens": 374787539.0, + "step": 14494 + }, + { + "epoch": 1.591807599385021, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1084787845611572, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.707729697227478, + "num_tokens": 374815548.0, + "step": 14495 + }, + { + "epoch": 1.5919174170876347, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.182433843612671, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7067195773124695, + "num_tokens": 374842855.0, + "step": 14496 + }, + { + "epoch": 1.5920272347902482, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9654937982559204, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7115857601165771, + "num_tokens": 374874927.0, + "step": 14497 + }, + { + "epoch": 1.5921370524928617, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4595816135406494, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.7426919341087341, + "num_tokens": 374895034.0, + "step": 14498 + }, + { + "epoch": 1.5922468701954755, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3471715450286865, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7096278071403503, + "num_tokens": 374919496.0, + "step": 14499 + }, + { + "epoch": 1.5923566878980893, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.36629319190979, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7131717205047607, + "num_tokens": 374943244.0, + "step": 14500 + }, + { + "epoch": 1.5924665056007028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.170788526535034, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7140473127365112, + "num_tokens": 374971146.0, + "step": 14501 + }, + { + "epoch": 1.5925763233033163, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1229779720306396, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7248122692108154, + "num_tokens": 374997752.0, + "step": 14502 + }, + { + "epoch": 1.59268614100593, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2202885150909424, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7038076519966125, + "num_tokens": 375025536.0, + "step": 14503 + }, + { + "epoch": 1.5927959587085438, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3760292530059814, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7233227491378784, + "num_tokens": 375049295.0, + "step": 14504 + }, + { + "epoch": 1.5929057764111576, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2510645389556885, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7339794635772705, + "num_tokens": 375073356.0, + "step": 14505 + }, + { + "epoch": 1.5930155941137711, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3224141597747803, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7344052791595459, + "num_tokens": 375097520.0, + "step": 14506 + }, + { + "epoch": 1.5931254118163847, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1476683616638184, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7157177925109863, + "num_tokens": 375124237.0, + "step": 14507 + }, + { + "epoch": 1.5932352295189984, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2664411067962646, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7478983998298645, + "num_tokens": 375148049.0, + "step": 14508 + }, + { + "epoch": 1.5933450472216122, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1160097122192383, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7177075147628784, + "num_tokens": 375177807.0, + "step": 14509 + }, + { + "epoch": 1.593454864924226, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2834298610687256, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7360668182373047, + "num_tokens": 375201645.0, + "step": 14510 + }, + { + "epoch": 1.5935646826268395, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0873491764068604, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7122694253921509, + "num_tokens": 375229559.0, + "step": 14511 + }, + { + "epoch": 1.593674500329453, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.070807695388794, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.711614727973938, + "num_tokens": 375259333.0, + "step": 14512 + }, + { + "epoch": 1.5937843180320668, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.116276979446411, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6964150667190552, + "num_tokens": 375288550.0, + "step": 14513 + }, + { + "epoch": 1.5938941357346805, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9410392045974731, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7345333695411682, + "num_tokens": 375320715.0, + "step": 14514 + }, + { + "epoch": 1.594003953437294, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.170853614807129, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6934593915939331, + "num_tokens": 375350974.0, + "step": 14515 + }, + { + "epoch": 1.5941137711399076, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.071368932723999, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7070345878601074, + "num_tokens": 375380124.0, + "step": 14516 + }, + { + "epoch": 1.5942235888425214, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4068410396575928, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7396237850189209, + "num_tokens": 375403688.0, + "step": 14517 + }, + { + "epoch": 1.5943334065451351, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.472911834716797, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7285728454589844, + "num_tokens": 375424204.0, + "step": 14518 + }, + { + "epoch": 1.5944432242477489, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0701112747192383, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6975206136703491, + "num_tokens": 375454822.0, + "step": 14519 + }, + { + "epoch": 1.5945530419503624, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.267110586166382, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7111173868179321, + "num_tokens": 375481295.0, + "step": 14520 + }, + { + "epoch": 1.594662859652976, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.253214120864868, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7261934280395508, + "num_tokens": 375505553.0, + "step": 14521 + }, + { + "epoch": 1.5947726773555897, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3538706302642822, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7213943600654602, + "num_tokens": 375529493.0, + "step": 14522 + }, + { + "epoch": 1.5948824950582035, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.022461414337158, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7173300981521606, + "num_tokens": 375560012.0, + "step": 14523 + }, + { + "epoch": 1.5949923127608172, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3019063472747803, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7217051982879639, + "num_tokens": 375584353.0, + "step": 14524 + }, + { + "epoch": 1.5951021304634307, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1918816566467285, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7268975377082825, + "num_tokens": 375610895.0, + "step": 14525 + }, + { + "epoch": 1.5952119481660443, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4379169940948486, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7287544012069702, + "num_tokens": 375632870.0, + "step": 14526 + }, + { + "epoch": 1.595321765868658, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.236771583557129, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.708764910697937, + "num_tokens": 375658316.0, + "step": 14527 + }, + { + "epoch": 1.5954315835712718, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.176621437072754, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7121095657348633, + "num_tokens": 375684916.0, + "step": 14528 + }, + { + "epoch": 1.5955414012738853, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.217787742614746, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7155236601829529, + "num_tokens": 375711341.0, + "step": 14529 + }, + { + "epoch": 1.5956512189764989, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.193917989730835, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6999082565307617, + "num_tokens": 375738416.0, + "step": 14530 + }, + { + "epoch": 1.5957610366791126, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4445719718933105, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7425482869148254, + "num_tokens": 375760713.0, + "step": 14531 + }, + { + "epoch": 1.5958708543817264, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1622467041015625, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7423558831214905, + "num_tokens": 375786840.0, + "step": 14532 + }, + { + "epoch": 1.5959806720843401, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3642079830169678, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7182464003562927, + "num_tokens": 375811614.0, + "step": 14533 + }, + { + "epoch": 1.5960904897869537, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.244765520095825, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7180798053741455, + "num_tokens": 375838020.0, + "step": 14534 + }, + { + "epoch": 1.5962003074895672, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3664863109588623, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7189604043960571, + "num_tokens": 375861796.0, + "step": 14535 + }, + { + "epoch": 1.596310125192181, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2968974113464355, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7212519645690918, + "num_tokens": 375887156.0, + "step": 14536 + }, + { + "epoch": 1.5964199428947947, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1946582794189453, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7383391261100769, + "num_tokens": 375911343.0, + "step": 14537 + }, + { + "epoch": 1.5965297605974083, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.323927402496338, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6966776847839355, + "num_tokens": 375936167.0, + "step": 14538 + }, + { + "epoch": 1.596639578300022, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9729893207550049, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7154595851898193, + "num_tokens": 375967462.0, + "step": 14539 + }, + { + "epoch": 1.5967493960026355, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2730419635772705, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7313560247421265, + "num_tokens": 375990868.0, + "step": 14540 + }, + { + "epoch": 1.5968592137052493, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3089568614959717, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7043068408966064, + "num_tokens": 376016493.0, + "step": 14541 + }, + { + "epoch": 1.596969031407863, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5297739505767822, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.72063148021698, + "num_tokens": 376039072.0, + "step": 14542 + }, + { + "epoch": 1.5970788491104766, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.15763521194458, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7220208644866943, + "num_tokens": 376067183.0, + "step": 14543 + }, + { + "epoch": 1.5971886668130901, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.339496374130249, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7294045686721802, + "num_tokens": 376091007.0, + "step": 14544 + }, + { + "epoch": 1.5972984845157039, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5380616188049316, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7423995733261108, + "num_tokens": 376110858.0, + "step": 14545 + }, + { + "epoch": 1.5974083022183176, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4333901405334473, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7445255517959595, + "num_tokens": 376131065.0, + "step": 14546 + }, + { + "epoch": 1.5975181199209314, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2483980655670166, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7147359848022461, + "num_tokens": 376156691.0, + "step": 14547 + }, + { + "epoch": 1.597627937623545, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3344058990478516, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7103667259216309, + "num_tokens": 376180701.0, + "step": 14548 + }, + { + "epoch": 1.5977377553261585, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.175079822540283, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7106658220291138, + "num_tokens": 376209886.0, + "step": 14549 + }, + { + "epoch": 1.5978475730287722, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2696173191070557, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7348369359970093, + "num_tokens": 376234719.0, + "step": 14550 + }, + { + "epoch": 1.597957390731386, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.49088978767395, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7287397980690002, + "num_tokens": 376255215.0, + "step": 14551 + }, + { + "epoch": 1.5980672084339995, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.20585036277771, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7186146974563599, + "num_tokens": 376280985.0, + "step": 14552 + }, + { + "epoch": 1.5981770261366133, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.404442310333252, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7211117148399353, + "num_tokens": 376305600.0, + "step": 14553 + }, + { + "epoch": 1.5982868438392268, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4690492153167725, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7161245346069336, + "num_tokens": 376329147.0, + "step": 14554 + }, + { + "epoch": 1.5983966615418406, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.239816188812256, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7178517580032349, + "num_tokens": 376353320.0, + "step": 14555 + }, + { + "epoch": 1.5985064792444543, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.166457414627075, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7039331793785095, + "num_tokens": 376380025.0, + "step": 14556 + }, + { + "epoch": 1.5986162969470679, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0070621967315674, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7112722396850586, + "num_tokens": 376411452.0, + "step": 14557 + }, + { + "epoch": 1.5987261146496814, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9690399169921875, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.704150378704071, + "num_tokens": 376444923.0, + "step": 14558 + }, + { + "epoch": 1.5988359323522952, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4312942028045654, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7153032422065735, + "num_tokens": 376467832.0, + "step": 14559 + }, + { + "epoch": 1.598945750054909, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2666380405426025, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7163342833518982, + "num_tokens": 376492742.0, + "step": 14560 + }, + { + "epoch": 1.5990555677575227, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.01025652885437, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7198731899261475, + "num_tokens": 376525201.0, + "step": 14561 + }, + { + "epoch": 1.5991653854601362, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3790295124053955, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7312710881233215, + "num_tokens": 376549642.0, + "step": 14562 + }, + { + "epoch": 1.5992752031627497, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.82258939743042, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7131170034408569, + "num_tokens": 376568606.0, + "step": 14563 + }, + { + "epoch": 1.5993850208653635, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3412013053894043, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7026797533035278, + "num_tokens": 376593160.0, + "step": 14564 + }, + { + "epoch": 1.5994948385679773, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.342952013015747, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7394063472747803, + "num_tokens": 376615368.0, + "step": 14565 + }, + { + "epoch": 1.5996046562705908, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.399003267288208, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6956788897514343, + "num_tokens": 376639264.0, + "step": 14566 + }, + { + "epoch": 1.5997144739732043, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0686283111572266, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.7075458765029907, + "num_tokens": 376669428.0, + "step": 14567 + }, + { + "epoch": 1.599824291675818, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.13167142868042, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7373347878456116, + "num_tokens": 376694719.0, + "step": 14568 + }, + { + "epoch": 1.5999341093784318, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3135082721710205, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7163315415382385, + "num_tokens": 376718948.0, + "step": 14569 + }, + { + "epoch": 1.6000439270810456, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1678919792175293, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.71889728307724, + "num_tokens": 376747031.0, + "step": 14570 + }, + { + "epoch": 1.6001537447836591, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3718502521514893, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7334908246994019, + "num_tokens": 376770942.0, + "step": 14571 + }, + { + "epoch": 1.6002635624862727, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.126385450363159, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7221676111221313, + "num_tokens": 376800528.0, + "step": 14572 + }, + { + "epoch": 1.6003733801888864, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.94550621509552, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7136245369911194, + "num_tokens": 376833529.0, + "step": 14573 + }, + { + "epoch": 1.6004831978915002, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2614657878875732, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7247133255004883, + "num_tokens": 376856645.0, + "step": 14574 + }, + { + "epoch": 1.600593015594114, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.049895763397217, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7215129137039185, + "num_tokens": 376888611.0, + "step": 14575 + }, + { + "epoch": 1.6007028332967275, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3106937408447266, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7185045480728149, + "num_tokens": 376912936.0, + "step": 14576 + }, + { + "epoch": 1.600812650999341, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.261406898498535, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7087913155555725, + "num_tokens": 376937269.0, + "step": 14577 + }, + { + "epoch": 1.6009224687019548, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3451995849609375, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.725007176399231, + "num_tokens": 376960702.0, + "step": 14578 + }, + { + "epoch": 1.6010322864045685, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.104846954345703, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7246730327606201, + "num_tokens": 376988619.0, + "step": 14579 + }, + { + "epoch": 1.601142104107182, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9881876707077026, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6885584592819214, + "num_tokens": 377020953.0, + "step": 14580 + }, + { + "epoch": 1.6012519218097956, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3316619396209717, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.726242184638977, + "num_tokens": 377044028.0, + "step": 14581 + }, + { + "epoch": 1.6013617395124093, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1877965927124023, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7126846313476562, + "num_tokens": 377076356.0, + "step": 14582 + }, + { + "epoch": 1.601471557215023, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2987656593322754, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7240077257156372, + "num_tokens": 377100286.0, + "step": 14583 + }, + { + "epoch": 1.6015813749176369, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1069326400756836, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7275680303573608, + "num_tokens": 377129006.0, + "step": 14584 + }, + { + "epoch": 1.6016911926202504, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.7658724784851074, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7387334108352661, + "num_tokens": 377147317.0, + "step": 14585 + }, + { + "epoch": 1.601801010322864, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2703089714050293, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7096884846687317, + "num_tokens": 377174880.0, + "step": 14586 + }, + { + "epoch": 1.6019108280254777, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4481165409088135, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7195113897323608, + "num_tokens": 377197324.0, + "step": 14587 + }, + { + "epoch": 1.6020206457280914, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.522721290588379, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7144399881362915, + "num_tokens": 377218568.0, + "step": 14588 + }, + { + "epoch": 1.602130463430705, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.301464796066284, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7197338938713074, + "num_tokens": 377242710.0, + "step": 14589 + }, + { + "epoch": 1.6022402811333187, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.906676173210144, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7166315317153931, + "num_tokens": 377282627.0, + "step": 14590 + }, + { + "epoch": 1.6023500988359323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4684886932373047, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7332162261009216, + "num_tokens": 377304607.0, + "step": 14591 + }, + { + "epoch": 1.602459916538546, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6606252193450928, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.738582968711853, + "num_tokens": 377324519.0, + "step": 14592 + }, + { + "epoch": 1.6025697342411598, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.847961664199829, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6895390748977661, + "num_tokens": 377360265.0, + "step": 14593 + }, + { + "epoch": 1.6026795519437733, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.8058319091796875, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.720752477645874, + "num_tokens": 377377975.0, + "step": 14594 + }, + { + "epoch": 1.6027893696463869, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.09382963180542, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7233612537384033, + "num_tokens": 377407738.0, + "step": 14595 + }, + { + "epoch": 1.6028991873490006, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.375993251800537, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7223371267318726, + "num_tokens": 377429190.0, + "step": 14596 + }, + { + "epoch": 1.6030090050516144, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3868725299835205, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7403007745742798, + "num_tokens": 377451639.0, + "step": 14597 + }, + { + "epoch": 1.6031188227542281, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3787434101104736, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.711252748966217, + "num_tokens": 377477193.0, + "step": 14598 + }, + { + "epoch": 1.6032286404568417, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4263105392456055, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7222697138786316, + "num_tokens": 377503375.0, + "step": 14599 + }, + { + "epoch": 1.6033384581594552, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5007171630859375, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7279693484306335, + "num_tokens": 377523637.0, + "step": 14600 + }, + { + "epoch": 1.603448275862069, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1976993083953857, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.698323667049408, + "num_tokens": 377553443.0, + "step": 14601 + }, + { + "epoch": 1.6035580935646827, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9807319641113281, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7543967962265015, + "num_tokens": 377582274.0, + "step": 14602 + }, + { + "epoch": 1.6036679112672962, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.444326162338257, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.709395170211792, + "num_tokens": 377604082.0, + "step": 14603 + }, + { + "epoch": 1.60377772896991, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1273696422576904, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.739275336265564, + "num_tokens": 377629865.0, + "step": 14604 + }, + { + "epoch": 1.6038875466725235, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3987088203430176, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7402746677398682, + "num_tokens": 377651745.0, + "step": 14605 + }, + { + "epoch": 1.6039973643751373, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0097455978393555, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7027031183242798, + "num_tokens": 377683487.0, + "step": 14606 + }, + { + "epoch": 1.604107182077751, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3258414268493652, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.6995453834533691, + "num_tokens": 377708915.0, + "step": 14607 + }, + { + "epoch": 1.6042169997803646, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1729958057403564, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.720683753490448, + "num_tokens": 377733857.0, + "step": 14608 + }, + { + "epoch": 1.6043268174829781, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5650103092193604, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7461607456207275, + "num_tokens": 377752263.0, + "step": 14609 + }, + { + "epoch": 1.6044366351855919, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3030574321746826, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7192872762680054, + "num_tokens": 377775949.0, + "step": 14610 + }, + { + "epoch": 1.6045464528882056, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4595210552215576, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7351309657096863, + "num_tokens": 377797937.0, + "step": 14611 + }, + { + "epoch": 1.6046562705908194, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1620054244995117, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7227598428726196, + "num_tokens": 377828982.0, + "step": 14612 + }, + { + "epoch": 1.604766088293433, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9290741682052612, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7037181854248047, + "num_tokens": 377863360.0, + "step": 14613 + }, + { + "epoch": 1.6048759059960465, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1414599418640137, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7014992237091064, + "num_tokens": 377891909.0, + "step": 14614 + }, + { + "epoch": 1.6049857236986602, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2812907695770264, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7201692461967468, + "num_tokens": 377916869.0, + "step": 14615 + }, + { + "epoch": 1.605095541401274, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3661744594573975, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6894839406013489, + "num_tokens": 377941425.0, + "step": 14616 + }, + { + "epoch": 1.6052053591038875, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.234254837036133, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7351148128509521, + "num_tokens": 377966567.0, + "step": 14617 + }, + { + "epoch": 1.605315176806501, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2464075088500977, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.6961110830307007, + "num_tokens": 377993994.0, + "step": 14618 + }, + { + "epoch": 1.6054249945091148, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3516221046447754, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7227463722229004, + "num_tokens": 378018392.0, + "step": 14619 + }, + { + "epoch": 1.6055348122117286, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.560960531234741, + "learning_rate": 1e-06, + "loss": 0.7511, + "mean_token_accuracy": 0.7571114301681519, + "num_tokens": 378038143.0, + "step": 14620 + }, + { + "epoch": 1.6056446299143423, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.288334846496582, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7102762460708618, + "num_tokens": 378063431.0, + "step": 14621 + }, + { + "epoch": 1.6057544476169558, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2607288360595703, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.746387779712677, + "num_tokens": 378086320.0, + "step": 14622 + }, + { + "epoch": 1.6058642653195694, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6189184188842773, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7203975319862366, + "num_tokens": 378107634.0, + "step": 14623 + }, + { + "epoch": 1.6059740830221831, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.326631784439087, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7229930758476257, + "num_tokens": 378134425.0, + "step": 14624 + }, + { + "epoch": 1.606083900724797, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.019663095474243, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.6997266411781311, + "num_tokens": 378163654.0, + "step": 14625 + }, + { + "epoch": 1.6061937184274107, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.179898977279663, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7207905054092407, + "num_tokens": 378190317.0, + "step": 14626 + }, + { + "epoch": 1.6063035361300242, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.165493965148926, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6920315623283386, + "num_tokens": 378219657.0, + "step": 14627 + }, + { + "epoch": 1.6064133538326377, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5859556198120117, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7253508567810059, + "num_tokens": 378239882.0, + "step": 14628 + }, + { + "epoch": 1.6065231715352515, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1904051303863525, + "learning_rate": 1e-06, + "loss": 1.0316, + "mean_token_accuracy": 0.693301796913147, + "num_tokens": 378267995.0, + "step": 14629 + }, + { + "epoch": 1.6066329892378652, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.364828586578369, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7124276161193848, + "num_tokens": 378291931.0, + "step": 14630 + }, + { + "epoch": 1.6067428069404788, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.442934989929199, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.713203489780426, + "num_tokens": 378315418.0, + "step": 14631 + }, + { + "epoch": 1.6068526246430923, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.378915786743164, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7270786166191101, + "num_tokens": 378336854.0, + "step": 14632 + }, + { + "epoch": 1.606962442345706, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.079057216644287, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7148648500442505, + "num_tokens": 378365320.0, + "step": 14633 + }, + { + "epoch": 1.6070722600483198, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6403961181640625, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7279447317123413, + "num_tokens": 378387196.0, + "step": 14634 + }, + { + "epoch": 1.6071820777509336, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.668381690979004, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7378756999969482, + "num_tokens": 378406099.0, + "step": 14635 + }, + { + "epoch": 1.6072918954535471, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0590739250183105, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6969183683395386, + "num_tokens": 378434516.0, + "step": 14636 + }, + { + "epoch": 1.6074017131561606, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0521559715270996, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7331660985946655, + "num_tokens": 378463211.0, + "step": 14637 + }, + { + "epoch": 1.6075115308587744, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.389626979827881, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7077544331550598, + "num_tokens": 378487206.0, + "step": 14638 + }, + { + "epoch": 1.6076213485613882, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3413562774658203, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7188451886177063, + "num_tokens": 378509810.0, + "step": 14639 + }, + { + "epoch": 1.607731166264002, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2996208667755127, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.711138904094696, + "num_tokens": 378535236.0, + "step": 14640 + }, + { + "epoch": 1.6078409839666155, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1248059272766113, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7170852422714233, + "num_tokens": 378564805.0, + "step": 14641 + }, + { + "epoch": 1.607950801669229, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2735393047332764, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7143633365631104, + "num_tokens": 378590294.0, + "step": 14642 + }, + { + "epoch": 1.6080606193718427, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2622244358062744, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6967485547065735, + "num_tokens": 378616209.0, + "step": 14643 + }, + { + "epoch": 1.6081704370744565, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.229377031326294, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7054370641708374, + "num_tokens": 378642574.0, + "step": 14644 + }, + { + "epoch": 1.60828025477707, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2471706867218018, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7160561680793762, + "num_tokens": 378667975.0, + "step": 14645 + }, + { + "epoch": 1.6083900724796836, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1570465564727783, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7312231659889221, + "num_tokens": 378697179.0, + "step": 14646 + }, + { + "epoch": 1.6084998901822973, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5882794857025146, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.725963830947876, + "num_tokens": 378718268.0, + "step": 14647 + }, + { + "epoch": 1.608609707884911, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2741377353668213, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7143765091896057, + "num_tokens": 378745182.0, + "step": 14648 + }, + { + "epoch": 1.6087195255875248, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2967188358306885, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7359510660171509, + "num_tokens": 378770063.0, + "step": 14649 + }, + { + "epoch": 1.6088293432901384, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.294848918914795, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7316073179244995, + "num_tokens": 378794104.0, + "step": 14650 + }, + { + "epoch": 1.608939160992752, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1933979988098145, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.718019962310791, + "num_tokens": 378820354.0, + "step": 14651 + }, + { + "epoch": 1.6090489786953657, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.249898910522461, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7169318199157715, + "num_tokens": 378847049.0, + "step": 14652 + }, + { + "epoch": 1.6091587963979794, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2729856967926025, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7367321848869324, + "num_tokens": 378871601.0, + "step": 14653 + }, + { + "epoch": 1.609268614100593, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.253349781036377, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7332707643508911, + "num_tokens": 378895536.0, + "step": 14654 + }, + { + "epoch": 1.6093784318032067, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1513166427612305, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7174543142318726, + "num_tokens": 378923499.0, + "step": 14655 + }, + { + "epoch": 1.6094882495058203, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0738062858581543, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.709265947341919, + "num_tokens": 378952595.0, + "step": 14656 + }, + { + "epoch": 1.609598067208434, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1130876541137695, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7036396861076355, + "num_tokens": 378981185.0, + "step": 14657 + }, + { + "epoch": 1.6097078849110478, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.219998836517334, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.711158812046051, + "num_tokens": 379009117.0, + "step": 14658 + }, + { + "epoch": 1.6098177026136613, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6377780437469482, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7367051839828491, + "num_tokens": 379028374.0, + "step": 14659 + }, + { + "epoch": 1.6099275203162748, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.030482769012451, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6969527006149292, + "num_tokens": 379058265.0, + "step": 14660 + }, + { + "epoch": 1.6100373380188886, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4930825233459473, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7162209749221802, + "num_tokens": 379078961.0, + "step": 14661 + }, + { + "epoch": 1.6101471557215024, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3248291015625, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7231578826904297, + "num_tokens": 379104371.0, + "step": 14662 + }, + { + "epoch": 1.6102569734241161, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.500728130340576, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.729613184928894, + "num_tokens": 379125368.0, + "step": 14663 + }, + { + "epoch": 1.6103667911267296, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.217360258102417, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7166235446929932, + "num_tokens": 379148661.0, + "step": 14664 + }, + { + "epoch": 1.6104766088293432, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.029818534851074, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7061004638671875, + "num_tokens": 379180087.0, + "step": 14665 + }, + { + "epoch": 1.610586426531957, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1984915733337402, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6949247717857361, + "num_tokens": 379208009.0, + "step": 14666 + }, + { + "epoch": 1.6106962442345707, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4069812297821045, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7187170386314392, + "num_tokens": 379229757.0, + "step": 14667 + }, + { + "epoch": 1.6108060619371842, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3761301040649414, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7326804399490356, + "num_tokens": 379252466.0, + "step": 14668 + }, + { + "epoch": 1.610915879639798, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.182163715362549, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7291691303253174, + "num_tokens": 379281949.0, + "step": 14669 + }, + { + "epoch": 1.6110256973424115, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.185112714767456, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6949766874313354, + "num_tokens": 379314100.0, + "step": 14670 + }, + { + "epoch": 1.6111355150450253, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.193098545074463, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7213906645774841, + "num_tokens": 379341840.0, + "step": 14671 + }, + { + "epoch": 1.611245332747639, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1296534538269043, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7293703556060791, + "num_tokens": 379368847.0, + "step": 14672 + }, + { + "epoch": 1.6113551504502526, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4033658504486084, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7387051582336426, + "num_tokens": 379391704.0, + "step": 14673 + }, + { + "epoch": 1.611464968152866, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4633889198303223, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7305971384048462, + "num_tokens": 379415996.0, + "step": 14674 + }, + { + "epoch": 1.6115747858554799, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.028837203979492, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7271355986595154, + "num_tokens": 379445398.0, + "step": 14675 + }, + { + "epoch": 1.6116846035580936, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1260037422180176, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.721976101398468, + "num_tokens": 379473108.0, + "step": 14676 + }, + { + "epoch": 1.6117944212607074, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2958483695983887, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7285282611846924, + "num_tokens": 379497755.0, + "step": 14677 + }, + { + "epoch": 1.611904238963321, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5010416507720947, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7286533713340759, + "num_tokens": 379520189.0, + "step": 14678 + }, + { + "epoch": 1.6120140566659344, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.443265438079834, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.714746356010437, + "num_tokens": 379543320.0, + "step": 14679 + }, + { + "epoch": 1.6121238743685482, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2690091133117676, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7103021144866943, + "num_tokens": 379566938.0, + "step": 14680 + }, + { + "epoch": 1.612233692071162, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2774739265441895, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.719866931438446, + "num_tokens": 379591567.0, + "step": 14681 + }, + { + "epoch": 1.6123435097737755, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.18422794342041, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7313966155052185, + "num_tokens": 379617964.0, + "step": 14682 + }, + { + "epoch": 1.612453327476389, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1454076766967773, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7162150144577026, + "num_tokens": 379645750.0, + "step": 14683 + }, + { + "epoch": 1.6125631451790028, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.269599199295044, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7395520210266113, + "num_tokens": 379669620.0, + "step": 14684 + }, + { + "epoch": 1.6126729628816165, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.257931709289551, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.71623694896698, + "num_tokens": 379695559.0, + "step": 14685 + }, + { + "epoch": 1.6127827805842303, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3975443840026855, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.719635009765625, + "num_tokens": 379719060.0, + "step": 14686 + }, + { + "epoch": 1.6128925982868438, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2321717739105225, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.703884482383728, + "num_tokens": 379746301.0, + "step": 14687 + }, + { + "epoch": 1.6130024159894574, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.124096632003784, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7059446573257446, + "num_tokens": 379774412.0, + "step": 14688 + }, + { + "epoch": 1.6131122336920711, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.093442440032959, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6977342963218689, + "num_tokens": 379802433.0, + "step": 14689 + }, + { + "epoch": 1.6132220513946849, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9548033475875854, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7132272720336914, + "num_tokens": 379834583.0, + "step": 14690 + }, + { + "epoch": 1.6133318690972986, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6934823989868164, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7227921485900879, + "num_tokens": 379854369.0, + "step": 14691 + }, + { + "epoch": 1.6134416867999122, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2357468605041504, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7191066741943359, + "num_tokens": 379880685.0, + "step": 14692 + }, + { + "epoch": 1.6135515045025257, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1304707527160645, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6878036856651306, + "num_tokens": 379910731.0, + "step": 14693 + }, + { + "epoch": 1.6136613222051395, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.269407272338867, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7297694683074951, + "num_tokens": 379934116.0, + "step": 14694 + }, + { + "epoch": 1.6137711399077532, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.255420684814453, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7298216819763184, + "num_tokens": 379960684.0, + "step": 14695 + }, + { + "epoch": 1.6138809576103668, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.259239673614502, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7085109949111938, + "num_tokens": 379987816.0, + "step": 14696 + }, + { + "epoch": 1.6139907753129803, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0432586669921875, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7184048891067505, + "num_tokens": 380017144.0, + "step": 14697 + }, + { + "epoch": 1.614100593015594, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3984639644622803, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7262401580810547, + "num_tokens": 380039497.0, + "step": 14698 + }, + { + "epoch": 1.6142104107182078, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.106689214706421, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7158485054969788, + "num_tokens": 380068009.0, + "step": 14699 + }, + { + "epoch": 1.6143202284208216, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0305557250976562, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7062708735466003, + "num_tokens": 380101244.0, + "step": 14700 + }, + { + "epoch": 1.614430046123435, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1633076667785645, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7341669797897339, + "num_tokens": 380129242.0, + "step": 14701 + }, + { + "epoch": 1.6145398638260486, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4800806045532227, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7037917971611023, + "num_tokens": 380151774.0, + "step": 14702 + }, + { + "epoch": 1.6146496815286624, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.080397844314575, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7346802353858948, + "num_tokens": 380178855.0, + "step": 14703 + }, + { + "epoch": 1.6147594992312762, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3934288024902344, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7205460667610168, + "num_tokens": 380200391.0, + "step": 14704 + }, + { + "epoch": 1.61486931693389, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0583627223968506, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7016041278839111, + "num_tokens": 380231561.0, + "step": 14705 + }, + { + "epoch": 1.6149791346365034, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1975889205932617, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7352650165557861, + "num_tokens": 380255188.0, + "step": 14706 + }, + { + "epoch": 1.615088952339117, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.453305721282959, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7250493764877319, + "num_tokens": 380278009.0, + "step": 14707 + }, + { + "epoch": 1.6151987700417307, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9428526163101196, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6971859931945801, + "num_tokens": 380311382.0, + "step": 14708 + }, + { + "epoch": 1.6153085877443445, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.8500546216964722, + "learning_rate": 1e-06, + "loss": 1.0293, + "mean_token_accuracy": 0.6858126521110535, + "num_tokens": 380348900.0, + "step": 14709 + }, + { + "epoch": 1.615418405446958, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.270554780960083, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7380768060684204, + "num_tokens": 380373273.0, + "step": 14710 + }, + { + "epoch": 1.6155282231495716, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.020749807357788, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7252329587936401, + "num_tokens": 380404588.0, + "step": 14711 + }, + { + "epoch": 1.6156380408521853, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.031938076019287, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.701608419418335, + "num_tokens": 380435705.0, + "step": 14712 + }, + { + "epoch": 1.615747858554799, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3476266860961914, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.708756148815155, + "num_tokens": 380459583.0, + "step": 14713 + }, + { + "epoch": 1.6158576762574128, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3291280269622803, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7476319074630737, + "num_tokens": 380483648.0, + "step": 14714 + }, + { + "epoch": 1.6159674939600264, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.231098175048828, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.7008236050605774, + "num_tokens": 380510835.0, + "step": 14715 + }, + { + "epoch": 1.61607731166264, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4566502571105957, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7050939798355103, + "num_tokens": 380534479.0, + "step": 14716 + }, + { + "epoch": 1.6161871293652537, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.25839900970459, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7150063514709473, + "num_tokens": 380558985.0, + "step": 14717 + }, + { + "epoch": 1.6162969470678674, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4607911109924316, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7127341032028198, + "num_tokens": 380582954.0, + "step": 14718 + }, + { + "epoch": 1.616406764770481, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.16089129447937, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7083282470703125, + "num_tokens": 380609904.0, + "step": 14719 + }, + { + "epoch": 1.6165165824730947, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4568300247192383, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7162036895751953, + "num_tokens": 380633853.0, + "step": 14720 + }, + { + "epoch": 1.6166264001757082, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3553178310394287, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7166987657546997, + "num_tokens": 380658581.0, + "step": 14721 + }, + { + "epoch": 1.616736217878322, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1595661640167236, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7326387763023376, + "num_tokens": 380684309.0, + "step": 14722 + }, + { + "epoch": 1.6168460355809358, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1508123874664307, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7089970707893372, + "num_tokens": 380714010.0, + "step": 14723 + }, + { + "epoch": 1.6169558532835493, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.229022979736328, + "learning_rate": 1e-06, + "loss": 0.7436, + "mean_token_accuracy": 0.7653520107269287, + "num_tokens": 380736775.0, + "step": 14724 + }, + { + "epoch": 1.6170656709861628, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.442133903503418, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7414624691009521, + "num_tokens": 380758733.0, + "step": 14725 + }, + { + "epoch": 1.6171754886887766, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.394836902618408, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7324370741844177, + "num_tokens": 380781061.0, + "step": 14726 + }, + { + "epoch": 1.6172853063913903, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.5497066974639893, + "learning_rate": 1e-06, + "loss": 0.7823, + "mean_token_accuracy": 0.7479296326637268, + "num_tokens": 380800537.0, + "step": 14727 + }, + { + "epoch": 1.617395124094004, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4971070289611816, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7112125754356384, + "num_tokens": 380822551.0, + "step": 14728 + }, + { + "epoch": 1.6175049417966176, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.124070167541504, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6926345229148865, + "num_tokens": 380850508.0, + "step": 14729 + }, + { + "epoch": 1.6176147594992312, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2977101802825928, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7163108587265015, + "num_tokens": 380873596.0, + "step": 14730 + }, + { + "epoch": 1.617724577201845, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1283082962036133, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7161698341369629, + "num_tokens": 380900552.0, + "step": 14731 + }, + { + "epoch": 1.6178343949044587, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1582632064819336, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7242411375045776, + "num_tokens": 380927255.0, + "step": 14732 + }, + { + "epoch": 1.6179442126070722, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.228182315826416, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7020575404167175, + "num_tokens": 380952323.0, + "step": 14733 + }, + { + "epoch": 1.618054030309686, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.455777168273926, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6924733519554138, + "num_tokens": 380978041.0, + "step": 14734 + }, + { + "epoch": 1.6181638480122995, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3994617462158203, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7445930242538452, + "num_tokens": 380998558.0, + "step": 14735 + }, + { + "epoch": 1.6182736657149133, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.032451868057251, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7038766145706177, + "num_tokens": 381029442.0, + "step": 14736 + }, + { + "epoch": 1.618383483417527, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.359428644180298, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7353219985961914, + "num_tokens": 381053022.0, + "step": 14737 + }, + { + "epoch": 1.6184933011201406, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.404778242111206, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.7018344402313232, + "num_tokens": 381078358.0, + "step": 14738 + }, + { + "epoch": 1.618603118822754, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1589996814727783, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7191840410232544, + "num_tokens": 381105573.0, + "step": 14739 + }, + { + "epoch": 1.6187129365253679, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.100663900375366, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7124021053314209, + "num_tokens": 381136566.0, + "step": 14740 + }, + { + "epoch": 1.6188227542279816, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3721983432769775, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7162551283836365, + "num_tokens": 381159490.0, + "step": 14741 + }, + { + "epoch": 1.6189325719305954, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.171251058578491, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.740199089050293, + "num_tokens": 381184986.0, + "step": 14742 + }, + { + "epoch": 1.619042389633209, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1409881114959717, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7226026058197021, + "num_tokens": 381210746.0, + "step": 14743 + }, + { + "epoch": 1.6191522073358224, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.9828389883041382, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7101857662200928, + "num_tokens": 381241557.0, + "step": 14744 + }, + { + "epoch": 1.6192620250384362, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2420706748962402, + "learning_rate": 1e-06, + "loss": 0.7976, + "mean_token_accuracy": 0.7453428506851196, + "num_tokens": 381266486.0, + "step": 14745 + }, + { + "epoch": 1.61937184274105, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.207271099090576, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.731601893901825, + "num_tokens": 381292874.0, + "step": 14746 + }, + { + "epoch": 1.6194816604436635, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0955545902252197, + "learning_rate": 1e-06, + "loss": 1.0451, + "mean_token_accuracy": 0.6956538558006287, + "num_tokens": 381323286.0, + "step": 14747 + }, + { + "epoch": 1.619591478146277, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3645999431610107, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7314758896827698, + "num_tokens": 381344594.0, + "step": 14748 + }, + { + "epoch": 1.6197012958488908, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.231154203414917, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.719862699508667, + "num_tokens": 381371462.0, + "step": 14749 + }, + { + "epoch": 1.6198111135515045, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.313101053237915, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7005569934844971, + "num_tokens": 381397383.0, + "step": 14750 + }, + { + "epoch": 1.6199209312541183, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0698111057281494, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7208433747291565, + "num_tokens": 381426112.0, + "step": 14751 + }, + { + "epoch": 1.6200307489567318, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2841947078704834, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7093019485473633, + "num_tokens": 381452303.0, + "step": 14752 + }, + { + "epoch": 1.6201405666593454, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.365382671356201, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7086254358291626, + "num_tokens": 381476583.0, + "step": 14753 + }, + { + "epoch": 1.6202503843619591, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0256755352020264, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7077846527099609, + "num_tokens": 381509030.0, + "step": 14754 + }, + { + "epoch": 1.6203602020645729, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4973106384277344, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7495855689048767, + "num_tokens": 381529557.0, + "step": 14755 + }, + { + "epoch": 1.6204700197671866, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0529558658599854, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7006129622459412, + "num_tokens": 381558586.0, + "step": 14756 + }, + { + "epoch": 1.6205798374698002, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1229822635650635, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6952698230743408, + "num_tokens": 381587177.0, + "step": 14757 + }, + { + "epoch": 1.6206896551724137, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.030411958694458, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7062715291976929, + "num_tokens": 381617923.0, + "step": 14758 + }, + { + "epoch": 1.6207994728750275, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1328132152557373, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7059493660926819, + "num_tokens": 381647988.0, + "step": 14759 + }, + { + "epoch": 1.6209092905776412, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3335514068603516, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7128964066505432, + "num_tokens": 381671221.0, + "step": 14760 + }, + { + "epoch": 1.6210191082802548, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2521655559539795, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6987935900688171, + "num_tokens": 381698408.0, + "step": 14761 + }, + { + "epoch": 1.6211289259828683, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.6534039974212646, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7282065153121948, + "num_tokens": 381717742.0, + "step": 14762 + }, + { + "epoch": 1.621238743685482, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0804967880249023, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7238771319389343, + "num_tokens": 381748107.0, + "step": 14763 + }, + { + "epoch": 1.6213485613880958, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3795132637023926, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7306291460990906, + "num_tokens": 381770856.0, + "step": 14764 + }, + { + "epoch": 1.6214583790907096, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0775887966156006, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7170827984809875, + "num_tokens": 381799450.0, + "step": 14765 + }, + { + "epoch": 1.621568196793323, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.221978187561035, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7116743326187134, + "num_tokens": 381825078.0, + "step": 14766 + }, + { + "epoch": 1.6216780144959366, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1620593070983887, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7092447280883789, + "num_tokens": 381853370.0, + "step": 14767 + }, + { + "epoch": 1.6217878321985504, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.192333221435547, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7247883081436157, + "num_tokens": 381880615.0, + "step": 14768 + }, + { + "epoch": 1.6218976499011641, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.444864511489868, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7178203463554382, + "num_tokens": 381904268.0, + "step": 14769 + }, + { + "epoch": 1.6220074676037777, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.291372060775757, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7128624320030212, + "num_tokens": 381928517.0, + "step": 14770 + }, + { + "epoch": 1.6221172853063914, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4332470893859863, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7170466184616089, + "num_tokens": 381951430.0, + "step": 14771 + }, + { + "epoch": 1.622227103009005, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.28426194190979, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7209944128990173, + "num_tokens": 381976894.0, + "step": 14772 + }, + { + "epoch": 1.6223369207116187, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3167624473571777, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7358112931251526, + "num_tokens": 382002349.0, + "step": 14773 + }, + { + "epoch": 1.6224467384142325, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.481053113937378, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7145774364471436, + "num_tokens": 382025309.0, + "step": 14774 + }, + { + "epoch": 1.622556556116846, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4274234771728516, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.711367666721344, + "num_tokens": 382047531.0, + "step": 14775 + }, + { + "epoch": 1.6226663738194596, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.4668936729431152, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.698910117149353, + "num_tokens": 382070872.0, + "step": 14776 + }, + { + "epoch": 1.6227761915220733, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.0378878116607666, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7192937135696411, + "num_tokens": 382100230.0, + "step": 14777 + }, + { + "epoch": 1.622886009224687, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 1.98623788356781, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7476350665092468, + "num_tokens": 382129822.0, + "step": 14778 + }, + { + "epoch": 1.6229958269273008, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.232159376144409, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7385026216506958, + "num_tokens": 382155243.0, + "step": 14779 + }, + { + "epoch": 1.6231056446299144, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.267442226409912, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7370519638061523, + "num_tokens": 382179428.0, + "step": 14780 + }, + { + "epoch": 1.623215462332528, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2352209091186523, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7150710225105286, + "num_tokens": 382205016.0, + "step": 14781 + }, + { + "epoch": 1.6233252800351416, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.1306960582733154, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.6981068253517151, + "num_tokens": 382236986.0, + "step": 14782 + }, + { + "epoch": 1.6234350977377554, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.0513734817504883, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7041476964950562, + "num_tokens": 382268483.0, + "step": 14783 + }, + { + "epoch": 1.623544915440369, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.3139114379882812, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7151460647583008, + "num_tokens": 382294735.0, + "step": 14784 + }, + { + "epoch": 1.6236547331429827, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.190572738647461, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7498542070388794, + "num_tokens": 382319364.0, + "step": 14785 + }, + { + "epoch": 1.6237645508455962, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.2145004272460938, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7089051008224487, + "num_tokens": 382343621.0, + "step": 14786 + }, + { + "epoch": 1.62387436854821, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.220343828201294, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7089395523071289, + "num_tokens": 382371121.0, + "step": 14787 + }, + { + "epoch": 1.6239841862508237, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2770862579345703, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7138583660125732, + "num_tokens": 382397503.0, + "step": 14788 + }, + { + "epoch": 1.6240940039534373, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.18176531791687, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.6996181011199951, + "num_tokens": 382424574.0, + "step": 14789 + }, + { + "epoch": 1.6242038216560508, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.128422260284424, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7114036083221436, + "num_tokens": 382450711.0, + "step": 14790 + }, + { + "epoch": 1.6243136393586646, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.2372093200683594, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7215635776519775, + "num_tokens": 382476821.0, + "step": 14791 + }, + { + "epoch": 1.6244234570612783, + "ewc_loss": 0.0001239776611328125, + "grad_norm": 2.216353178024292, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.71638423204422, + "num_tokens": 382503988.0, + "step": 14792 + }, + { + "epoch": 1.624533274763892, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.031568765640259, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7099936008453369, + "num_tokens": 382533558.0, + "step": 14793 + }, + { + "epoch": 1.6246430924665056, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.430326461791992, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7162293195724487, + "num_tokens": 382557276.0, + "step": 14794 + }, + { + "epoch": 1.6247529101691192, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.3617405891418457, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7062854170799255, + "num_tokens": 382582157.0, + "step": 14795 + }, + { + "epoch": 1.624862727871733, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.0746238231658936, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6994094848632812, + "num_tokens": 382612789.0, + "step": 14796 + }, + { + "epoch": 1.6249725455743467, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 1.9878836870193481, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7015232443809509, + "num_tokens": 382646381.0, + "step": 14797 + }, + { + "epoch": 1.6250823632769602, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.3357431888580322, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7170711755752563, + "num_tokens": 382670294.0, + "step": 14798 + }, + { + "epoch": 1.6251921809795737, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.2688910961151123, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7280350923538208, + "num_tokens": 382694925.0, + "step": 14799 + }, + { + "epoch": 1.6253019986821875, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.0884218215942383, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7030948400497437, + "num_tokens": 382724352.0, + "step": 14800 + }, + { + "epoch": 1.6254118163848013, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.2462592124938965, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7231969833374023, + "num_tokens": 382751825.0, + "step": 14801 + }, + { + "epoch": 1.625521634087415, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.2799735069274902, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7208995223045349, + "num_tokens": 382777700.0, + "step": 14802 + }, + { + "epoch": 1.6256314517900285, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.3700125217437744, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7146108150482178, + "num_tokens": 382801147.0, + "step": 14803 + }, + { + "epoch": 1.625741269492642, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.423762321472168, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.707908034324646, + "num_tokens": 382823777.0, + "step": 14804 + }, + { + "epoch": 1.6258510871952558, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.126690626144409, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7144112586975098, + "num_tokens": 382851424.0, + "step": 14805 + }, + { + "epoch": 1.6259609048978696, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.1736814975738525, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7325172424316406, + "num_tokens": 382879282.0, + "step": 14806 + }, + { + "epoch": 1.6260707226004834, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.420616626739502, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7097248435020447, + "num_tokens": 382905436.0, + "step": 14807 + }, + { + "epoch": 1.6261805403030969, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.1844735145568848, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7141880989074707, + "num_tokens": 382933757.0, + "step": 14808 + }, + { + "epoch": 1.6262903580057104, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.4625325202941895, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7117710113525391, + "num_tokens": 382956239.0, + "step": 14809 + }, + { + "epoch": 1.6264001757083242, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.0143184661865234, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7197455167770386, + "num_tokens": 382988165.0, + "step": 14810 + }, + { + "epoch": 1.626509993410938, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.3519296646118164, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7169690132141113, + "num_tokens": 383014012.0, + "step": 14811 + }, + { + "epoch": 1.6266198111135515, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.3147366046905518, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7088749408721924, + "num_tokens": 383039558.0, + "step": 14812 + }, + { + "epoch": 1.626729628816165, + "ewc_loss": 0.00012493133544921875, + "grad_norm": 2.0612449645996094, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7081446647644043, + "num_tokens": 383068124.0, + "step": 14813 + }, + { + "epoch": 1.6268394465187788, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.3449416160583496, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7284619212150574, + "num_tokens": 383091127.0, + "step": 14814 + }, + { + "epoch": 1.6269492642213925, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.205331802368164, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7208805084228516, + "num_tokens": 383117348.0, + "step": 14815 + }, + { + "epoch": 1.6270590819240063, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.2433109283447266, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7263954877853394, + "num_tokens": 383144594.0, + "step": 14816 + }, + { + "epoch": 1.6271688996266198, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.230928659439087, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7074477672576904, + "num_tokens": 383173211.0, + "step": 14817 + }, + { + "epoch": 1.6272787173292333, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.312624931335449, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7090116739273071, + "num_tokens": 383196574.0, + "step": 14818 + }, + { + "epoch": 1.627388535031847, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.1433629989624023, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.713001012802124, + "num_tokens": 383225233.0, + "step": 14819 + }, + { + "epoch": 1.6274983527344609, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.4585437774658203, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7261790037155151, + "num_tokens": 383246301.0, + "step": 14820 + }, + { + "epoch": 1.6276081704370746, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.0867154598236084, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7113743424415588, + "num_tokens": 383276684.0, + "step": 14821 + }, + { + "epoch": 1.6277179881396882, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.4582266807556152, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.733629047870636, + "num_tokens": 383298194.0, + "step": 14822 + }, + { + "epoch": 1.6278278058423017, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.421987771987915, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7482370734214783, + "num_tokens": 383319361.0, + "step": 14823 + }, + { + "epoch": 1.6279376235449154, + "ewc_loss": 0.000125885009765625, + "grad_norm": 1.9912939071655273, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7243568897247314, + "num_tokens": 383349452.0, + "step": 14824 + }, + { + "epoch": 1.6280474412475292, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.222505569458008, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7069774866104126, + "num_tokens": 383380037.0, + "step": 14825 + }, + { + "epoch": 1.6281572589501427, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.2149100303649902, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7386395931243896, + "num_tokens": 383405249.0, + "step": 14826 + }, + { + "epoch": 1.6282670766527563, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.242063522338867, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7031562924385071, + "num_tokens": 383430728.0, + "step": 14827 + }, + { + "epoch": 1.62837689435537, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.0494132041931152, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7154260873794556, + "num_tokens": 383460877.0, + "step": 14828 + }, + { + "epoch": 1.6284867120579838, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.47857928276062, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7263619899749756, + "num_tokens": 383484181.0, + "step": 14829 + }, + { + "epoch": 1.6285965297605975, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.0919673442840576, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7168290615081787, + "num_tokens": 383513323.0, + "step": 14830 + }, + { + "epoch": 1.628706347463211, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.1398613452911377, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7410250306129456, + "num_tokens": 383538472.0, + "step": 14831 + }, + { + "epoch": 1.6288161651658246, + "ewc_loss": 0.000125885009765625, + "grad_norm": 2.2635746002197266, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7314389944076538, + "num_tokens": 383563054.0, + "step": 14832 + }, + { + "epoch": 1.6289259828684384, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4618582725524902, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7158054113388062, + "num_tokens": 383586167.0, + "step": 14833 + }, + { + "epoch": 1.6290358005710521, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.346052408218384, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7391359210014343, + "num_tokens": 383609904.0, + "step": 14834 + }, + { + "epoch": 1.6291456182736657, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4186644554138184, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7227731943130493, + "num_tokens": 383631601.0, + "step": 14835 + }, + { + "epoch": 1.6292554359762794, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4074501991271973, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7341557741165161, + "num_tokens": 383653670.0, + "step": 14836 + }, + { + "epoch": 1.629365253678893, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1004583835601807, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.704306960105896, + "num_tokens": 383683590.0, + "step": 14837 + }, + { + "epoch": 1.6294750713815067, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 1.948127269744873, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.732286810874939, + "num_tokens": 383716512.0, + "step": 14838 + }, + { + "epoch": 1.6295848890841205, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0832808017730713, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.699242115020752, + "num_tokens": 383748673.0, + "step": 14839 + }, + { + "epoch": 1.629694706786734, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0625505447387695, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7357109785079956, + "num_tokens": 383775985.0, + "step": 14840 + }, + { + "epoch": 1.6298045244893475, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1916956901550293, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7087048292160034, + "num_tokens": 383803403.0, + "step": 14841 + }, + { + "epoch": 1.6299143421919613, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1765127182006836, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7396061420440674, + "num_tokens": 383829954.0, + "step": 14842 + }, + { + "epoch": 1.630024159894575, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3595805168151855, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7181153297424316, + "num_tokens": 383852254.0, + "step": 14843 + }, + { + "epoch": 1.6301339775971888, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.515842914581299, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7121080756187439, + "num_tokens": 383873300.0, + "step": 14844 + }, + { + "epoch": 1.6302437952998023, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2828261852264404, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7277634143829346, + "num_tokens": 383898671.0, + "step": 14845 + }, + { + "epoch": 1.6303536130024159, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.264143466949463, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7201719284057617, + "num_tokens": 383924445.0, + "step": 14846 + }, + { + "epoch": 1.6304634307050296, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1481122970581055, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7211947441101074, + "num_tokens": 383952449.0, + "step": 14847 + }, + { + "epoch": 1.6305732484076434, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3866355419158936, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7221892476081848, + "num_tokens": 383976174.0, + "step": 14848 + }, + { + "epoch": 1.630683066110257, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1901633739471436, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7230687141418457, + "num_tokens": 384003191.0, + "step": 14849 + }, + { + "epoch": 1.6307928838128707, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1178338527679443, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7171170711517334, + "num_tokens": 384030783.0, + "step": 14850 + }, + { + "epoch": 1.6309027015154842, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.040477752685547, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6952890157699585, + "num_tokens": 384061920.0, + "step": 14851 + }, + { + "epoch": 1.631012519218098, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3052940368652344, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7343921661376953, + "num_tokens": 384086460.0, + "step": 14852 + }, + { + "epoch": 1.6311223369207117, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.342771291732788, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7302249670028687, + "num_tokens": 384110573.0, + "step": 14853 + }, + { + "epoch": 1.6312321546233253, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2968432903289795, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7257766723632812, + "num_tokens": 384135820.0, + "step": 14854 + }, + { + "epoch": 1.6313419723259388, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.142023801803589, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7152369618415833, + "num_tokens": 384161870.0, + "step": 14855 + }, + { + "epoch": 1.6314517900285526, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.341428756713867, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7156343460083008, + "num_tokens": 384185167.0, + "step": 14856 + }, + { + "epoch": 1.6315616077311663, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.286614179611206, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7343417406082153, + "num_tokens": 384209478.0, + "step": 14857 + }, + { + "epoch": 1.63167142543378, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0960192680358887, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7143216729164124, + "num_tokens": 384238523.0, + "step": 14858 + }, + { + "epoch": 1.6317812431363936, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1484858989715576, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7332451343536377, + "num_tokens": 384264608.0, + "step": 14859 + }, + { + "epoch": 1.6318910608390071, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1688945293426514, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7160834074020386, + "num_tokens": 384292304.0, + "step": 14860 + }, + { + "epoch": 1.632000878541621, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1546242237091064, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6871790885925293, + "num_tokens": 384322313.0, + "step": 14861 + }, + { + "epoch": 1.6321106962442347, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0811831951141357, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7064266204833984, + "num_tokens": 384351098.0, + "step": 14862 + }, + { + "epoch": 1.6322205139468482, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.279067039489746, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7183312177658081, + "num_tokens": 384376359.0, + "step": 14863 + }, + { + "epoch": 1.6323303316494617, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0885536670684814, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7182291150093079, + "num_tokens": 384404628.0, + "step": 14864 + }, + { + "epoch": 1.6324401493520755, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.218862771987915, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7082564830780029, + "num_tokens": 384431541.0, + "step": 14865 + }, + { + "epoch": 1.6325499670546892, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1764893531799316, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6998807191848755, + "num_tokens": 384460588.0, + "step": 14866 + }, + { + "epoch": 1.632659784757303, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.5298593044281006, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7343773245811462, + "num_tokens": 384481212.0, + "step": 14867 + }, + { + "epoch": 1.6327696024599165, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1481850147247314, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7455434203147888, + "num_tokens": 384507741.0, + "step": 14868 + }, + { + "epoch": 1.63287942016253, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1675636768341064, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7235434055328369, + "num_tokens": 384532661.0, + "step": 14869 + }, + { + "epoch": 1.6329892378651438, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1915252208709717, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6894938945770264, + "num_tokens": 384560894.0, + "step": 14870 + }, + { + "epoch": 1.6330990555677576, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1161656379699707, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7051958441734314, + "num_tokens": 384588914.0, + "step": 14871 + }, + { + "epoch": 1.6332088732703713, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1833789348602295, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.706100583076477, + "num_tokens": 384616629.0, + "step": 14872 + }, + { + "epoch": 1.6333186909729849, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.087751865386963, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7077606916427612, + "num_tokens": 384647781.0, + "step": 14873 + }, + { + "epoch": 1.6334285086755984, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.454610586166382, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7084933519363403, + "num_tokens": 384670700.0, + "step": 14874 + }, + { + "epoch": 1.6335383263782122, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.608490467071533, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7297766208648682, + "num_tokens": 384689528.0, + "step": 14875 + }, + { + "epoch": 1.633648144080826, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 1.932762622833252, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7003780603408813, + "num_tokens": 384723911.0, + "step": 14876 + }, + { + "epoch": 1.6337579617834395, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1523938179016113, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7162544131278992, + "num_tokens": 384750167.0, + "step": 14877 + }, + { + "epoch": 1.633867779486053, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.114039659500122, + "learning_rate": 1e-06, + "loss": 0.7467, + "mean_token_accuracy": 0.7613615989685059, + "num_tokens": 384775381.0, + "step": 14878 + }, + { + "epoch": 1.6339775971886668, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.515378952026367, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7257093787193298, + "num_tokens": 384795819.0, + "step": 14879 + }, + { + "epoch": 1.6340874148912805, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.33817195892334, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7155619263648987, + "num_tokens": 384818800.0, + "step": 14880 + }, + { + "epoch": 1.6341972325938943, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.230692148208618, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7068920135498047, + "num_tokens": 384844704.0, + "step": 14881 + }, + { + "epoch": 1.6343070502965078, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2727293968200684, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.719111442565918, + "num_tokens": 384871538.0, + "step": 14882 + }, + { + "epoch": 1.6344168679991213, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2578988075256348, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7056692838668823, + "num_tokens": 384899619.0, + "step": 14883 + }, + { + "epoch": 1.634526685701735, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.212820529937744, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7186352014541626, + "num_tokens": 384927967.0, + "step": 14884 + }, + { + "epoch": 1.6346365034043489, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.5323855876922607, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7406994104385376, + "num_tokens": 384950049.0, + "step": 14885 + }, + { + "epoch": 1.6347463211069626, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3338441848754883, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.720049262046814, + "num_tokens": 384974721.0, + "step": 14886 + }, + { + "epoch": 1.6348561388095761, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.540029764175415, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.725805401802063, + "num_tokens": 384997463.0, + "step": 14887 + }, + { + "epoch": 1.6349659565121897, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.8509302139282227, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7310774922370911, + "num_tokens": 385015660.0, + "step": 14888 + }, + { + "epoch": 1.6350757742148034, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3667190074920654, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7304742336273193, + "num_tokens": 385040549.0, + "step": 14889 + }, + { + "epoch": 1.6351855919174172, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4038548469543457, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7547073364257812, + "num_tokens": 385061746.0, + "step": 14890 + }, + { + "epoch": 1.6352954096200307, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.587873935699463, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7123964428901672, + "num_tokens": 385082821.0, + "step": 14891 + }, + { + "epoch": 1.6354052273226443, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 1.9863426685333252, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7053449153900146, + "num_tokens": 385115248.0, + "step": 14892 + }, + { + "epoch": 1.635515045025258, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3625407218933105, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7232690453529358, + "num_tokens": 385139636.0, + "step": 14893 + }, + { + "epoch": 1.6356248627278718, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3060476779937744, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.715579628944397, + "num_tokens": 385164473.0, + "step": 14894 + }, + { + "epoch": 1.6357346804304855, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.175184488296509, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6962324976921082, + "num_tokens": 385192452.0, + "step": 14895 + }, + { + "epoch": 1.635844498133099, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.213564872741699, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7255758047103882, + "num_tokens": 385215938.0, + "step": 14896 + }, + { + "epoch": 1.6359543158357126, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3145158290863037, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7284420132637024, + "num_tokens": 385239699.0, + "step": 14897 + }, + { + "epoch": 1.6360641335383264, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.022103786468506, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6972416043281555, + "num_tokens": 385271485.0, + "step": 14898 + }, + { + "epoch": 1.6361739512409401, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.5457637310028076, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7126209735870361, + "num_tokens": 385292962.0, + "step": 14899 + }, + { + "epoch": 1.6362837689435537, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.276705265045166, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7206910848617554, + "num_tokens": 385318331.0, + "step": 14900 + }, + { + "epoch": 1.6363935866461674, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4267210960388184, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7150758504867554, + "num_tokens": 385342660.0, + "step": 14901 + }, + { + "epoch": 1.636503404348781, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.134749174118042, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7224045991897583, + "num_tokens": 385371735.0, + "step": 14902 + }, + { + "epoch": 1.6366132220513947, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.284632682800293, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7306370735168457, + "num_tokens": 385396048.0, + "step": 14903 + }, + { + "epoch": 1.6367230397540085, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.289719343185425, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7138594388961792, + "num_tokens": 385421690.0, + "step": 14904 + }, + { + "epoch": 1.636832857456622, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.431072950363159, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7175992727279663, + "num_tokens": 385444381.0, + "step": 14905 + }, + { + "epoch": 1.6369426751592355, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.223137140274048, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7262798547744751, + "num_tokens": 385470342.0, + "step": 14906 + }, + { + "epoch": 1.6370524928618493, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3087399005889893, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7391799688339233, + "num_tokens": 385493926.0, + "step": 14907 + }, + { + "epoch": 1.637162310564463, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3176825046539307, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7230926752090454, + "num_tokens": 385520209.0, + "step": 14908 + }, + { + "epoch": 1.6372721282670768, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2207961082458496, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7192407250404358, + "num_tokens": 385546552.0, + "step": 14909 + }, + { + "epoch": 1.6373819459696903, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.168598175048828, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7046713829040527, + "num_tokens": 385573161.0, + "step": 14910 + }, + { + "epoch": 1.6374917636723039, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2744300365448, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7596660852432251, + "num_tokens": 385597104.0, + "step": 14911 + }, + { + "epoch": 1.6376015813749176, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3046233654022217, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7342013716697693, + "num_tokens": 385619874.0, + "step": 14912 + }, + { + "epoch": 1.6377113990775314, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3365302085876465, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.697022557258606, + "num_tokens": 385646407.0, + "step": 14913 + }, + { + "epoch": 1.637821216780145, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.364938974380493, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7186134457588196, + "num_tokens": 385671319.0, + "step": 14914 + }, + { + "epoch": 1.6379310344827587, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1583974361419678, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.6974801421165466, + "num_tokens": 385701247.0, + "step": 14915 + }, + { + "epoch": 1.6380408521853722, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0825510025024414, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.696211576461792, + "num_tokens": 385732322.0, + "step": 14916 + }, + { + "epoch": 1.638150669887986, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.5167508125305176, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7269026041030884, + "num_tokens": 385754702.0, + "step": 14917 + }, + { + "epoch": 1.6382604875905997, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0629003047943115, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7059910297393799, + "num_tokens": 385785464.0, + "step": 14918 + }, + { + "epoch": 1.6383703052932133, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3322134017944336, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7233735918998718, + "num_tokens": 385810503.0, + "step": 14919 + }, + { + "epoch": 1.6384801229958268, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.6601569652557373, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7334048748016357, + "num_tokens": 385829758.0, + "step": 14920 + }, + { + "epoch": 1.6385899406984406, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.654702663421631, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7248588800430298, + "num_tokens": 385849368.0, + "step": 14921 + }, + { + "epoch": 1.6386997584010543, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.1234946250915527, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7123721837997437, + "num_tokens": 385876430.0, + "step": 14922 + }, + { + "epoch": 1.638809576103668, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.336925983428955, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7315598726272583, + "num_tokens": 385899573.0, + "step": 14923 + }, + { + "epoch": 1.6389193938062816, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.3889219760894775, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7262022495269775, + "num_tokens": 385922626.0, + "step": 14924 + }, + { + "epoch": 1.6390292115088951, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.24239182472229, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7253369092941284, + "num_tokens": 385947999.0, + "step": 14925 + }, + { + "epoch": 1.639139029211509, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.591876745223999, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7240749597549438, + "num_tokens": 385968647.0, + "step": 14926 + }, + { + "epoch": 1.6392488469141226, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.599069595336914, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7371891736984253, + "num_tokens": 385988918.0, + "step": 14927 + }, + { + "epoch": 1.6393586646167362, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.2499911785125732, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7231864333152771, + "num_tokens": 386014473.0, + "step": 14928 + }, + { + "epoch": 1.6394684823193497, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 1.9894498586654663, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7182935476303101, + "num_tokens": 386044083.0, + "step": 14929 + }, + { + "epoch": 1.6395783000219635, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.4440224170684814, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7199260592460632, + "num_tokens": 386066503.0, + "step": 14930 + }, + { + "epoch": 1.6396881177245772, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 1.901581883430481, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.704439640045166, + "num_tokens": 386100803.0, + "step": 14931 + }, + { + "epoch": 1.639797935427191, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.002255439758301, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.7068331837654114, + "num_tokens": 386133291.0, + "step": 14932 + }, + { + "epoch": 1.6399077531298045, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.0114259719848633, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.6994378566741943, + "num_tokens": 386165398.0, + "step": 14933 + }, + { + "epoch": 1.640017570832418, + "ewc_loss": 0.00012683868408203125, + "grad_norm": 2.020264148712158, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7059139013290405, + "num_tokens": 386196459.0, + "step": 14934 + }, + { + "epoch": 1.6401273885350318, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.385789632797241, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7061138153076172, + "num_tokens": 386221217.0, + "step": 14935 + }, + { + "epoch": 1.6402372062376456, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.373504400253296, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7235513925552368, + "num_tokens": 386244454.0, + "step": 14936 + }, + { + "epoch": 1.6403470239402593, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3380179405212402, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7132705450057983, + "num_tokens": 386271152.0, + "step": 14937 + }, + { + "epoch": 1.6404568416428729, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2352795600891113, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7069767713546753, + "num_tokens": 386296627.0, + "step": 14938 + }, + { + "epoch": 1.6405666593454864, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1974940299987793, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7285971641540527, + "num_tokens": 386322958.0, + "step": 14939 + }, + { + "epoch": 1.6406764770481002, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.299332618713379, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7338664531707764, + "num_tokens": 386345027.0, + "step": 14940 + }, + { + "epoch": 1.640786294750714, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.118557929992676, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7242274284362793, + "num_tokens": 386375754.0, + "step": 14941 + }, + { + "epoch": 1.6408961124533274, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.265079975128174, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7044388055801392, + "num_tokens": 386405499.0, + "step": 14942 + }, + { + "epoch": 1.641005930155941, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.102595806121826, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7260285019874573, + "num_tokens": 386435079.0, + "step": 14943 + }, + { + "epoch": 1.6411157478585547, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.181380033493042, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7135680913925171, + "num_tokens": 386461966.0, + "step": 14944 + }, + { + "epoch": 1.6412255655611685, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2068886756896973, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7246220111846924, + "num_tokens": 386487258.0, + "step": 14945 + }, + { + "epoch": 1.6413353832637823, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.435810089111328, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7088096141815186, + "num_tokens": 386511094.0, + "step": 14946 + }, + { + "epoch": 1.6414452009663958, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1768181324005127, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.710126519203186, + "num_tokens": 386537644.0, + "step": 14947 + }, + { + "epoch": 1.6415550186690093, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4290411472320557, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7096585631370544, + "num_tokens": 386563414.0, + "step": 14948 + }, + { + "epoch": 1.641664836371623, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3143374919891357, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.705790638923645, + "num_tokens": 386588261.0, + "step": 14949 + }, + { + "epoch": 1.6417746540742368, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.114250898361206, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7252141833305359, + "num_tokens": 386618288.0, + "step": 14950 + }, + { + "epoch": 1.6418844717768504, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.378840446472168, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7133399844169617, + "num_tokens": 386641683.0, + "step": 14951 + }, + { + "epoch": 1.6419942894794641, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.191094398498535, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7334445714950562, + "num_tokens": 386669054.0, + "step": 14952 + }, + { + "epoch": 1.6421041071820777, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0810139179229736, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7262645363807678, + "num_tokens": 386698654.0, + "step": 14953 + }, + { + "epoch": 1.6422139248846914, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.241766929626465, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7091987133026123, + "num_tokens": 386725386.0, + "step": 14954 + }, + { + "epoch": 1.6423237425873052, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.303952693939209, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7217233777046204, + "num_tokens": 386748385.0, + "step": 14955 + }, + { + "epoch": 1.6424335602899187, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2034316062927246, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7260435819625854, + "num_tokens": 386775453.0, + "step": 14956 + }, + { + "epoch": 1.6425433779925322, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1845786571502686, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.730688750743866, + "num_tokens": 386802118.0, + "step": 14957 + }, + { + "epoch": 1.642653195695146, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1562108993530273, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.695541501045227, + "num_tokens": 386833173.0, + "step": 14958 + }, + { + "epoch": 1.6427630133977598, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2049150466918945, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7072131037712097, + "num_tokens": 386858346.0, + "step": 14959 + }, + { + "epoch": 1.6428728311003735, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.22406005859375, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.718867301940918, + "num_tokens": 386886388.0, + "step": 14960 + }, + { + "epoch": 1.642982648802987, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.53593373298645, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7396450042724609, + "num_tokens": 386906007.0, + "step": 14961 + }, + { + "epoch": 1.6430924665056006, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.101583957672119, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7087937593460083, + "num_tokens": 386933428.0, + "step": 14962 + }, + { + "epoch": 1.6432022842082143, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1062614917755127, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7049950361251831, + "num_tokens": 386964853.0, + "step": 14963 + }, + { + "epoch": 1.643312101910828, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0619852542877197, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.704967737197876, + "num_tokens": 386993593.0, + "step": 14964 + }, + { + "epoch": 1.6434219196134416, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3588333129882812, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7092682123184204, + "num_tokens": 387017013.0, + "step": 14965 + }, + { + "epoch": 1.6435317373160554, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.464958667755127, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7395824790000916, + "num_tokens": 387037142.0, + "step": 14966 + }, + { + "epoch": 1.643641555018669, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.492304563522339, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7322832942008972, + "num_tokens": 387058297.0, + "step": 14967 + }, + { + "epoch": 1.6437513727212827, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.430591344833374, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.722763180732727, + "num_tokens": 387079291.0, + "step": 14968 + }, + { + "epoch": 1.6438611904238964, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.535057544708252, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7113622426986694, + "num_tokens": 387100506.0, + "step": 14969 + }, + { + "epoch": 1.64397100812651, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.192915916442871, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7070115804672241, + "num_tokens": 387128650.0, + "step": 14970 + }, + { + "epoch": 1.6440808258291235, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.450134038925171, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7237628698348999, + "num_tokens": 387151054.0, + "step": 14971 + }, + { + "epoch": 1.6441906435317373, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.421783208847046, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7237245440483093, + "num_tokens": 387173317.0, + "step": 14972 + }, + { + "epoch": 1.644300461234351, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5357019901275635, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7078360319137573, + "num_tokens": 387195398.0, + "step": 14973 + }, + { + "epoch": 1.6444102789369648, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5645036697387695, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7276289463043213, + "num_tokens": 387216806.0, + "step": 14974 + }, + { + "epoch": 1.6445200966395783, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1267106533050537, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.709113597869873, + "num_tokens": 387244637.0, + "step": 14975 + }, + { + "epoch": 1.6446299143421919, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.209364891052246, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7176842093467712, + "num_tokens": 387271443.0, + "step": 14976 + }, + { + "epoch": 1.6447397320448056, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1298139095306396, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.704085111618042, + "num_tokens": 387300901.0, + "step": 14977 + }, + { + "epoch": 1.6448495497474194, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.824463129043579, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7442666888237, + "num_tokens": 387319865.0, + "step": 14978 + }, + { + "epoch": 1.644959367450033, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2909746170043945, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7137229442596436, + "num_tokens": 387343059.0, + "step": 14979 + }, + { + "epoch": 1.6450691851526464, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1177430152893066, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.722210168838501, + "num_tokens": 387372396.0, + "step": 14980 + }, + { + "epoch": 1.6451790028552602, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.8572198152542114, + "learning_rate": 1e-06, + "loss": 1.032, + "mean_token_accuracy": 0.6806944608688354, + "num_tokens": 387405565.0, + "step": 14981 + }, + { + "epoch": 1.645288820557874, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.519216299057007, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7399488091468811, + "num_tokens": 387425517.0, + "step": 14982 + }, + { + "epoch": 1.6453986382604877, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.564276933670044, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7415976524353027, + "num_tokens": 387445500.0, + "step": 14983 + }, + { + "epoch": 1.6455084559631012, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.645859718322754, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7341253757476807, + "num_tokens": 387464702.0, + "step": 14984 + }, + { + "epoch": 1.6456182736657148, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.266625165939331, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7097175717353821, + "num_tokens": 387489147.0, + "step": 14985 + }, + { + "epoch": 1.6457280913683285, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0657215118408203, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.707905113697052, + "num_tokens": 387516359.0, + "step": 14986 + }, + { + "epoch": 1.6458379090709423, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3837757110595703, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.719346284866333, + "num_tokens": 387539169.0, + "step": 14987 + }, + { + "epoch": 1.645947726773556, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4563608169555664, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.743377685546875, + "num_tokens": 387561416.0, + "step": 14988 + }, + { + "epoch": 1.6460575444761696, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1559410095214844, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6942702531814575, + "num_tokens": 387588704.0, + "step": 14989 + }, + { + "epoch": 1.6461673621787831, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.60843563079834, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7189381122589111, + "num_tokens": 387609671.0, + "step": 14990 + }, + { + "epoch": 1.6462771798813969, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9705361127853394, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6988359689712524, + "num_tokens": 387644406.0, + "step": 14991 + }, + { + "epoch": 1.6463869975840106, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9844623804092407, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7185996770858765, + "num_tokens": 387675385.0, + "step": 14992 + }, + { + "epoch": 1.6464968152866242, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0789642333984375, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7291709184646606, + "num_tokens": 387702653.0, + "step": 14993 + }, + { + "epoch": 1.6466066329892377, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.324049711227417, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7420543432235718, + "num_tokens": 387726744.0, + "step": 14994 + }, + { + "epoch": 1.6467164506918515, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4077441692352295, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.72724848985672, + "num_tokens": 387749051.0, + "step": 14995 + }, + { + "epoch": 1.6468262683944652, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2757627964019775, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7118775844573975, + "num_tokens": 387773399.0, + "step": 14996 + }, + { + "epoch": 1.646936086097079, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.326418399810791, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7229008674621582, + "num_tokens": 387798562.0, + "step": 14997 + }, + { + "epoch": 1.6470459037996925, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3118746280670166, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7140805125236511, + "num_tokens": 387825123.0, + "step": 14998 + }, + { + "epoch": 1.647155721502306, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.065823793411255, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7079377174377441, + "num_tokens": 387855785.0, + "step": 14999 + }, + { + "epoch": 1.6472655392049198, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1296863555908203, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7131410837173462, + "num_tokens": 387885037.0, + "step": 15000 + }, + { + "epoch": 1.6473753569075336, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.7398858070373535, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7367163896560669, + "num_tokens": 387903513.0, + "step": 15001 + }, + { + "epoch": 1.6474851746101473, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9788788557052612, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7002468109130859, + "num_tokens": 387935721.0, + "step": 15002 + }, + { + "epoch": 1.6475949923127609, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1731064319610596, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7247109413146973, + "num_tokens": 387960680.0, + "step": 15003 + }, + { + "epoch": 1.6477048100153744, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1661272048950195, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7003297209739685, + "num_tokens": 387990376.0, + "step": 15004 + }, + { + "epoch": 1.6478146277179881, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3208134174346924, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7318114042282104, + "num_tokens": 388013280.0, + "step": 15005 + }, + { + "epoch": 1.647924445420602, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0911569595336914, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7031527757644653, + "num_tokens": 388044228.0, + "step": 15006 + }, + { + "epoch": 1.6480342631232154, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.161423444747925, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7193666100502014, + "num_tokens": 388069822.0, + "step": 15007 + }, + { + "epoch": 1.648144080825829, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9764519929885864, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7057487964630127, + "num_tokens": 388102749.0, + "step": 15008 + }, + { + "epoch": 1.6482538985284427, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6217901706695557, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7372138500213623, + "num_tokens": 388122017.0, + "step": 15009 + }, + { + "epoch": 1.6483637162310565, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4395296573638916, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7085052728652954, + "num_tokens": 388146361.0, + "step": 15010 + }, + { + "epoch": 1.6484735339336702, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.38107967376709, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.698027491569519, + "num_tokens": 388170621.0, + "step": 15011 + }, + { + "epoch": 1.6485833516362838, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1821000576019287, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7239314317703247, + "num_tokens": 388198618.0, + "step": 15012 + }, + { + "epoch": 1.6486931693388973, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.271732807159424, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7091233730316162, + "num_tokens": 388224203.0, + "step": 15013 + }, + { + "epoch": 1.648802987041511, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1244406700134277, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6943936944007874, + "num_tokens": 388256049.0, + "step": 15014 + }, + { + "epoch": 1.6489128047441248, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.282010793685913, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.720308244228363, + "num_tokens": 388280606.0, + "step": 15015 + }, + { + "epoch": 1.6490226224467384, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4708099365234375, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7054160833358765, + "num_tokens": 388305205.0, + "step": 15016 + }, + { + "epoch": 1.6491324401493521, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1215615272521973, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.714780330657959, + "num_tokens": 388335252.0, + "step": 15017 + }, + { + "epoch": 1.6492422578519657, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.559744119644165, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7308794260025024, + "num_tokens": 388358098.0, + "step": 15018 + }, + { + "epoch": 1.6493520755545794, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2971909046173096, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7441916465759277, + "num_tokens": 388381800.0, + "step": 15019 + }, + { + "epoch": 1.6494618932571932, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2719976902008057, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7049851417541504, + "num_tokens": 388406651.0, + "step": 15020 + }, + { + "epoch": 1.6495717109598067, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.371066093444824, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7471967339515686, + "num_tokens": 388431031.0, + "step": 15021 + }, + { + "epoch": 1.6496815286624202, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.139897346496582, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.703116774559021, + "num_tokens": 388459428.0, + "step": 15022 + }, + { + "epoch": 1.649791346365034, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0998237133026123, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7285200357437134, + "num_tokens": 388488077.0, + "step": 15023 + }, + { + "epoch": 1.6499011640676478, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.122645616531372, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7211238145828247, + "num_tokens": 388516556.0, + "step": 15024 + }, + { + "epoch": 1.6500109817702615, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.478350877761841, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7064462900161743, + "num_tokens": 388542273.0, + "step": 15025 + }, + { + "epoch": 1.650120799472875, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.184785842895508, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7216041684150696, + "num_tokens": 388568517.0, + "step": 15026 + }, + { + "epoch": 1.6502306171754886, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3494205474853516, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7240645885467529, + "num_tokens": 388590410.0, + "step": 15027 + }, + { + "epoch": 1.6503404348781023, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2547662258148193, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7466899752616882, + "num_tokens": 388614421.0, + "step": 15028 + }, + { + "epoch": 1.650450252580716, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.011775255203247, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7347453832626343, + "num_tokens": 388643445.0, + "step": 15029 + }, + { + "epoch": 1.6505600702833296, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.244443893432617, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.6998761892318726, + "num_tokens": 388669637.0, + "step": 15030 + }, + { + "epoch": 1.6506698879859434, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3603875637054443, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7292147874832153, + "num_tokens": 388692316.0, + "step": 15031 + }, + { + "epoch": 1.650779705688557, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5281317234039307, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7380869388580322, + "num_tokens": 388714361.0, + "step": 15032 + }, + { + "epoch": 1.6508895233911707, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1294424533843994, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7135063409805298, + "num_tokens": 388742386.0, + "step": 15033 + }, + { + "epoch": 1.6509993410937844, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4732534885406494, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7105076313018799, + "num_tokens": 388765011.0, + "step": 15034 + }, + { + "epoch": 1.651109158796398, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3171186447143555, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7098652124404907, + "num_tokens": 388790939.0, + "step": 15035 + }, + { + "epoch": 1.6512189764990115, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.309911012649536, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.720229983329773, + "num_tokens": 388816139.0, + "step": 15036 + }, + { + "epoch": 1.6513287942016253, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.375493288040161, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7188321352005005, + "num_tokens": 388841709.0, + "step": 15037 + }, + { + "epoch": 1.651438611904239, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3460779190063477, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7116090655326843, + "num_tokens": 388868760.0, + "step": 15038 + }, + { + "epoch": 1.6515484296068528, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.403287172317505, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7274084091186523, + "num_tokens": 388889984.0, + "step": 15039 + }, + { + "epoch": 1.6516582473094663, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1410038471221924, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7241086959838867, + "num_tokens": 388915538.0, + "step": 15040 + }, + { + "epoch": 1.6517680650120798, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.50909161567688, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.740847647190094, + "num_tokens": 388935827.0, + "step": 15041 + }, + { + "epoch": 1.6518778827146936, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4177520275115967, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7155587077140808, + "num_tokens": 388959234.0, + "step": 15042 + }, + { + "epoch": 1.6519877004173074, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.286118268966675, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7093609571456909, + "num_tokens": 388985059.0, + "step": 15043 + }, + { + "epoch": 1.652097518119921, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.416687250137329, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6968308687210083, + "num_tokens": 389008956.0, + "step": 15044 + }, + { + "epoch": 1.6522073358225344, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.302842140197754, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7026482820510864, + "num_tokens": 389035428.0, + "step": 15045 + }, + { + "epoch": 1.6523171535251482, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.302562952041626, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7171931266784668, + "num_tokens": 389060170.0, + "step": 15046 + }, + { + "epoch": 1.652426971227762, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.175849676132202, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7053979635238647, + "num_tokens": 389087000.0, + "step": 15047 + }, + { + "epoch": 1.6525367889303757, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.371983766555786, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7033271193504333, + "num_tokens": 389110650.0, + "step": 15048 + }, + { + "epoch": 1.6526466066329892, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.512543201446533, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7274094820022583, + "num_tokens": 389131713.0, + "step": 15049 + }, + { + "epoch": 1.6527564243356028, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3488433361053467, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7003499269485474, + "num_tokens": 389155984.0, + "step": 15050 + }, + { + "epoch": 1.6528662420382165, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.598545789718628, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.721350908279419, + "num_tokens": 389175024.0, + "step": 15051 + }, + { + "epoch": 1.6529760597408303, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3727781772613525, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7223292589187622, + "num_tokens": 389198396.0, + "step": 15052 + }, + { + "epoch": 1.653085877443444, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2344319820404053, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7168499231338501, + "num_tokens": 389224011.0, + "step": 15053 + }, + { + "epoch": 1.6531956951460576, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.09264874458313, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7258786559104919, + "num_tokens": 389250356.0, + "step": 15054 + }, + { + "epoch": 1.653305512848671, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2751944065093994, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7258060574531555, + "num_tokens": 389274008.0, + "step": 15055 + }, + { + "epoch": 1.6534153305512849, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4014315605163574, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6867660284042358, + "num_tokens": 389297790.0, + "step": 15056 + }, + { + "epoch": 1.6535251482538986, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9362541437149048, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7047814726829529, + "num_tokens": 389329161.0, + "step": 15057 + }, + { + "epoch": 1.6536349659565122, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1683833599090576, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.728017270565033, + "num_tokens": 389355768.0, + "step": 15058 + }, + { + "epoch": 1.6537447836591257, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9760838747024536, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.71497642993927, + "num_tokens": 389386013.0, + "step": 15059 + }, + { + "epoch": 1.6538546013617395, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2445626258850098, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7054339647293091, + "num_tokens": 389412209.0, + "step": 15060 + }, + { + "epoch": 1.6539644190643532, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.895403504371643, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7005829215049744, + "num_tokens": 389446399.0, + "step": 15061 + }, + { + "epoch": 1.654074236766967, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1603198051452637, + "learning_rate": 1e-06, + "loss": 0.7609, + "mean_token_accuracy": 0.7572690844535828, + "num_tokens": 389471148.0, + "step": 15062 + }, + { + "epoch": 1.6541840544695805, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.886390209197998, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6971426010131836, + "num_tokens": 389508397.0, + "step": 15063 + }, + { + "epoch": 1.654293872172194, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2783775329589844, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7132105827331543, + "num_tokens": 389535562.0, + "step": 15064 + }, + { + "epoch": 1.6544036898748078, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.235750913619995, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7357910871505737, + "num_tokens": 389563004.0, + "step": 15065 + }, + { + "epoch": 1.6545135075774215, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.322596788406372, + "learning_rate": 1e-06, + "loss": 0.7626, + "mean_token_accuracy": 0.756945788860321, + "num_tokens": 389586088.0, + "step": 15066 + }, + { + "epoch": 1.6546233252800353, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2088406085968018, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7331253886222839, + "num_tokens": 389612021.0, + "step": 15067 + }, + { + "epoch": 1.6547331429826488, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.954447865486145, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7238668203353882, + "num_tokens": 389642765.0, + "step": 15068 + }, + { + "epoch": 1.6548429606852624, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.310988664627075, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7392539978027344, + "num_tokens": 389665514.0, + "step": 15069 + }, + { + "epoch": 1.6549527783878761, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.240971565246582, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7090511322021484, + "num_tokens": 389692152.0, + "step": 15070 + }, + { + "epoch": 1.65506259609049, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3988687992095947, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7227369546890259, + "num_tokens": 389717493.0, + "step": 15071 + }, + { + "epoch": 1.6551724137931034, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.378612995147705, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.710345983505249, + "num_tokens": 389741451.0, + "step": 15072 + }, + { + "epoch": 1.655282231495717, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3372201919555664, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7189551591873169, + "num_tokens": 389767007.0, + "step": 15073 + }, + { + "epoch": 1.6553920491983307, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5851550102233887, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.6990848779678345, + "num_tokens": 389788434.0, + "step": 15074 + }, + { + "epoch": 1.6555018669009445, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1633639335632324, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7028902769088745, + "num_tokens": 389815287.0, + "step": 15075 + }, + { + "epoch": 1.6556116846035582, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.454815626144409, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7399143576622009, + "num_tokens": 389836107.0, + "step": 15076 + }, + { + "epoch": 1.6557215023061718, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.250319480895996, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7286849617958069, + "num_tokens": 389860566.0, + "step": 15077 + }, + { + "epoch": 1.6558313200087853, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9652390480041504, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7292677164077759, + "num_tokens": 389891930.0, + "step": 15078 + }, + { + "epoch": 1.655941137711399, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3282318115234375, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7097578644752502, + "num_tokens": 389916670.0, + "step": 15079 + }, + { + "epoch": 1.6560509554140128, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.105456829071045, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.698172926902771, + "num_tokens": 389945835.0, + "step": 15080 + }, + { + "epoch": 1.6561607731166264, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1481266021728516, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7558701634407043, + "num_tokens": 389972301.0, + "step": 15081 + }, + { + "epoch": 1.65627059081924, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6233773231506348, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7481239438056946, + "num_tokens": 389992002.0, + "step": 15082 + }, + { + "epoch": 1.6563804085218536, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3324403762817383, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7015138864517212, + "num_tokens": 390018505.0, + "step": 15083 + }, + { + "epoch": 1.6564902262244674, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4429855346679688, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7302486896514893, + "num_tokens": 390040048.0, + "step": 15084 + }, + { + "epoch": 1.6566000439270812, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.072445869445801, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7101930975914001, + "num_tokens": 390068920.0, + "step": 15085 + }, + { + "epoch": 1.6567098616296947, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3580760955810547, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.6967202425003052, + "num_tokens": 390095080.0, + "step": 15086 + }, + { + "epoch": 1.6568196793323082, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.088254451751709, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7268460988998413, + "num_tokens": 390123902.0, + "step": 15087 + }, + { + "epoch": 1.656929497034922, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.225585460662842, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7046492099761963, + "num_tokens": 390149633.0, + "step": 15088 + }, + { + "epoch": 1.6570393147375357, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.44960618019104, + "learning_rate": 1e-06, + "loss": 1.0333, + "mean_token_accuracy": 0.7022587656974792, + "num_tokens": 390173020.0, + "step": 15089 + }, + { + "epoch": 1.6571491324401495, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3343279361724854, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7124190330505371, + "num_tokens": 390198722.0, + "step": 15090 + }, + { + "epoch": 1.657258950142763, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2428464889526367, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7292571067810059, + "num_tokens": 390226112.0, + "step": 15091 + }, + { + "epoch": 1.6573687678453766, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2413442134857178, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7338770627975464, + "num_tokens": 390250826.0, + "step": 15092 + }, + { + "epoch": 1.6574785855479903, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3807718753814697, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7266075015068054, + "num_tokens": 390273521.0, + "step": 15093 + }, + { + "epoch": 1.657588403250604, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.298342704772949, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7259820103645325, + "num_tokens": 390300333.0, + "step": 15094 + }, + { + "epoch": 1.6576982209532176, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1709234714508057, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7501653432846069, + "num_tokens": 390327090.0, + "step": 15095 + }, + { + "epoch": 1.6578080386558314, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.362591505050659, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.7056379318237305, + "num_tokens": 390352112.0, + "step": 15096 + }, + { + "epoch": 1.657917856358445, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1691977977752686, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7379158735275269, + "num_tokens": 390379128.0, + "step": 15097 + }, + { + "epoch": 1.6580276740610587, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3807201385498047, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7185507416725159, + "num_tokens": 390402702.0, + "step": 15098 + }, + { + "epoch": 1.6581374917636724, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.601544141769409, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7086585164070129, + "num_tokens": 390425482.0, + "step": 15099 + }, + { + "epoch": 1.658247309466286, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3776450157165527, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7228654623031616, + "num_tokens": 390448680.0, + "step": 15100 + }, + { + "epoch": 1.6583571271688995, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1548266410827637, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6961226463317871, + "num_tokens": 390475776.0, + "step": 15101 + }, + { + "epoch": 1.6584669448715132, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3340978622436523, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7202688455581665, + "num_tokens": 390499965.0, + "step": 15102 + }, + { + "epoch": 1.658576762574127, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.248106002807617, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7287039160728455, + "num_tokens": 390525663.0, + "step": 15103 + }, + { + "epoch": 1.6586865802767408, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.359624147415161, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7072227001190186, + "num_tokens": 390549509.0, + "step": 15104 + }, + { + "epoch": 1.6587963979793543, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.219546318054199, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.6997638940811157, + "num_tokens": 390577154.0, + "step": 15105 + }, + { + "epoch": 1.6589062156819678, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1022942066192627, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.712737500667572, + "num_tokens": 390606950.0, + "step": 15106 + }, + { + "epoch": 1.6590160333845816, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3963112831115723, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.705636203289032, + "num_tokens": 390630059.0, + "step": 15107 + }, + { + "epoch": 1.6591258510871953, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1480228900909424, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7185503840446472, + "num_tokens": 390656855.0, + "step": 15108 + }, + { + "epoch": 1.6592356687898089, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3878612518310547, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7464026212692261, + "num_tokens": 390681132.0, + "step": 15109 + }, + { + "epoch": 1.6593454864924224, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1470627784729004, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7173848152160645, + "num_tokens": 390710201.0, + "step": 15110 + }, + { + "epoch": 1.6594553041950362, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2252988815307617, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7381166219711304, + "num_tokens": 390735305.0, + "step": 15111 + }, + { + "epoch": 1.65956512189765, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.475647211074829, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7190302014350891, + "num_tokens": 390760313.0, + "step": 15112 + }, + { + "epoch": 1.6596749396002637, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.328392267227173, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.723487377166748, + "num_tokens": 390782847.0, + "step": 15113 + }, + { + "epoch": 1.6597847573028772, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3143811225891113, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7144492864608765, + "num_tokens": 390808668.0, + "step": 15114 + }, + { + "epoch": 1.6598945750054908, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2800865173339844, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7100621461868286, + "num_tokens": 390835133.0, + "step": 15115 + }, + { + "epoch": 1.6600043927081045, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.231243371963501, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7270114421844482, + "num_tokens": 390860844.0, + "step": 15116 + }, + { + "epoch": 1.6601142104107183, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6494271755218506, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7187548875808716, + "num_tokens": 390881447.0, + "step": 15117 + }, + { + "epoch": 1.660224028113332, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9891319274902344, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7166403532028198, + "num_tokens": 390914311.0, + "step": 15118 + }, + { + "epoch": 1.6603338458159456, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5319790840148926, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7428097724914551, + "num_tokens": 390936130.0, + "step": 15119 + }, + { + "epoch": 1.660443663518559, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.042025327682495, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7333682775497437, + "num_tokens": 390963900.0, + "step": 15120 + }, + { + "epoch": 1.6605534812211729, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1607625484466553, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7421308755874634, + "num_tokens": 390988318.0, + "step": 15121 + }, + { + "epoch": 1.6606632989237866, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3355445861816406, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7172439098358154, + "num_tokens": 391013088.0, + "step": 15122 + }, + { + "epoch": 1.6607731166264001, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.276259660720825, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7136531472206116, + "num_tokens": 391038533.0, + "step": 15123 + }, + { + "epoch": 1.6608829343290137, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3289718627929688, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7160850167274475, + "num_tokens": 391063810.0, + "step": 15124 + }, + { + "epoch": 1.6609927520316274, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2801194190979004, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7031409740447998, + "num_tokens": 391089542.0, + "step": 15125 + }, + { + "epoch": 1.6611025697342412, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2671501636505127, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7409751415252686, + "num_tokens": 391113833.0, + "step": 15126 + }, + { + "epoch": 1.661212387436855, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.281752347946167, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.727960467338562, + "num_tokens": 391138425.0, + "step": 15127 + }, + { + "epoch": 1.6613222051394685, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3532323837280273, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7411825656890869, + "num_tokens": 391161315.0, + "step": 15128 + }, + { + "epoch": 1.661432022842082, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2678964138031006, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7273492813110352, + "num_tokens": 391185436.0, + "step": 15129 + }, + { + "epoch": 1.6615418405446958, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.385873317718506, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7241618037223816, + "num_tokens": 391209066.0, + "step": 15130 + }, + { + "epoch": 1.6616516582473095, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5658302307128906, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7212235927581787, + "num_tokens": 391229470.0, + "step": 15131 + }, + { + "epoch": 1.661761475949923, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.037827730178833, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7177349925041199, + "num_tokens": 391259599.0, + "step": 15132 + }, + { + "epoch": 1.6618712936525368, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2880969047546387, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7278767228126526, + "num_tokens": 391284759.0, + "step": 15133 + }, + { + "epoch": 1.6619811113551504, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.274585485458374, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7239365577697754, + "num_tokens": 391310964.0, + "step": 15134 + }, + { + "epoch": 1.6620909290577641, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2486090660095215, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.727439284324646, + "num_tokens": 391336718.0, + "step": 15135 + }, + { + "epoch": 1.6622007467603779, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1934869289398193, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.707068145275116, + "num_tokens": 391363522.0, + "step": 15136 + }, + { + "epoch": 1.6623105644629914, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1406490802764893, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7220872640609741, + "num_tokens": 391390172.0, + "step": 15137 + }, + { + "epoch": 1.662420382165605, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1905245780944824, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7121171951293945, + "num_tokens": 391416941.0, + "step": 15138 + }, + { + "epoch": 1.6625301998682187, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0637476444244385, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7045378684997559, + "num_tokens": 391444599.0, + "step": 15139 + }, + { + "epoch": 1.6626400175708325, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.174961566925049, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7059481143951416, + "num_tokens": 391471679.0, + "step": 15140 + }, + { + "epoch": 1.6627498352734462, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1318492889404297, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7141409516334534, + "num_tokens": 391500427.0, + "step": 15141 + }, + { + "epoch": 1.6628596529760598, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1885933876037598, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7501933574676514, + "num_tokens": 391526531.0, + "step": 15142 + }, + { + "epoch": 1.6629694706786733, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.192016839981079, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7473524808883667, + "num_tokens": 391549471.0, + "step": 15143 + }, + { + "epoch": 1.663079288381287, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3097028732299805, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7264804840087891, + "num_tokens": 391572180.0, + "step": 15144 + }, + { + "epoch": 1.6631891060839008, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.236347198486328, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7231342196464539, + "num_tokens": 391598234.0, + "step": 15145 + }, + { + "epoch": 1.6632989237865143, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0873374938964844, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7020723223686218, + "num_tokens": 391627804.0, + "step": 15146 + }, + { + "epoch": 1.663408741489128, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6415209770202637, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7300519943237305, + "num_tokens": 391650198.0, + "step": 15147 + }, + { + "epoch": 1.6635185591917416, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.136526107788086, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.72019362449646, + "num_tokens": 391679974.0, + "step": 15148 + }, + { + "epoch": 1.6636283768943554, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2428035736083984, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7145153284072876, + "num_tokens": 391706421.0, + "step": 15149 + }, + { + "epoch": 1.6637381945969691, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0212595462799072, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.71403968334198, + "num_tokens": 391737789.0, + "step": 15150 + }, + { + "epoch": 1.6638480122995827, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.013640880584717, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.730732798576355, + "num_tokens": 391767656.0, + "step": 15151 + }, + { + "epoch": 1.6639578300021962, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3581795692443848, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7273661494255066, + "num_tokens": 391788356.0, + "step": 15152 + }, + { + "epoch": 1.66406764770481, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1209237575531006, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7154799699783325, + "num_tokens": 391817753.0, + "step": 15153 + }, + { + "epoch": 1.6641774654074237, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2811098098754883, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.722411036491394, + "num_tokens": 391842091.0, + "step": 15154 + }, + { + "epoch": 1.6642872831100375, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.17091703414917, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7055323123931885, + "num_tokens": 391870270.0, + "step": 15155 + }, + { + "epoch": 1.664397100812651, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.067474603652954, + "learning_rate": 1e-06, + "loss": 1.0784, + "mean_token_accuracy": 0.679688572883606, + "num_tokens": 391903803.0, + "step": 15156 + }, + { + "epoch": 1.6645069185152646, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.36661434173584, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7271634340286255, + "num_tokens": 391926968.0, + "step": 15157 + }, + { + "epoch": 1.6646167362178783, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2176389694213867, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7057790756225586, + "num_tokens": 391954503.0, + "step": 15158 + }, + { + "epoch": 1.664726553920492, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5468573570251465, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7185897827148438, + "num_tokens": 391976220.0, + "step": 15159 + }, + { + "epoch": 1.6648363716231056, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.155890703201294, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.715045690536499, + "num_tokens": 392003237.0, + "step": 15160 + }, + { + "epoch": 1.6649461893257194, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9878880977630615, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7217432856559753, + "num_tokens": 392032988.0, + "step": 15161 + }, + { + "epoch": 1.665056007028333, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.231717348098755, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7384288311004639, + "num_tokens": 392057474.0, + "step": 15162 + }, + { + "epoch": 1.6651658247309467, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.229985237121582, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7187424898147583, + "num_tokens": 392084909.0, + "step": 15163 + }, + { + "epoch": 1.6652756424335604, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2048861980438232, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7060086131095886, + "num_tokens": 392111931.0, + "step": 15164 + }, + { + "epoch": 1.665385460136174, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.190387725830078, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.733197033405304, + "num_tokens": 392138413.0, + "step": 15165 + }, + { + "epoch": 1.6654952778387875, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.232780933380127, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7255085706710815, + "num_tokens": 392164030.0, + "step": 15166 + }, + { + "epoch": 1.6656050955414012, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.222989082336426, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.718481183052063, + "num_tokens": 392189971.0, + "step": 15167 + }, + { + "epoch": 1.665714913244015, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1235828399658203, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7227303981781006, + "num_tokens": 392217168.0, + "step": 15168 + }, + { + "epoch": 1.6658247309466288, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1283841133117676, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7185099124908447, + "num_tokens": 392245303.0, + "step": 15169 + }, + { + "epoch": 1.6659345486492423, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1712865829467773, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.731546938419342, + "num_tokens": 392270791.0, + "step": 15170 + }, + { + "epoch": 1.6660443663518558, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3265907764434814, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7148058414459229, + "num_tokens": 392295622.0, + "step": 15171 + }, + { + "epoch": 1.6661541840544696, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2204456329345703, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7109130620956421, + "num_tokens": 392322785.0, + "step": 15172 + }, + { + "epoch": 1.6662640017570833, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.5980546474456787, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.727645754814148, + "num_tokens": 392343448.0, + "step": 15173 + }, + { + "epoch": 1.6663738194596969, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6791725158691406, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7336667776107788, + "num_tokens": 392362200.0, + "step": 15174 + }, + { + "epoch": 1.6664836371623104, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.074390411376953, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7173739671707153, + "num_tokens": 392390515.0, + "step": 15175 + }, + { + "epoch": 1.6665934548649242, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.198744297027588, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7184361815452576, + "num_tokens": 392416679.0, + "step": 15176 + }, + { + "epoch": 1.666703272567538, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4113361835479736, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7154839038848877, + "num_tokens": 392441347.0, + "step": 15177 + }, + { + "epoch": 1.6668130902701517, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.079946756362915, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7107059955596924, + "num_tokens": 392472060.0, + "step": 15178 + }, + { + "epoch": 1.6669229079727652, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3177998065948486, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.700336217880249, + "num_tokens": 392494595.0, + "step": 15179 + }, + { + "epoch": 1.6670327256753787, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.32261061668396, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7228748798370361, + "num_tokens": 392521890.0, + "step": 15180 + }, + { + "epoch": 1.6671425433779925, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.43572998046875, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.717311441898346, + "num_tokens": 392543064.0, + "step": 15181 + }, + { + "epoch": 1.6672523610806063, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3793067932128906, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7291797399520874, + "num_tokens": 392565624.0, + "step": 15182 + }, + { + "epoch": 1.66736217878322, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.407341718673706, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7339204549789429, + "num_tokens": 392588081.0, + "step": 15183 + }, + { + "epoch": 1.6674719964858336, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.222411632537842, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7231179475784302, + "num_tokens": 392615830.0, + "step": 15184 + }, + { + "epoch": 1.667581814188447, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0947840213775635, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7342733144760132, + "num_tokens": 392643642.0, + "step": 15185 + }, + { + "epoch": 1.6676916318910608, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2212085723876953, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7137681245803833, + "num_tokens": 392668842.0, + "step": 15186 + }, + { + "epoch": 1.6678014495936746, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4174294471740723, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7328035831451416, + "num_tokens": 392691727.0, + "step": 15187 + }, + { + "epoch": 1.6679112672962881, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2104551792144775, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7157961130142212, + "num_tokens": 392718206.0, + "step": 15188 + }, + { + "epoch": 1.6680210849989017, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9174315929412842, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7193906307220459, + "num_tokens": 392749720.0, + "step": 15189 + }, + { + "epoch": 1.6681309027015154, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.241044044494629, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7039548754692078, + "num_tokens": 392775450.0, + "step": 15190 + }, + { + "epoch": 1.6682407204041292, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.158552885055542, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7110718488693237, + "num_tokens": 392805725.0, + "step": 15191 + }, + { + "epoch": 1.668350538106743, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.329810857772827, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7133508324623108, + "num_tokens": 392828935.0, + "step": 15192 + }, + { + "epoch": 1.6684603558093565, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3519370555877686, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7336460947990417, + "num_tokens": 392852371.0, + "step": 15193 + }, + { + "epoch": 1.66857017351197, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.264641761779785, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7230087518692017, + "num_tokens": 392876515.0, + "step": 15194 + }, + { + "epoch": 1.6686799912145838, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0344343185424805, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7124453186988831, + "num_tokens": 392909674.0, + "step": 15195 + }, + { + "epoch": 1.6687898089171975, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0171351432800293, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7050642967224121, + "num_tokens": 392942641.0, + "step": 15196 + }, + { + "epoch": 1.668899626619811, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.8857654333114624, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7054948806762695, + "num_tokens": 392976520.0, + "step": 15197 + }, + { + "epoch": 1.6690094443224248, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4448373317718506, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7118567824363708, + "num_tokens": 392999207.0, + "step": 15198 + }, + { + "epoch": 1.6691192620250384, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.9899057149887085, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7115827202796936, + "num_tokens": 393029558.0, + "step": 15199 + }, + { + "epoch": 1.669229079727652, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2670280933380127, + "learning_rate": 1e-06, + "loss": 0.774, + "mean_token_accuracy": 0.7527908682823181, + "num_tokens": 393051392.0, + "step": 15200 + }, + { + "epoch": 1.6693388974302659, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2143781185150146, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7479676008224487, + "num_tokens": 393076234.0, + "step": 15201 + }, + { + "epoch": 1.6694487151328794, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2955596446990967, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7400338053703308, + "num_tokens": 393099892.0, + "step": 15202 + }, + { + "epoch": 1.669558532835493, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1260478496551514, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7127777934074402, + "num_tokens": 393128596.0, + "step": 15203 + }, + { + "epoch": 1.6696683505381067, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.067187547683716, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7118790149688721, + "num_tokens": 393159301.0, + "step": 15204 + }, + { + "epoch": 1.6697781682407205, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.432915210723877, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7258087396621704, + "num_tokens": 393181250.0, + "step": 15205 + }, + { + "epoch": 1.6698879859433342, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.198164701461792, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7035906910896301, + "num_tokens": 393208502.0, + "step": 15206 + }, + { + "epoch": 1.6699978036459477, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1320929527282715, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.719784140586853, + "num_tokens": 393234328.0, + "step": 15207 + }, + { + "epoch": 1.6701076213485613, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.38423228263855, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.722232460975647, + "num_tokens": 393257177.0, + "step": 15208 + }, + { + "epoch": 1.670217439051175, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.293827533721924, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6917287111282349, + "num_tokens": 393283471.0, + "step": 15209 + }, + { + "epoch": 1.6703272567537888, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.464782476425171, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7326076626777649, + "num_tokens": 393305955.0, + "step": 15210 + }, + { + "epoch": 1.6704370744564023, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.199270486831665, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7213218212127686, + "num_tokens": 393333716.0, + "step": 15211 + }, + { + "epoch": 1.670546892159016, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.094019889831543, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7293468117713928, + "num_tokens": 393362875.0, + "step": 15212 + }, + { + "epoch": 1.6706567098616296, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4842584133148193, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7239950299263, + "num_tokens": 393383648.0, + "step": 15213 + }, + { + "epoch": 1.6707665275642434, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2613508701324463, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7149418592453003, + "num_tokens": 393410602.0, + "step": 15214 + }, + { + "epoch": 1.6708763452668571, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.367671489715576, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7274975180625916, + "num_tokens": 393433729.0, + "step": 15215 + }, + { + "epoch": 1.6709861629694707, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.338928461074829, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7141909599304199, + "num_tokens": 393457823.0, + "step": 15216 + }, + { + "epoch": 1.6710959806720842, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0779542922973633, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.70367431640625, + "num_tokens": 393487977.0, + "step": 15217 + }, + { + "epoch": 1.671205798374698, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0916688442230225, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6891316175460815, + "num_tokens": 393520805.0, + "step": 15218 + }, + { + "epoch": 1.6713156160773117, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.040339231491089, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7073361277580261, + "num_tokens": 393550646.0, + "step": 15219 + }, + { + "epoch": 1.6714254337799255, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3222720623016357, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7349337339401245, + "num_tokens": 393572080.0, + "step": 15220 + }, + { + "epoch": 1.671535251482539, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.224350929260254, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7078202366828918, + "num_tokens": 393599908.0, + "step": 15221 + }, + { + "epoch": 1.6716450691851525, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.345438003540039, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7321572303771973, + "num_tokens": 393622182.0, + "step": 15222 + }, + { + "epoch": 1.6717548868877663, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1988556385040283, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7208721041679382, + "num_tokens": 393647899.0, + "step": 15223 + }, + { + "epoch": 1.67186470459038, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3560023307800293, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.73372483253479, + "num_tokens": 393671432.0, + "step": 15224 + }, + { + "epoch": 1.6719745222929936, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2601215839385986, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7320664525032043, + "num_tokens": 393695873.0, + "step": 15225 + }, + { + "epoch": 1.6720843399956071, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.348377227783203, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7235819101333618, + "num_tokens": 393721029.0, + "step": 15226 + }, + { + "epoch": 1.6721941576982209, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0217106342315674, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7165477871894836, + "num_tokens": 393751345.0, + "step": 15227 + }, + { + "epoch": 1.6723039754008346, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.197650909423828, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7129183411598206, + "num_tokens": 393777302.0, + "step": 15228 + }, + { + "epoch": 1.6724137931034484, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0924386978149414, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.6998003721237183, + "num_tokens": 393805391.0, + "step": 15229 + }, + { + "epoch": 1.672523610806062, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.643444776535034, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7084434032440186, + "num_tokens": 393825653.0, + "step": 15230 + }, + { + "epoch": 1.6726334285086755, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1635327339172363, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.70321124792099, + "num_tokens": 393852131.0, + "step": 15231 + }, + { + "epoch": 1.6727432462112892, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.150848627090454, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.72182297706604, + "num_tokens": 393879563.0, + "step": 15232 + }, + { + "epoch": 1.672853063913903, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0781829357147217, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.709320068359375, + "num_tokens": 393908384.0, + "step": 15233 + }, + { + "epoch": 1.6729628816165167, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.553999185562134, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7185966968536377, + "num_tokens": 393929500.0, + "step": 15234 + }, + { + "epoch": 1.6730726993191303, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.296645402908325, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7472164630889893, + "num_tokens": 393952337.0, + "step": 15235 + }, + { + "epoch": 1.6731825170217438, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2013096809387207, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7040042877197266, + "num_tokens": 393978801.0, + "step": 15236 + }, + { + "epoch": 1.6732923347243576, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3253707885742188, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7125789523124695, + "num_tokens": 394006803.0, + "step": 15237 + }, + { + "epoch": 1.6734021524269713, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.256772041320801, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7111546993255615, + "num_tokens": 394031596.0, + "step": 15238 + }, + { + "epoch": 1.6735119701295849, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.561891794204712, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7387721538543701, + "num_tokens": 394053285.0, + "step": 15239 + }, + { + "epoch": 1.6736217878321984, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.074474334716797, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7333698868751526, + "num_tokens": 394081476.0, + "step": 15240 + }, + { + "epoch": 1.6737316055348122, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.041830539703369, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7221139669418335, + "num_tokens": 394109698.0, + "step": 15241 + }, + { + "epoch": 1.673841423237426, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.250965118408203, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7537273168563843, + "num_tokens": 394133480.0, + "step": 15242 + }, + { + "epoch": 1.6739512409400397, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3026747703552246, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7298545837402344, + "num_tokens": 394159412.0, + "step": 15243 + }, + { + "epoch": 1.6740610586426532, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.230794906616211, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.712770938873291, + "num_tokens": 394185616.0, + "step": 15244 + }, + { + "epoch": 1.6741708763452667, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0766212940216064, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7308707237243652, + "num_tokens": 394214586.0, + "step": 15245 + }, + { + "epoch": 1.6742806940478805, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.178978681564331, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.709639310836792, + "num_tokens": 394241550.0, + "step": 15246 + }, + { + "epoch": 1.6743905117504942, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3328611850738525, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7318832874298096, + "num_tokens": 394266073.0, + "step": 15247 + }, + { + "epoch": 1.674500329453108, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3335719108581543, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7179057598114014, + "num_tokens": 394292595.0, + "step": 15248 + }, + { + "epoch": 1.6746101471557215, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.068781614303589, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6942031383514404, + "num_tokens": 394323940.0, + "step": 15249 + }, + { + "epoch": 1.674719964858335, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3282594680786133, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7079934477806091, + "num_tokens": 394349290.0, + "step": 15250 + }, + { + "epoch": 1.6748297825609488, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.6301586627960205, + "learning_rate": 1e-06, + "loss": 0.7817, + "mean_token_accuracy": 0.7593165636062622, + "num_tokens": 394368900.0, + "step": 15251 + }, + { + "epoch": 1.6749396002635626, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4257426261901855, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7255878448486328, + "num_tokens": 394390688.0, + "step": 15252 + }, + { + "epoch": 1.6750494179661761, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1146576404571533, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7066159844398499, + "num_tokens": 394419230.0, + "step": 15253 + }, + { + "epoch": 1.6751592356687897, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 1.903459072113037, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6957534551620483, + "num_tokens": 394458588.0, + "step": 15254 + }, + { + "epoch": 1.6752690533714034, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.2494702339172363, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7027034759521484, + "num_tokens": 394486642.0, + "step": 15255 + }, + { + "epoch": 1.6753788710740172, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.244493007659912, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7139600515365601, + "num_tokens": 394513361.0, + "step": 15256 + }, + { + "epoch": 1.675488688776631, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.15307879447937, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7187025547027588, + "num_tokens": 394539867.0, + "step": 15257 + }, + { + "epoch": 1.6755985064792445, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.0231704711914062, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7136774063110352, + "num_tokens": 394571484.0, + "step": 15258 + }, + { + "epoch": 1.675708324181858, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.4108803272247314, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7508674263954163, + "num_tokens": 394592336.0, + "step": 15259 + }, + { + "epoch": 1.6758181418844718, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.297877073287964, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7137364745140076, + "num_tokens": 394617743.0, + "step": 15260 + }, + { + "epoch": 1.6759279595870855, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.288255453109741, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7285549640655518, + "num_tokens": 394642515.0, + "step": 15261 + }, + { + "epoch": 1.676037777289699, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.335664749145508, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7131232023239136, + "num_tokens": 394665608.0, + "step": 15262 + }, + { + "epoch": 1.6761475949923128, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3742640018463135, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7252702116966248, + "num_tokens": 394690613.0, + "step": 15263 + }, + { + "epoch": 1.6762574126949263, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1599597930908203, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6977168321609497, + "num_tokens": 394720380.0, + "step": 15264 + }, + { + "epoch": 1.67636723039754, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.1387643814086914, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7118966579437256, + "num_tokens": 394750186.0, + "step": 15265 + }, + { + "epoch": 1.6764770481001539, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.11785888671875, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7176148295402527, + "num_tokens": 394780990.0, + "step": 15266 + }, + { + "epoch": 1.6765868658027674, + "ewc_loss": 0.00012874603271484375, + "grad_norm": 2.3998615741729736, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7420752048492432, + "num_tokens": 394802607.0, + "step": 15267 + }, + { + "epoch": 1.676696683505381, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.5037519931793213, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7266136407852173, + "num_tokens": 394825914.0, + "step": 15268 + }, + { + "epoch": 1.6768065012079947, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.0368759632110596, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7271050214767456, + "num_tokens": 394855125.0, + "step": 15269 + }, + { + "epoch": 1.6769163189106084, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2403135299682617, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6996419429779053, + "num_tokens": 394883159.0, + "step": 15270 + }, + { + "epoch": 1.6770261366132222, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2630951404571533, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.702692985534668, + "num_tokens": 394909074.0, + "step": 15271 + }, + { + "epoch": 1.6771359543158357, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.316876173019409, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7179861664772034, + "num_tokens": 394932896.0, + "step": 15272 + }, + { + "epoch": 1.6772457720184493, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.422971725463867, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7074382305145264, + "num_tokens": 394957311.0, + "step": 15273 + }, + { + "epoch": 1.677355589721063, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3021485805511475, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.703903317451477, + "num_tokens": 394981625.0, + "step": 15274 + }, + { + "epoch": 1.6774654074236768, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1080591678619385, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7152059078216553, + "num_tokens": 395011025.0, + "step": 15275 + }, + { + "epoch": 1.6775752251262903, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.5910394191741943, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7166900634765625, + "num_tokens": 395031310.0, + "step": 15276 + }, + { + "epoch": 1.677685042828904, + "ewc_loss": 0.00012969970703125, + "grad_norm": 1.8990980386734009, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7067864537239075, + "num_tokens": 395064109.0, + "step": 15277 + }, + { + "epoch": 1.6777948605315176, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.12353777885437, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.708419680595398, + "num_tokens": 395092223.0, + "step": 15278 + }, + { + "epoch": 1.6779046782341314, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.080019474029541, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7178281545639038, + "num_tokens": 395121767.0, + "step": 15279 + }, + { + "epoch": 1.6780144959367451, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.305194854736328, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7389307022094727, + "num_tokens": 395145604.0, + "step": 15280 + }, + { + "epoch": 1.6781243136393587, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1512391567230225, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7186042070388794, + "num_tokens": 395172840.0, + "step": 15281 + }, + { + "epoch": 1.6782341313419722, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.581264019012451, + "learning_rate": 1e-06, + "loss": 0.7339, + "mean_token_accuracy": 0.7731090784072876, + "num_tokens": 395192119.0, + "step": 15282 + }, + { + "epoch": 1.678343949044586, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1477367877960205, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7048331499099731, + "num_tokens": 395220787.0, + "step": 15283 + }, + { + "epoch": 1.6784537667471997, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2311947345733643, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.698577344417572, + "num_tokens": 395246050.0, + "step": 15284 + }, + { + "epoch": 1.6785635844498135, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.4942262172698975, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7269387245178223, + "num_tokens": 395267040.0, + "step": 15285 + }, + { + "epoch": 1.678673402152427, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.220193386077881, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7130163908004761, + "num_tokens": 395293994.0, + "step": 15286 + }, + { + "epoch": 1.6787832198550405, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3589587211608887, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7227056622505188, + "num_tokens": 395319169.0, + "step": 15287 + }, + { + "epoch": 1.6788930375576543, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.5525920391082764, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7147355675697327, + "num_tokens": 395341689.0, + "step": 15288 + }, + { + "epoch": 1.679002855260268, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1474061012268066, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7016211748123169, + "num_tokens": 395369341.0, + "step": 15289 + }, + { + "epoch": 1.6791126729628816, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3631415367126465, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7186634540557861, + "num_tokens": 395391825.0, + "step": 15290 + }, + { + "epoch": 1.6792224906654951, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3521921634674072, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7160927057266235, + "num_tokens": 395416109.0, + "step": 15291 + }, + { + "epoch": 1.6793323083681089, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3789947032928467, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7023828029632568, + "num_tokens": 395440664.0, + "step": 15292 + }, + { + "epoch": 1.6794421260707226, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.486912250518799, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7431098222732544, + "num_tokens": 395460977.0, + "step": 15293 + }, + { + "epoch": 1.6795519437733364, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.4269776344299316, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.694623589515686, + "num_tokens": 395486277.0, + "step": 15294 + }, + { + "epoch": 1.67966176147595, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2222321033477783, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7220840454101562, + "num_tokens": 395511382.0, + "step": 15295 + }, + { + "epoch": 1.6797715791785635, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.195258855819702, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7396659851074219, + "num_tokens": 395535975.0, + "step": 15296 + }, + { + "epoch": 1.6798813968811772, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.209545612335205, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7096808552742004, + "num_tokens": 395561926.0, + "step": 15297 + }, + { + "epoch": 1.679991214583791, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.115488290786743, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.695250928401947, + "num_tokens": 395589193.0, + "step": 15298 + }, + { + "epoch": 1.6801010322864047, + "ewc_loss": 0.00012969970703125, + "grad_norm": 1.9555542469024658, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7104782462120056, + "num_tokens": 395622093.0, + "step": 15299 + }, + { + "epoch": 1.6802108499890183, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1072113513946533, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7378586530685425, + "num_tokens": 395649011.0, + "step": 15300 + }, + { + "epoch": 1.6803206676916318, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3097805976867676, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7277190685272217, + "num_tokens": 395673215.0, + "step": 15301 + }, + { + "epoch": 1.6804304853942456, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2549631595611572, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7054295539855957, + "num_tokens": 395699330.0, + "step": 15302 + }, + { + "epoch": 1.6805403030968593, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2962520122528076, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7255492210388184, + "num_tokens": 395723199.0, + "step": 15303 + }, + { + "epoch": 1.6806501207994728, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.264601469039917, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7086615562438965, + "num_tokens": 395748689.0, + "step": 15304 + }, + { + "epoch": 1.6807599385020864, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.2197697162628174, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7404825687408447, + "num_tokens": 395775038.0, + "step": 15305 + }, + { + "epoch": 1.6808697562047001, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.041978359222412, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7488653063774109, + "num_tokens": 395803491.0, + "step": 15306 + }, + { + "epoch": 1.680979573907314, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.147063970565796, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7327350378036499, + "num_tokens": 395829483.0, + "step": 15307 + }, + { + "epoch": 1.6810893916099277, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.5471415519714355, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7180807590484619, + "num_tokens": 395849479.0, + "step": 15308 + }, + { + "epoch": 1.6811992093125412, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.206411838531494, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6930903196334839, + "num_tokens": 395877741.0, + "step": 15309 + }, + { + "epoch": 1.6813090270151547, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.3920211791992188, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.729120135307312, + "num_tokens": 395900303.0, + "step": 15310 + }, + { + "epoch": 1.6814188447177685, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.1706738471984863, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7170900702476501, + "num_tokens": 395928543.0, + "step": 15311 + }, + { + "epoch": 1.6815286624203822, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.162134885787964, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7302453517913818, + "num_tokens": 395954915.0, + "step": 15312 + }, + { + "epoch": 1.681638480122996, + "ewc_loss": 0.00012969970703125, + "grad_norm": 2.354076623916626, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7479872107505798, + "num_tokens": 395977319.0, + "step": 15313 + }, + { + "epoch": 1.6817482978256095, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.336312770843506, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7159084677696228, + "num_tokens": 396000471.0, + "step": 15314 + }, + { + "epoch": 1.681858115528223, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.002271890640259, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7125080227851868, + "num_tokens": 396033624.0, + "step": 15315 + }, + { + "epoch": 1.6819679332308368, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.2635059356689453, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.722379207611084, + "num_tokens": 396060174.0, + "step": 15316 + }, + { + "epoch": 1.6820777509334506, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.593951463699341, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7115639448165894, + "num_tokens": 396081819.0, + "step": 15317 + }, + { + "epoch": 1.6821875686360641, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.1699655055999756, + "learning_rate": 1e-06, + "loss": 1.0542, + "mean_token_accuracy": 0.6790792942047119, + "num_tokens": 396112089.0, + "step": 15318 + }, + { + "epoch": 1.6822973863386776, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.429063320159912, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7283749580383301, + "num_tokens": 396134208.0, + "step": 15319 + }, + { + "epoch": 1.6824072040412914, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.170731782913208, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.6857818365097046, + "num_tokens": 396165284.0, + "step": 15320 + }, + { + "epoch": 1.6825170217439052, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.578552484512329, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7211288213729858, + "num_tokens": 396185911.0, + "step": 15321 + }, + { + "epoch": 1.682626839446519, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.369216203689575, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7196660041809082, + "num_tokens": 396208581.0, + "step": 15322 + }, + { + "epoch": 1.6827366571491325, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.0562479496002197, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7386206388473511, + "num_tokens": 396235996.0, + "step": 15323 + }, + { + "epoch": 1.682846474851746, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.3486223220825195, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7076436281204224, + "num_tokens": 396259865.0, + "step": 15324 + }, + { + "epoch": 1.6829562925543597, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.142413854598999, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7119715213775635, + "num_tokens": 396288236.0, + "step": 15325 + }, + { + "epoch": 1.6830661102569735, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.2823891639709473, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7276591658592224, + "num_tokens": 396310576.0, + "step": 15326 + }, + { + "epoch": 1.683175927959587, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.330045700073242, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.717467188835144, + "num_tokens": 396334790.0, + "step": 15327 + }, + { + "epoch": 1.6832857456622008, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.15170955657959, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7132240533828735, + "num_tokens": 396362040.0, + "step": 15328 + }, + { + "epoch": 1.6833955633648143, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.4478116035461426, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6930970549583435, + "num_tokens": 396385734.0, + "step": 15329 + }, + { + "epoch": 1.683505381067428, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.158705234527588, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7242192029953003, + "num_tokens": 396410754.0, + "step": 15330 + }, + { + "epoch": 1.6836151987700418, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.042595386505127, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.6980711221694946, + "num_tokens": 396440831.0, + "step": 15331 + }, + { + "epoch": 1.6837250164726554, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.293771743774414, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6993764638900757, + "num_tokens": 396466952.0, + "step": 15332 + }, + { + "epoch": 1.683834834175269, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.1387853622436523, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7244893908500671, + "num_tokens": 396496122.0, + "step": 15333 + }, + { + "epoch": 1.6839446518778827, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.3474314212799072, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7181451320648193, + "num_tokens": 396518979.0, + "step": 15334 + }, + { + "epoch": 1.6840544695804964, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.105893135070801, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7133474946022034, + "num_tokens": 396549076.0, + "step": 15335 + }, + { + "epoch": 1.6841642872831102, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.0343592166900635, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7097997665405273, + "num_tokens": 396579410.0, + "step": 15336 + }, + { + "epoch": 1.6842741049857237, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.216926336288452, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7234587669372559, + "num_tokens": 396605417.0, + "step": 15337 + }, + { + "epoch": 1.6843839226883373, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.309175729751587, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7226693034172058, + "num_tokens": 396629173.0, + "step": 15338 + }, + { + "epoch": 1.684493740390951, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.1636385917663574, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7131397724151611, + "num_tokens": 396657129.0, + "step": 15339 + }, + { + "epoch": 1.6846035580935648, + "ewc_loss": 0.00013065338134765625, + "grad_norm": 2.47886323928833, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7367929220199585, + "num_tokens": 396678250.0, + "step": 15340 + }, + { + "epoch": 1.6847133757961783, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.241882085800171, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7390720844268799, + "num_tokens": 396702668.0, + "step": 15341 + }, + { + "epoch": 1.684823193498792, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3184897899627686, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7259645462036133, + "num_tokens": 396727824.0, + "step": 15342 + }, + { + "epoch": 1.6849330112014056, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1478774547576904, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7182208895683289, + "num_tokens": 396756753.0, + "step": 15343 + }, + { + "epoch": 1.6850428289040194, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3910179138183594, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7200738787651062, + "num_tokens": 396780860.0, + "step": 15344 + }, + { + "epoch": 1.685152646606633, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.4488790035247803, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.704370379447937, + "num_tokens": 396804809.0, + "step": 15345 + }, + { + "epoch": 1.6852624643092466, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.472641944885254, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7395204901695251, + "num_tokens": 396827140.0, + "step": 15346 + }, + { + "epoch": 1.6853722820118602, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.350175380706787, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7146828174591064, + "num_tokens": 396851950.0, + "step": 15347 + }, + { + "epoch": 1.685482099714474, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.211529016494751, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7325633764266968, + "num_tokens": 396876087.0, + "step": 15348 + }, + { + "epoch": 1.6855919174170877, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0239639282226562, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7129887342453003, + "num_tokens": 396905161.0, + "step": 15349 + }, + { + "epoch": 1.6857017351197014, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3900184631347656, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7201518416404724, + "num_tokens": 396928353.0, + "step": 15350 + }, + { + "epoch": 1.685811552822315, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2865421772003174, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7240312099456787, + "num_tokens": 396952381.0, + "step": 15351 + }, + { + "epoch": 1.6859213705249285, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1707189083099365, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6884089708328247, + "num_tokens": 396981463.0, + "step": 15352 + }, + { + "epoch": 1.6860311882275423, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.253722667694092, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6905953884124756, + "num_tokens": 397008751.0, + "step": 15353 + }, + { + "epoch": 1.686141005930156, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.517313241958618, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7210078239440918, + "num_tokens": 397030367.0, + "step": 15354 + }, + { + "epoch": 1.6862508236327696, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.8874374628067017, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7150553464889526, + "num_tokens": 397063919.0, + "step": 15355 + }, + { + "epoch": 1.686360641335383, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.185250759124756, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7231177091598511, + "num_tokens": 397090232.0, + "step": 15356 + }, + { + "epoch": 1.6864704590379969, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3853983879089355, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7315431833267212, + "num_tokens": 397111520.0, + "step": 15357 + }, + { + "epoch": 1.6865802767406106, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.9341485500335693, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.7006576061248779, + "num_tokens": 397144790.0, + "step": 15358 + }, + { + "epoch": 1.6866900944432244, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.149104356765747, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7185661792755127, + "num_tokens": 397171378.0, + "step": 15359 + }, + { + "epoch": 1.686799912145838, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.412081003189087, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7249864339828491, + "num_tokens": 397192336.0, + "step": 15360 + }, + { + "epoch": 1.6869097298484514, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2590179443359375, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7389619946479797, + "num_tokens": 397217011.0, + "step": 15361 + }, + { + "epoch": 1.6870195475510652, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5869033336639404, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7076613903045654, + "num_tokens": 397236937.0, + "step": 15362 + }, + { + "epoch": 1.687129365253679, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2959275245666504, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7127774357795715, + "num_tokens": 397262375.0, + "step": 15363 + }, + { + "epoch": 1.6872391829562927, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.236699104309082, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7276579737663269, + "num_tokens": 397287748.0, + "step": 15364 + }, + { + "epoch": 1.6873490006589063, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0791513919830322, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7002177238464355, + "num_tokens": 397318339.0, + "step": 15365 + }, + { + "epoch": 1.6874588183615198, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.582331418991089, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7392880320549011, + "num_tokens": 397339436.0, + "step": 15366 + }, + { + "epoch": 1.6875686360641335, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.221003770828247, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7323433756828308, + "num_tokens": 397364289.0, + "step": 15367 + }, + { + "epoch": 1.6876784537667473, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.212634801864624, + "learning_rate": 1e-06, + "loss": 1.0247, + "mean_token_accuracy": 0.6909303665161133, + "num_tokens": 397390891.0, + "step": 15368 + }, + { + "epoch": 1.6877882714693608, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.286323308944702, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.726475715637207, + "num_tokens": 397415729.0, + "step": 15369 + }, + { + "epoch": 1.6878980891719744, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0801403522491455, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7161378860473633, + "num_tokens": 397444613.0, + "step": 15370 + }, + { + "epoch": 1.6880079068745881, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.251101493835449, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7230832576751709, + "num_tokens": 397470442.0, + "step": 15371 + }, + { + "epoch": 1.6881177245772019, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3619773387908936, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7337301969528198, + "num_tokens": 397493162.0, + "step": 15372 + }, + { + "epoch": 1.6882275422798156, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.8910590410232544, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.7000632286071777, + "num_tokens": 397527873.0, + "step": 15373 + }, + { + "epoch": 1.6883373599824292, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3160276412963867, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7359663248062134, + "num_tokens": 397551717.0, + "step": 15374 + }, + { + "epoch": 1.6884471776850427, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.288233518600464, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7229580879211426, + "num_tokens": 397576725.0, + "step": 15375 + }, + { + "epoch": 1.6885569953876565, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.4478509426116943, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7225959300994873, + "num_tokens": 397597859.0, + "step": 15376 + }, + { + "epoch": 1.6886668130902702, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3477704524993896, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7187244892120361, + "num_tokens": 397620983.0, + "step": 15377 + }, + { + "epoch": 1.6887766307928838, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.9538687467575073, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.726891279220581, + "num_tokens": 397654202.0, + "step": 15378 + }, + { + "epoch": 1.6888864484954975, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0520238876342773, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7329574227333069, + "num_tokens": 397684205.0, + "step": 15379 + }, + { + "epoch": 1.688996266198111, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3455169200897217, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7218964099884033, + "num_tokens": 397709024.0, + "step": 15380 + }, + { + "epoch": 1.6891060839007248, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3049516677856445, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.73462975025177, + "num_tokens": 397734698.0, + "step": 15381 + }, + { + "epoch": 1.6892159016033386, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.390059471130371, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7240598201751709, + "num_tokens": 397757587.0, + "step": 15382 + }, + { + "epoch": 1.689325719305952, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.619744300842285, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7351039052009583, + "num_tokens": 397776130.0, + "step": 15383 + }, + { + "epoch": 1.6894355370085656, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2229716777801514, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.722412109375, + "num_tokens": 397801847.0, + "step": 15384 + }, + { + "epoch": 1.6895453547111794, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0515594482421875, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7302958965301514, + "num_tokens": 397830107.0, + "step": 15385 + }, + { + "epoch": 1.6896551724137931, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.195685863494873, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.706955075263977, + "num_tokens": 397856490.0, + "step": 15386 + }, + { + "epoch": 1.689764990116407, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0447678565979004, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7107440233230591, + "num_tokens": 397886156.0, + "step": 15387 + }, + { + "epoch": 1.6898748078190204, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2986764907836914, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.73589026927948, + "num_tokens": 397910565.0, + "step": 15388 + }, + { + "epoch": 1.689984625521634, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.6154861450195312, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7412124872207642, + "num_tokens": 397931254.0, + "step": 15389 + }, + { + "epoch": 1.6900944432242477, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1031153202056885, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7230170369148254, + "num_tokens": 397960044.0, + "step": 15390 + }, + { + "epoch": 1.6902042609268615, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1411001682281494, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7064939141273499, + "num_tokens": 397988078.0, + "step": 15391 + }, + { + "epoch": 1.690314078629475, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.299765110015869, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.701971709728241, + "num_tokens": 398014185.0, + "step": 15392 + }, + { + "epoch": 1.6904238963320888, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.117612361907959, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7197448015213013, + "num_tokens": 398041080.0, + "step": 15393 + }, + { + "epoch": 1.6905337140347023, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.33544921875, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.725228488445282, + "num_tokens": 398066396.0, + "step": 15394 + }, + { + "epoch": 1.690643531737316, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3613624572753906, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7215089797973633, + "num_tokens": 398091231.0, + "step": 15395 + }, + { + "epoch": 1.6907533494399298, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.361171245574951, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7204997539520264, + "num_tokens": 398113988.0, + "step": 15396 + }, + { + "epoch": 1.6908631671425434, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.172914743423462, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7048412561416626, + "num_tokens": 398139097.0, + "step": 15397 + }, + { + "epoch": 1.690972984845157, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.237105131149292, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7327150702476501, + "num_tokens": 398166310.0, + "step": 15398 + }, + { + "epoch": 1.6910828025477707, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.373345375061035, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7300961017608643, + "num_tokens": 398188479.0, + "step": 15399 + }, + { + "epoch": 1.6911926202503844, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2148866653442383, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.72221839427948, + "num_tokens": 398214429.0, + "step": 15400 + }, + { + "epoch": 1.6913024379529982, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.9865798950195312, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7116169929504395, + "num_tokens": 398246406.0, + "step": 15401 + }, + { + "epoch": 1.6914122556556117, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1528079509735107, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7166156768798828, + "num_tokens": 398273894.0, + "step": 15402 + }, + { + "epoch": 1.6915220733582252, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.491164445877075, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7262234091758728, + "num_tokens": 398295936.0, + "step": 15403 + }, + { + "epoch": 1.691631891060839, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2052090167999268, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7128305435180664, + "num_tokens": 398322470.0, + "step": 15404 + }, + { + "epoch": 1.6917417087634528, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3253254890441895, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7027745246887207, + "num_tokens": 398347501.0, + "step": 15405 + }, + { + "epoch": 1.6918515264660663, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1076200008392334, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7167813181877136, + "num_tokens": 398377375.0, + "step": 15406 + }, + { + "epoch": 1.6919613441686798, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5600690841674805, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7278382778167725, + "num_tokens": 398397998.0, + "step": 15407 + }, + { + "epoch": 1.6920711618712936, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.8713494539260864, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7114955186843872, + "num_tokens": 398433069.0, + "step": 15408 + }, + { + "epoch": 1.6921809795739073, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.4440741539001465, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7300158739089966, + "num_tokens": 398455663.0, + "step": 15409 + }, + { + "epoch": 1.692290797276521, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.35953950881958, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7190092206001282, + "num_tokens": 398478737.0, + "step": 15410 + }, + { + "epoch": 1.6924006149791346, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1779651641845703, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7286527156829834, + "num_tokens": 398506106.0, + "step": 15411 + }, + { + "epoch": 1.6925104326817482, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.437067985534668, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7495184540748596, + "num_tokens": 398526706.0, + "step": 15412 + }, + { + "epoch": 1.692620250384362, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1055386066436768, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.692180871963501, + "num_tokens": 398558623.0, + "step": 15413 + }, + { + "epoch": 1.6927300680869757, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5380592346191406, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7341345548629761, + "num_tokens": 398580334.0, + "step": 15414 + }, + { + "epoch": 1.6928398857895894, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.548903465270996, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7512327432632446, + "num_tokens": 398601379.0, + "step": 15415 + }, + { + "epoch": 1.692949703492203, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.163665294647217, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7400343418121338, + "num_tokens": 398629504.0, + "step": 15416 + }, + { + "epoch": 1.6930595211948165, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3550639152526855, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6929255127906799, + "num_tokens": 398655967.0, + "step": 15417 + }, + { + "epoch": 1.6931693388974303, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.4491195678710938, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7145965099334717, + "num_tokens": 398677510.0, + "step": 15418 + }, + { + "epoch": 1.693279156600044, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0740854740142822, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7137073278427124, + "num_tokens": 398705020.0, + "step": 15419 + }, + { + "epoch": 1.6933889743026576, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.423515796661377, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7197235822677612, + "num_tokens": 398730620.0, + "step": 15420 + }, + { + "epoch": 1.693498792005271, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3506641387939453, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7442482709884644, + "num_tokens": 398752590.0, + "step": 15421 + }, + { + "epoch": 1.6936086097078848, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2520601749420166, + "learning_rate": 1e-06, + "loss": 0.8019, + "mean_token_accuracy": 0.7465435266494751, + "num_tokens": 398775873.0, + "step": 15422 + }, + { + "epoch": 1.6937184274104986, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.368899345397949, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7295781970024109, + "num_tokens": 398798566.0, + "step": 15423 + }, + { + "epoch": 1.6938282451131124, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.399319648742676, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7149785757064819, + "num_tokens": 398822146.0, + "step": 15424 + }, + { + "epoch": 1.693938062815726, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.113116979598999, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.6968270540237427, + "num_tokens": 398851007.0, + "step": 15425 + }, + { + "epoch": 1.6940478805183394, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3391566276550293, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7283644080162048, + "num_tokens": 398874032.0, + "step": 15426 + }, + { + "epoch": 1.6941576982209532, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.9507710933685303, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7136754393577576, + "num_tokens": 398907392.0, + "step": 15427 + }, + { + "epoch": 1.694267515923567, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2836952209472656, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.7030518054962158, + "num_tokens": 398932488.0, + "step": 15428 + }, + { + "epoch": 1.6943773336261807, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2984256744384766, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7404106855392456, + "num_tokens": 398955318.0, + "step": 15429 + }, + { + "epoch": 1.6944871513287942, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.7497284412384033, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7422159910202026, + "num_tokens": 398973510.0, + "step": 15430 + }, + { + "epoch": 1.6945969690314078, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2991247177124023, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.727724552154541, + "num_tokens": 398998123.0, + "step": 15431 + }, + { + "epoch": 1.6947067867340215, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.372966766357422, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7286766767501831, + "num_tokens": 399021861.0, + "step": 15432 + }, + { + "epoch": 1.6948166044366353, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.4960460662841797, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7371366620063782, + "num_tokens": 399042557.0, + "step": 15433 + }, + { + "epoch": 1.6949264221392488, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.084780216217041, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7285778522491455, + "num_tokens": 399072586.0, + "step": 15434 + }, + { + "epoch": 1.6950362398418624, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5581448078155518, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6989163160324097, + "num_tokens": 399099061.0, + "step": 15435 + }, + { + "epoch": 1.6951460575444761, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5818307399749756, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7224721312522888, + "num_tokens": 399117783.0, + "step": 15436 + }, + { + "epoch": 1.6952558752470899, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3754067420959473, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7214310169219971, + "num_tokens": 399141358.0, + "step": 15437 + }, + { + "epoch": 1.6953656929497036, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1542704105377197, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6969677209854126, + "num_tokens": 399169370.0, + "step": 15438 + }, + { + "epoch": 1.6954755106523172, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.05551815032959, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.726457417011261, + "num_tokens": 399197959.0, + "step": 15439 + }, + { + "epoch": 1.6955853283549307, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.37060284614563, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7201581001281738, + "num_tokens": 399220990.0, + "step": 15440 + }, + { + "epoch": 1.6956951460575445, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2823739051818848, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7087670564651489, + "num_tokens": 399246665.0, + "step": 15441 + }, + { + "epoch": 1.6958049637601582, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.0308756828308105, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6970226168632507, + "num_tokens": 399276337.0, + "step": 15442 + }, + { + "epoch": 1.6959147814627717, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.349149227142334, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7274258136749268, + "num_tokens": 399298346.0, + "step": 15443 + }, + { + "epoch": 1.6960245991653855, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.258345365524292, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.709295392036438, + "num_tokens": 399323963.0, + "step": 15444 + }, + { + "epoch": 1.696134416867999, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.5734221935272217, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7249796390533447, + "num_tokens": 399344041.0, + "step": 15445 + }, + { + "epoch": 1.6962442345706128, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.51904296875, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7176815867424011, + "num_tokens": 399365154.0, + "step": 15446 + }, + { + "epoch": 1.6963540522732266, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1794016361236572, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7283491492271423, + "num_tokens": 399390134.0, + "step": 15447 + }, + { + "epoch": 1.69646386997584, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.2547805309295654, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7230139970779419, + "num_tokens": 399414294.0, + "step": 15448 + }, + { + "epoch": 1.6965736876784536, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.29994535446167, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6939815878868103, + "num_tokens": 399442632.0, + "step": 15449 + }, + { + "epoch": 1.6966835053810674, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1642956733703613, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6866970658302307, + "num_tokens": 399471270.0, + "step": 15450 + }, + { + "epoch": 1.6967933230836811, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3159477710723877, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7231464982032776, + "num_tokens": 399495752.0, + "step": 15451 + }, + { + "epoch": 1.696903140786295, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1467325687408447, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.70895916223526, + "num_tokens": 399523581.0, + "step": 15452 + }, + { + "epoch": 1.6970129584889084, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.178523540496826, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7089902758598328, + "num_tokens": 399550092.0, + "step": 15453 + }, + { + "epoch": 1.697122776191522, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.9536008834838867, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.736202597618103, + "num_tokens": 399579278.0, + "step": 15454 + }, + { + "epoch": 1.6972325938941357, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 1.872005581855774, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7225431799888611, + "num_tokens": 399610000.0, + "step": 15455 + }, + { + "epoch": 1.6973424115967495, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1590778827667236, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.735527753829956, + "num_tokens": 399635600.0, + "step": 15456 + }, + { + "epoch": 1.697452229299363, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.328542947769165, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7191320657730103, + "num_tokens": 399660664.0, + "step": 15457 + }, + { + "epoch": 1.6975620470019768, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.3226094245910645, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7031106352806091, + "num_tokens": 399687216.0, + "step": 15458 + }, + { + "epoch": 1.6976718647045903, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.270362138748169, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7004932165145874, + "num_tokens": 399711844.0, + "step": 15459 + }, + { + "epoch": 1.697781682407204, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.007340431213379, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6937754154205322, + "num_tokens": 399743060.0, + "step": 15460 + }, + { + "epoch": 1.6978915001098178, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.172975778579712, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7068642973899841, + "num_tokens": 399769595.0, + "step": 15461 + }, + { + "epoch": 1.6980013178124314, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.33520770072937, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7172079086303711, + "num_tokens": 399794162.0, + "step": 15462 + }, + { + "epoch": 1.698111135515045, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.312051296234131, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.6974993944168091, + "num_tokens": 399820012.0, + "step": 15463 + }, + { + "epoch": 1.6982209532176586, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1757969856262207, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7109267711639404, + "num_tokens": 399846566.0, + "step": 15464 + }, + { + "epoch": 1.6983307709202724, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.1596996784210205, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7438193559646606, + "num_tokens": 399872193.0, + "step": 15465 + }, + { + "epoch": 1.6984405886228862, + "ewc_loss": 0.0001316070556640625, + "grad_norm": 2.178518056869507, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7142230272293091, + "num_tokens": 399897730.0, + "step": 15466 + }, + { + "epoch": 1.6985504063254997, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.174757719039917, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7075371146202087, + "num_tokens": 399926421.0, + "step": 15467 + }, + { + "epoch": 1.6986602240281132, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.461068868637085, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7201828956604004, + "num_tokens": 399946819.0, + "step": 15468 + }, + { + "epoch": 1.698770041730727, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.379995822906494, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7271401882171631, + "num_tokens": 399970656.0, + "step": 15469 + }, + { + "epoch": 1.6988798594333407, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0877761840820312, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7206903696060181, + "num_tokens": 399998962.0, + "step": 15470 + }, + { + "epoch": 1.6989896771359543, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.127577304840088, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7202610969543457, + "num_tokens": 400029421.0, + "step": 15471 + }, + { + "epoch": 1.6990994948385678, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3426976203918457, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7072518467903137, + "num_tokens": 400053607.0, + "step": 15472 + }, + { + "epoch": 1.6992093125411816, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0709803104400635, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.732945442199707, + "num_tokens": 400080343.0, + "step": 15473 + }, + { + "epoch": 1.6993191302437953, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0744917392730713, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7075880765914917, + "num_tokens": 400109965.0, + "step": 15474 + }, + { + "epoch": 1.699428947946409, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3455278873443604, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.72509765625, + "num_tokens": 400132559.0, + "step": 15475 + }, + { + "epoch": 1.6995387656490226, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1083602905273438, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7249987721443176, + "num_tokens": 400159520.0, + "step": 15476 + }, + { + "epoch": 1.6996485833516362, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.358341932296753, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7103050947189331, + "num_tokens": 400182574.0, + "step": 15477 + }, + { + "epoch": 1.69975840105425, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1093451976776123, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7107627391815186, + "num_tokens": 400212425.0, + "step": 15478 + }, + { + "epoch": 1.6998682187568637, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.343608856201172, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7347671985626221, + "num_tokens": 400234508.0, + "step": 15479 + }, + { + "epoch": 1.6999780364594774, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1431262493133545, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7266781330108643, + "num_tokens": 400261336.0, + "step": 15480 + }, + { + "epoch": 1.700087854162091, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3495993614196777, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7178308963775635, + "num_tokens": 400284616.0, + "step": 15481 + }, + { + "epoch": 1.7001976718647045, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2282798290252686, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7097002863883972, + "num_tokens": 400311134.0, + "step": 15482 + }, + { + "epoch": 1.7003074895673183, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.396503210067749, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.727524995803833, + "num_tokens": 400333231.0, + "step": 15483 + }, + { + "epoch": 1.700417307269932, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9895315170288086, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6936898231506348, + "num_tokens": 400366946.0, + "step": 15484 + }, + { + "epoch": 1.7005271249725455, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2415199279785156, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7248502373695374, + "num_tokens": 400391020.0, + "step": 15485 + }, + { + "epoch": 1.700636942675159, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0781171321868896, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7137153148651123, + "num_tokens": 400420854.0, + "step": 15486 + }, + { + "epoch": 1.7007467603777728, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2667901515960693, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7287701964378357, + "num_tokens": 400445400.0, + "step": 15487 + }, + { + "epoch": 1.7008565780803866, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2587194442749023, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7033723592758179, + "num_tokens": 400472258.0, + "step": 15488 + }, + { + "epoch": 1.7009663957830004, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.29166579246521, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7274861931800842, + "num_tokens": 400495777.0, + "step": 15489 + }, + { + "epoch": 1.7010762134856139, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.312751293182373, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7080583572387695, + "num_tokens": 400523789.0, + "step": 15490 + }, + { + "epoch": 1.7011860311882274, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.384139060974121, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7287497520446777, + "num_tokens": 400548564.0, + "step": 15491 + }, + { + "epoch": 1.7012958488908412, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2727925777435303, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7101913094520569, + "num_tokens": 400572430.0, + "step": 15492 + }, + { + "epoch": 1.701405666593455, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1057329177856445, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7196608185768127, + "num_tokens": 400600268.0, + "step": 15493 + }, + { + "epoch": 1.7015154842960687, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.376448392868042, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7388981580734253, + "num_tokens": 400622347.0, + "step": 15494 + }, + { + "epoch": 1.7016253019986822, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.7698967456817627, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7337230443954468, + "num_tokens": 400640295.0, + "step": 15495 + }, + { + "epoch": 1.7017351197012958, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.722398281097412, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7185896635055542, + "num_tokens": 400657890.0, + "step": 15496 + }, + { + "epoch": 1.7018449374039095, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2621893882751465, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7202809453010559, + "num_tokens": 400682174.0, + "step": 15497 + }, + { + "epoch": 1.7019547551065233, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.26537823677063, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7257825136184692, + "num_tokens": 400707258.0, + "step": 15498 + }, + { + "epoch": 1.7020645728091368, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1480891704559326, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6942316889762878, + "num_tokens": 400736267.0, + "step": 15499 + }, + { + "epoch": 1.7021743905117503, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5172159671783447, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7163865566253662, + "num_tokens": 400758113.0, + "step": 15500 + }, + { + "epoch": 1.702284208214364, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5776169300079346, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7369397878646851, + "num_tokens": 400779398.0, + "step": 15501 + }, + { + "epoch": 1.7023940259169779, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2923030853271484, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7287742495536804, + "num_tokens": 400803902.0, + "step": 15502 + }, + { + "epoch": 1.7025038436195916, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2161898612976074, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7183923721313477, + "num_tokens": 400828366.0, + "step": 15503 + }, + { + "epoch": 1.7026136613222052, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.054387092590332, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7084330320358276, + "num_tokens": 400857298.0, + "step": 15504 + }, + { + "epoch": 1.7027234790248187, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.6081008911132812, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.726285994052887, + "num_tokens": 400878272.0, + "step": 15505 + }, + { + "epoch": 1.7028332967274324, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.289419174194336, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.71708083152771, + "num_tokens": 400904791.0, + "step": 15506 + }, + { + "epoch": 1.7029431144300462, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0268633365631104, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7422593832015991, + "num_tokens": 400931359.0, + "step": 15507 + }, + { + "epoch": 1.7030529321326597, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.8883970975875854, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7058416604995728, + "num_tokens": 400968802.0, + "step": 15508 + }, + { + "epoch": 1.7031627498352735, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.363348960876465, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7134007215499878, + "num_tokens": 400993412.0, + "step": 15509 + }, + { + "epoch": 1.703272567537887, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.43322491645813, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7188236713409424, + "num_tokens": 401016039.0, + "step": 15510 + }, + { + "epoch": 1.7033823852405008, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4409451484680176, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7304115891456604, + "num_tokens": 401037545.0, + "step": 15511 + }, + { + "epoch": 1.7034922029431145, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.216012954711914, + "learning_rate": 1e-06, + "loss": 1.0488, + "mean_token_accuracy": 0.6812313795089722, + "num_tokens": 401064993.0, + "step": 15512 + }, + { + "epoch": 1.703602020645728, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1678452491760254, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7279185056686401, + "num_tokens": 401091084.0, + "step": 15513 + }, + { + "epoch": 1.7037118383483416, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9835195541381836, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6949523687362671, + "num_tokens": 401122728.0, + "step": 15514 + }, + { + "epoch": 1.7038216560509554, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5956456661224365, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7159841060638428, + "num_tokens": 401142675.0, + "step": 15515 + }, + { + "epoch": 1.7039314737535691, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.118940591812134, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.705022931098938, + "num_tokens": 401170929.0, + "step": 15516 + }, + { + "epoch": 1.7040412914561829, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1856772899627686, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.698070764541626, + "num_tokens": 401198278.0, + "step": 15517 + }, + { + "epoch": 1.7041511091587964, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1419832706451416, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7201722264289856, + "num_tokens": 401225270.0, + "step": 15518 + }, + { + "epoch": 1.70426092686141, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1820027828216553, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7375195622444153, + "num_tokens": 401249881.0, + "step": 15519 + }, + { + "epoch": 1.7043707445640237, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3476648330688477, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.723984956741333, + "num_tokens": 401272146.0, + "step": 15520 + }, + { + "epoch": 1.7044805622666375, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.48925518989563, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7241264581680298, + "num_tokens": 401293937.0, + "step": 15521 + }, + { + "epoch": 1.704590379969251, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2649824619293213, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7353401184082031, + "num_tokens": 401317738.0, + "step": 15522 + }, + { + "epoch": 1.7047001976718648, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1359622478485107, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.68316650390625, + "num_tokens": 401348387.0, + "step": 15523 + }, + { + "epoch": 1.7048100153744783, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.014935255050659, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7190597057342529, + "num_tokens": 401381340.0, + "step": 15524 + }, + { + "epoch": 1.704919833077092, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.468320608139038, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7235721349716187, + "num_tokens": 401402595.0, + "step": 15525 + }, + { + "epoch": 1.7050296507797058, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.340263605117798, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7306627035140991, + "num_tokens": 401425200.0, + "step": 15526 + }, + { + "epoch": 1.7051394684823193, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4059298038482666, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7066603899002075, + "num_tokens": 401448836.0, + "step": 15527 + }, + { + "epoch": 1.7052492861849329, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.460909843444824, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7243014574050903, + "num_tokens": 401470107.0, + "step": 15528 + }, + { + "epoch": 1.7053591038875466, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1654140949249268, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7052308917045593, + "num_tokens": 401497024.0, + "step": 15529 + }, + { + "epoch": 1.7054689215901604, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.016805648803711, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6817587614059448, + "num_tokens": 401529644.0, + "step": 15530 + }, + { + "epoch": 1.7055787392927741, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.26275634765625, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7154465317726135, + "num_tokens": 401553693.0, + "step": 15531 + }, + { + "epoch": 1.7056885569953877, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.296762228012085, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7151928544044495, + "num_tokens": 401577197.0, + "step": 15532 + }, + { + "epoch": 1.7057983746980012, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0720231533050537, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6950793862342834, + "num_tokens": 401608064.0, + "step": 15533 + }, + { + "epoch": 1.705908192400615, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5702552795410156, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7416128516197205, + "num_tokens": 401628407.0, + "step": 15534 + }, + { + "epoch": 1.7060180101032287, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.14593243598938, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6902148127555847, + "num_tokens": 401656515.0, + "step": 15535 + }, + { + "epoch": 1.7061278278058423, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2993128299713135, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7218578457832336, + "num_tokens": 401681436.0, + "step": 15536 + }, + { + "epoch": 1.7062376455084558, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2350270748138428, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7167460918426514, + "num_tokens": 401706904.0, + "step": 15537 + }, + { + "epoch": 1.7063474632110696, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.335017204284668, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.728000819683075, + "num_tokens": 401730977.0, + "step": 15538 + }, + { + "epoch": 1.7064572809136833, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0926547050476074, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7298672795295715, + "num_tokens": 401757877.0, + "step": 15539 + }, + { + "epoch": 1.706567098616297, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1717653274536133, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7254281044006348, + "num_tokens": 401784118.0, + "step": 15540 + }, + { + "epoch": 1.7066769163189106, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.250084161758423, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7535821199417114, + "num_tokens": 401809143.0, + "step": 15541 + }, + { + "epoch": 1.7067867340215241, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5135369300842285, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7152773141860962, + "num_tokens": 401830948.0, + "step": 15542 + }, + { + "epoch": 1.706896551724138, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.394742250442505, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7307329177856445, + "num_tokens": 401853041.0, + "step": 15543 + }, + { + "epoch": 1.7070063694267517, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1293447017669678, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7088080048561096, + "num_tokens": 401883021.0, + "step": 15544 + }, + { + "epoch": 1.7071161871293654, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4582884311676025, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7148606181144714, + "num_tokens": 401905419.0, + "step": 15545 + }, + { + "epoch": 1.707226004831979, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1097514629364014, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7044914364814758, + "num_tokens": 401934862.0, + "step": 15546 + }, + { + "epoch": 1.7073358225345925, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0087697505950928, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6929507851600647, + "num_tokens": 401965412.0, + "step": 15547 + }, + { + "epoch": 1.7074456402372062, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3209190368652344, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7038833498954773, + "num_tokens": 401990219.0, + "step": 15548 + }, + { + "epoch": 1.70755545793982, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.680126190185547, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7255175113677979, + "num_tokens": 402010596.0, + "step": 15549 + }, + { + "epoch": 1.7076652756424335, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3298799991607666, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7225569486618042, + "num_tokens": 402035555.0, + "step": 15550 + }, + { + "epoch": 1.707775093345047, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.085812568664551, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7055082321166992, + "num_tokens": 402064140.0, + "step": 15551 + }, + { + "epoch": 1.7078849110476608, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4616050720214844, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7210150957107544, + "num_tokens": 402085461.0, + "step": 15552 + }, + { + "epoch": 1.7079947287502746, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.125265121459961, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7249065041542053, + "num_tokens": 402113389.0, + "step": 15553 + }, + { + "epoch": 1.7081045464528883, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4653406143188477, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7200822830200195, + "num_tokens": 402134794.0, + "step": 15554 + }, + { + "epoch": 1.7082143641555019, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2017464637756348, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7027732729911804, + "num_tokens": 402162204.0, + "step": 15555 + }, + { + "epoch": 1.7083241818581154, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.6662991046905518, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.731313943862915, + "num_tokens": 402180576.0, + "step": 15556 + }, + { + "epoch": 1.7084339995607292, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3118669986724854, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.742989182472229, + "num_tokens": 402202089.0, + "step": 15557 + }, + { + "epoch": 1.708543817263343, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5445806980133057, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7424260377883911, + "num_tokens": 402221884.0, + "step": 15558 + }, + { + "epoch": 1.7086536349659565, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4286997318267822, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7179032564163208, + "num_tokens": 402243592.0, + "step": 15559 + }, + { + "epoch": 1.7087634526685702, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.276729106903076, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7031751871109009, + "num_tokens": 402268787.0, + "step": 15560 + }, + { + "epoch": 1.7088732703711838, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0596888065338135, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7105817198753357, + "num_tokens": 402299378.0, + "step": 15561 + }, + { + "epoch": 1.7089830880737975, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.135119676589966, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.727593183517456, + "num_tokens": 402324054.0, + "step": 15562 + }, + { + "epoch": 1.7090929057764113, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2739272117614746, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7261675596237183, + "num_tokens": 402346913.0, + "step": 15563 + }, + { + "epoch": 1.7092027234790248, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.247176170349121, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7078050374984741, + "num_tokens": 402373163.0, + "step": 15564 + }, + { + "epoch": 1.7093125411816383, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2567455768585205, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7424564957618713, + "num_tokens": 402396775.0, + "step": 15565 + }, + { + "epoch": 1.709422358884252, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1342687606811523, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7214182615280151, + "num_tokens": 402426353.0, + "step": 15566 + }, + { + "epoch": 1.7095321765868658, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.962255597114563, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.6964924335479736, + "num_tokens": 402457270.0, + "step": 15567 + }, + { + "epoch": 1.7096419942894796, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2060601711273193, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7149468660354614, + "num_tokens": 402483166.0, + "step": 15568 + }, + { + "epoch": 1.7097518119920931, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.12363338470459, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7303425669670105, + "num_tokens": 402510246.0, + "step": 15569 + }, + { + "epoch": 1.7098616296947067, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9859209060668945, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.728473961353302, + "num_tokens": 402540081.0, + "step": 15570 + }, + { + "epoch": 1.7099714473973204, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9849343299865723, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7028570771217346, + "num_tokens": 402570877.0, + "step": 15571 + }, + { + "epoch": 1.7100812650999342, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3394365310668945, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7293503284454346, + "num_tokens": 402594434.0, + "step": 15572 + }, + { + "epoch": 1.7101910828025477, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2813730239868164, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7273696064949036, + "num_tokens": 402618572.0, + "step": 15573 + }, + { + "epoch": 1.7103009005051615, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.6166188716888428, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7147517204284668, + "num_tokens": 402639406.0, + "step": 15574 + }, + { + "epoch": 1.710410718207775, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.678821563720703, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7121582627296448, + "num_tokens": 402658305.0, + "step": 15575 + }, + { + "epoch": 1.7105205359103888, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0237162113189697, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.6969457268714905, + "num_tokens": 402690186.0, + "step": 15576 + }, + { + "epoch": 1.7106303536130025, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.022326707839966, + "learning_rate": 1e-06, + "loss": 1.0451, + "mean_token_accuracy": 0.6864057779312134, + "num_tokens": 402722651.0, + "step": 15577 + }, + { + "epoch": 1.710740171315616, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.4396181106567383, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7391822934150696, + "num_tokens": 402746767.0, + "step": 15578 + }, + { + "epoch": 1.7108499890182296, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9918806552886963, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7255979776382446, + "num_tokens": 402777349.0, + "step": 15579 + }, + { + "epoch": 1.7109598067208434, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1274540424346924, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7142320275306702, + "num_tokens": 402806854.0, + "step": 15580 + }, + { + "epoch": 1.7110696244234571, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.090786933898926, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7054443359375, + "num_tokens": 402836077.0, + "step": 15581 + }, + { + "epoch": 1.7111794421260709, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2210538387298584, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7066009044647217, + "num_tokens": 402862507.0, + "step": 15582 + }, + { + "epoch": 1.7112892598286844, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.9162395000457764, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.720051646232605, + "num_tokens": 402895496.0, + "step": 15583 + }, + { + "epoch": 1.711399077531298, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1671392917633057, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7157686948776245, + "num_tokens": 402922000.0, + "step": 15584 + }, + { + "epoch": 1.7115088952339117, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2239255905151367, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7132277488708496, + "num_tokens": 402946708.0, + "step": 15585 + }, + { + "epoch": 1.7116187129365255, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.178563117980957, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7142335772514343, + "num_tokens": 402974797.0, + "step": 15586 + }, + { + "epoch": 1.711728530639139, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.247687339782715, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7243121862411499, + "num_tokens": 402999664.0, + "step": 15587 + }, + { + "epoch": 1.7118383483417525, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.105721950531006, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.6956905126571655, + "num_tokens": 403031461.0, + "step": 15588 + }, + { + "epoch": 1.7119481660443663, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1616148948669434, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7259653806686401, + "num_tokens": 403058704.0, + "step": 15589 + }, + { + "epoch": 1.71205798374698, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0176541805267334, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7352116703987122, + "num_tokens": 403089446.0, + "step": 15590 + }, + { + "epoch": 1.7121678014495938, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3029041290283203, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6949735879898071, + "num_tokens": 403116757.0, + "step": 15591 + }, + { + "epoch": 1.7122776191522073, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3373661041259766, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7254903316497803, + "num_tokens": 403138993.0, + "step": 15592 + }, + { + "epoch": 1.7123874368548209, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3662781715393066, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7189114093780518, + "num_tokens": 403162916.0, + "step": 15593 + }, + { + "epoch": 1.7124972545574346, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0186479091644287, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7234165072441101, + "num_tokens": 403191582.0, + "step": 15594 + }, + { + "epoch": 1.7126070722600484, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1577141284942627, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7055909633636475, + "num_tokens": 403218747.0, + "step": 15595 + }, + { + "epoch": 1.7127168899626621, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.1265199184417725, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7014829516410828, + "num_tokens": 403247716.0, + "step": 15596 + }, + { + "epoch": 1.7128267076652757, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.970312476158142, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7275664806365967, + "num_tokens": 403277059.0, + "step": 15597 + }, + { + "epoch": 1.7129365253678892, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.119894027709961, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7051900625228882, + "num_tokens": 403304477.0, + "step": 15598 + }, + { + "epoch": 1.713046343070503, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.450141191482544, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7350094318389893, + "num_tokens": 403327045.0, + "step": 15599 + }, + { + "epoch": 1.7131561607731167, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.6065633296966553, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7314993143081665, + "num_tokens": 403346736.0, + "step": 15600 + }, + { + "epoch": 1.7132659784757303, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3151493072509766, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.721418559551239, + "num_tokens": 403372578.0, + "step": 15601 + }, + { + "epoch": 1.7133757961783438, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.506592035293579, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7142544984817505, + "num_tokens": 403393365.0, + "step": 15602 + }, + { + "epoch": 1.7134856138809575, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2070553302764893, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7117919325828552, + "num_tokens": 403419727.0, + "step": 15603 + }, + { + "epoch": 1.7135954315835713, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.109107494354248, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7190098762512207, + "num_tokens": 403447141.0, + "step": 15604 + }, + { + "epoch": 1.713705249286185, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3681516647338867, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7111744284629822, + "num_tokens": 403471657.0, + "step": 15605 + }, + { + "epoch": 1.7138150669887986, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2310500144958496, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7226834297180176, + "num_tokens": 403496854.0, + "step": 15606 + }, + { + "epoch": 1.7139248846914121, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0773611068725586, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7165741920471191, + "num_tokens": 403524503.0, + "step": 15607 + }, + { + "epoch": 1.7140347023940259, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.0670900344848633, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6986078023910522, + "num_tokens": 403554962.0, + "step": 15608 + }, + { + "epoch": 1.7141445200966396, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.956796407699585, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7112306356430054, + "num_tokens": 403587714.0, + "step": 15609 + }, + { + "epoch": 1.7142543377992534, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.2275490760803223, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7003777027130127, + "num_tokens": 403616592.0, + "step": 15610 + }, + { + "epoch": 1.714364155501867, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.204827070236206, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7229742407798767, + "num_tokens": 403641132.0, + "step": 15611 + }, + { + "epoch": 1.7144739732044805, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 1.922995686531067, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7170295715332031, + "num_tokens": 403673395.0, + "step": 15612 + }, + { + "epoch": 1.7145837909070942, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.415731430053711, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7484440803527832, + "num_tokens": 403695074.0, + "step": 15613 + }, + { + "epoch": 1.714693608609708, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.3048393726348877, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7122633457183838, + "num_tokens": 403722761.0, + "step": 15614 + }, + { + "epoch": 1.7148034263123215, + "ewc_loss": 0.00013256072998046875, + "grad_norm": 2.5565755367279053, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7120766639709473, + "num_tokens": 403743307.0, + "step": 15615 + }, + { + "epoch": 1.714913244014935, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.353031635284424, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7148129343986511, + "num_tokens": 403767386.0, + "step": 15616 + }, + { + "epoch": 1.7150230617175488, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.147779941558838, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.725823700428009, + "num_tokens": 403795522.0, + "step": 15617 + }, + { + "epoch": 1.7151328794201626, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3055830001831055, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7151205539703369, + "num_tokens": 403821048.0, + "step": 15618 + }, + { + "epoch": 1.7152426971227763, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.8876020908355713, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7225850820541382, + "num_tokens": 403854081.0, + "step": 15619 + }, + { + "epoch": 1.7153525148253899, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.93244206905365, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7266759872436523, + "num_tokens": 403887105.0, + "step": 15620 + }, + { + "epoch": 1.7154623325280034, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3261053562164307, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7104780077934265, + "num_tokens": 403911293.0, + "step": 15621 + }, + { + "epoch": 1.7155721502306172, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.378793954849243, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7202428579330444, + "num_tokens": 403932423.0, + "step": 15622 + }, + { + "epoch": 1.715681967933231, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.011367082595825, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7078583240509033, + "num_tokens": 403964248.0, + "step": 15623 + }, + { + "epoch": 1.7157917856358444, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1608505249023438, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7343359589576721, + "num_tokens": 403990120.0, + "step": 15624 + }, + { + "epoch": 1.7159016033384582, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5718190670013428, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7385773062705994, + "num_tokens": 404009131.0, + "step": 15625 + }, + { + "epoch": 1.7160114210410717, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.102084159851074, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7117154598236084, + "num_tokens": 404039592.0, + "step": 15626 + }, + { + "epoch": 1.7161212387436855, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3818395137786865, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7260740995407104, + "num_tokens": 404061611.0, + "step": 15627 + }, + { + "epoch": 1.7162310564462993, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1567447185516357, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7053279280662537, + "num_tokens": 404089891.0, + "step": 15628 + }, + { + "epoch": 1.7163408741489128, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1048600673675537, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7101221084594727, + "num_tokens": 404118032.0, + "step": 15629 + }, + { + "epoch": 1.7164506918515263, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.6670892238616943, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7208348512649536, + "num_tokens": 404139583.0, + "step": 15630 + }, + { + "epoch": 1.71656050955414, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4903311729431152, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7138855457305908, + "num_tokens": 404162484.0, + "step": 15631 + }, + { + "epoch": 1.7166703272567538, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3208940029144287, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7355045080184937, + "num_tokens": 404185439.0, + "step": 15632 + }, + { + "epoch": 1.7167801449593676, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2547597885131836, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7252920269966125, + "num_tokens": 404209797.0, + "step": 15633 + }, + { + "epoch": 1.7168899626619811, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.401402235031128, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7418531179428101, + "num_tokens": 404230140.0, + "step": 15634 + }, + { + "epoch": 1.7169997803645947, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.425260543823242, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7087575793266296, + "num_tokens": 404254200.0, + "step": 15635 + }, + { + "epoch": 1.7171095980672084, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2776081562042236, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7218146324157715, + "num_tokens": 404278519.0, + "step": 15636 + }, + { + "epoch": 1.7172194157698222, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3152072429656982, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7058033347129822, + "num_tokens": 404303770.0, + "step": 15637 + }, + { + "epoch": 1.7173292334724357, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4879119396209717, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7295606136322021, + "num_tokens": 404325138.0, + "step": 15638 + }, + { + "epoch": 1.7174390511750495, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3133020401000977, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7049453258514404, + "num_tokens": 404349688.0, + "step": 15639 + }, + { + "epoch": 1.717548868877663, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9129940271377563, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7112722992897034, + "num_tokens": 404382347.0, + "step": 15640 + }, + { + "epoch": 1.7176586865802768, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.057436227798462, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7158708572387695, + "num_tokens": 404409840.0, + "step": 15641 + }, + { + "epoch": 1.7177685042828905, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.034287214279175, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6913714408874512, + "num_tokens": 404441008.0, + "step": 15642 + }, + { + "epoch": 1.717878321985504, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.060391664505005, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7079169750213623, + "num_tokens": 404468403.0, + "step": 15643 + }, + { + "epoch": 1.7179881396881176, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0951452255249023, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7193176746368408, + "num_tokens": 404497085.0, + "step": 15644 + }, + { + "epoch": 1.7180979573907313, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9668998718261719, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7034353017807007, + "num_tokens": 404529952.0, + "step": 15645 + }, + { + "epoch": 1.718207775093345, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.005686044692993, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.701815128326416, + "num_tokens": 404559855.0, + "step": 15646 + }, + { + "epoch": 1.7183175927959589, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1742732524871826, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7139010429382324, + "num_tokens": 404585447.0, + "step": 15647 + }, + { + "epoch": 1.7184274104985724, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0554285049438477, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7126880884170532, + "num_tokens": 404615258.0, + "step": 15648 + }, + { + "epoch": 1.718537228201186, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2602241039276123, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7108629941940308, + "num_tokens": 404640487.0, + "step": 15649 + }, + { + "epoch": 1.7186470459037997, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1464905738830566, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.7038666009902954, + "num_tokens": 404669984.0, + "step": 15650 + }, + { + "epoch": 1.7187568636064134, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.187345266342163, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7121973037719727, + "num_tokens": 404695795.0, + "step": 15651 + }, + { + "epoch": 1.718866681309027, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.380352258682251, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7322618961334229, + "num_tokens": 404716863.0, + "step": 15652 + }, + { + "epoch": 1.7189764990116405, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.7593419551849365, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7319215536117554, + "num_tokens": 404733635.0, + "step": 15653 + }, + { + "epoch": 1.7190863167142543, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5710699558258057, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.735653281211853, + "num_tokens": 404754534.0, + "step": 15654 + }, + { + "epoch": 1.719196134416868, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0524072647094727, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6989675164222717, + "num_tokens": 404786352.0, + "step": 15655 + }, + { + "epoch": 1.7193059521194818, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2660765647888184, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7151192426681519, + "num_tokens": 404812670.0, + "step": 15656 + }, + { + "epoch": 1.7194157698220953, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3870503902435303, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7270488739013672, + "num_tokens": 404833971.0, + "step": 15657 + }, + { + "epoch": 1.7195255875247089, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9403129816055298, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7139821648597717, + "num_tokens": 404865009.0, + "step": 15658 + }, + { + "epoch": 1.7196354052273226, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2250723838806152, + "learning_rate": 1e-06, + "loss": 1.0274, + "mean_token_accuracy": 0.6971824169158936, + "num_tokens": 404891573.0, + "step": 15659 + }, + { + "epoch": 1.7197452229299364, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.258105754852295, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.7108839750289917, + "num_tokens": 404918890.0, + "step": 15660 + }, + { + "epoch": 1.7198550406325501, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1889243125915527, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7465803027153015, + "num_tokens": 404943132.0, + "step": 15661 + }, + { + "epoch": 1.7199648583351637, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2475035190582275, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.709017276763916, + "num_tokens": 404970183.0, + "step": 15662 + }, + { + "epoch": 1.7200746760377772, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.233380079269409, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7121580243110657, + "num_tokens": 404996090.0, + "step": 15663 + }, + { + "epoch": 1.720184493740391, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.057192325592041, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7211048007011414, + "num_tokens": 405024775.0, + "step": 15664 + }, + { + "epoch": 1.7202943114430047, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9998167753219604, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7027822136878967, + "num_tokens": 405056691.0, + "step": 15665 + }, + { + "epoch": 1.7204041291456182, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3153340816497803, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7079976797103882, + "num_tokens": 405082271.0, + "step": 15666 + }, + { + "epoch": 1.7205139468482318, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5700502395629883, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7160095572471619, + "num_tokens": 405104542.0, + "step": 15667 + }, + { + "epoch": 1.7206237645508455, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.036649465560913, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.723544180393219, + "num_tokens": 405135485.0, + "step": 15668 + }, + { + "epoch": 1.7207335822534593, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.305673360824585, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7259278297424316, + "num_tokens": 405159707.0, + "step": 15669 + }, + { + "epoch": 1.720843399956073, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3422911167144775, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7196799516677856, + "num_tokens": 405185350.0, + "step": 15670 + }, + { + "epoch": 1.7209532176586866, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1709651947021484, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7315548658370972, + "num_tokens": 405214522.0, + "step": 15671 + }, + { + "epoch": 1.7210630353613001, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.608947277069092, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7214651703834534, + "num_tokens": 405236811.0, + "step": 15672 + }, + { + "epoch": 1.7211728530639139, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0325143337249756, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7445589303970337, + "num_tokens": 405265540.0, + "step": 15673 + }, + { + "epoch": 1.7212826707665276, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2100987434387207, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.706745982170105, + "num_tokens": 405295474.0, + "step": 15674 + }, + { + "epoch": 1.7213924884691414, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2402093410491943, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7231648564338684, + "num_tokens": 405321196.0, + "step": 15675 + }, + { + "epoch": 1.721502306171755, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.046670913696289, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7001823782920837, + "num_tokens": 405351944.0, + "step": 15676 + }, + { + "epoch": 1.7216121238743685, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3419432640075684, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7081383466720581, + "num_tokens": 405378336.0, + "step": 15677 + }, + { + "epoch": 1.7217219415769822, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.136080026626587, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7355064749717712, + "num_tokens": 405404115.0, + "step": 15678 + }, + { + "epoch": 1.721831759279596, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3743386268615723, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.712364673614502, + "num_tokens": 405427676.0, + "step": 15679 + }, + { + "epoch": 1.7219415769822095, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0816867351531982, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.698223352432251, + "num_tokens": 405456316.0, + "step": 15680 + }, + { + "epoch": 1.722051394684823, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.531477451324463, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7146346569061279, + "num_tokens": 405478827.0, + "step": 15681 + }, + { + "epoch": 1.7221612123874368, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2179205417633057, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.739709734916687, + "num_tokens": 405503775.0, + "step": 15682 + }, + { + "epoch": 1.7222710300900506, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.263913154602051, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7081969976425171, + "num_tokens": 405530284.0, + "step": 15683 + }, + { + "epoch": 1.7223808477926643, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1375646591186523, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7061064839363098, + "num_tokens": 405560487.0, + "step": 15684 + }, + { + "epoch": 1.7224906654952779, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.205274820327759, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7291204929351807, + "num_tokens": 405586029.0, + "step": 15685 + }, + { + "epoch": 1.7226004831978914, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.059659481048584, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7169895172119141, + "num_tokens": 405614303.0, + "step": 15686 + }, + { + "epoch": 1.7227103009005051, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.359135389328003, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7190014123916626, + "num_tokens": 405637574.0, + "step": 15687 + }, + { + "epoch": 1.722820118603119, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2314441204071045, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.695048451423645, + "num_tokens": 405664647.0, + "step": 15688 + }, + { + "epoch": 1.7229299363057324, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9696649312973022, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7294728755950928, + "num_tokens": 405695245.0, + "step": 15689 + }, + { + "epoch": 1.7230397540083462, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.127133369445801, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7379711270332336, + "num_tokens": 405721043.0, + "step": 15690 + }, + { + "epoch": 1.7231495717109597, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1657371520996094, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7058485746383667, + "num_tokens": 405748441.0, + "step": 15691 + }, + { + "epoch": 1.7232593894135735, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4559829235076904, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7146770358085632, + "num_tokens": 405770211.0, + "step": 15692 + }, + { + "epoch": 1.7233692071161872, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1742005348205566, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7349666357040405, + "num_tokens": 405799085.0, + "step": 15693 + }, + { + "epoch": 1.7234790248188008, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2238662242889404, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7301791310310364, + "num_tokens": 405825810.0, + "step": 15694 + }, + { + "epoch": 1.7235888425214143, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2175984382629395, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.7006514072418213, + "num_tokens": 405853812.0, + "step": 15695 + }, + { + "epoch": 1.723698660224028, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.355207681655884, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7295005917549133, + "num_tokens": 405879039.0, + "step": 15696 + }, + { + "epoch": 1.7238084779266418, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1731066703796387, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7124320268630981, + "num_tokens": 405908301.0, + "step": 15697 + }, + { + "epoch": 1.7239182956292556, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5041286945343018, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7377076148986816, + "num_tokens": 405928513.0, + "step": 15698 + }, + { + "epoch": 1.7240281133318691, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0530505180358887, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7068554759025574, + "num_tokens": 405958633.0, + "step": 15699 + }, + { + "epoch": 1.7241379310344827, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2740180492401123, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7299304604530334, + "num_tokens": 405983518.0, + "step": 15700 + }, + { + "epoch": 1.7242477487370964, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1723790168762207, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7270915508270264, + "num_tokens": 406009479.0, + "step": 15701 + }, + { + "epoch": 1.7243575664397102, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.634263753890991, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7200084328651428, + "num_tokens": 406030749.0, + "step": 15702 + }, + { + "epoch": 1.7244673841423237, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3103902339935303, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7170474529266357, + "num_tokens": 406056447.0, + "step": 15703 + }, + { + "epoch": 1.7245772018449375, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.76635479927063, + "learning_rate": 1e-06, + "loss": 0.78, + "mean_token_accuracy": 0.7538504600524902, + "num_tokens": 406074410.0, + "step": 15704 + }, + { + "epoch": 1.724687019547551, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5596792697906494, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7432475090026855, + "num_tokens": 406096397.0, + "step": 15705 + }, + { + "epoch": 1.7247968372501647, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.194600820541382, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7008441686630249, + "num_tokens": 406123869.0, + "step": 15706 + }, + { + "epoch": 1.7249066549527785, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3604965209960938, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7474789023399353, + "num_tokens": 406147337.0, + "step": 15707 + }, + { + "epoch": 1.725016472655392, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3423094749450684, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7132290601730347, + "num_tokens": 406171769.0, + "step": 15708 + }, + { + "epoch": 1.7251262903580056, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.270381450653076, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.712180495262146, + "num_tokens": 406196379.0, + "step": 15709 + }, + { + "epoch": 1.7252361080606193, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.110440731048584, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7168459892272949, + "num_tokens": 406226674.0, + "step": 15710 + }, + { + "epoch": 1.725345925763233, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.168647050857544, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7302924990653992, + "num_tokens": 406251785.0, + "step": 15711 + }, + { + "epoch": 1.7254557434658468, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4318625926971436, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7407851219177246, + "num_tokens": 406272446.0, + "step": 15712 + }, + { + "epoch": 1.7255655611684604, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.302530288696289, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7272260189056396, + "num_tokens": 406297309.0, + "step": 15713 + }, + { + "epoch": 1.725675378871074, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.326502561569214, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7166424989700317, + "num_tokens": 406322257.0, + "step": 15714 + }, + { + "epoch": 1.7257851965736877, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.308932304382324, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7168341875076294, + "num_tokens": 406346570.0, + "step": 15715 + }, + { + "epoch": 1.7258950142763014, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.139317274093628, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7229700088500977, + "num_tokens": 406372128.0, + "step": 15716 + }, + { + "epoch": 1.726004831978915, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.520432710647583, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7431068420410156, + "num_tokens": 406393200.0, + "step": 15717 + }, + { + "epoch": 1.7261146496815285, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2190306186676025, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6909418702125549, + "num_tokens": 406421258.0, + "step": 15718 + }, + { + "epoch": 1.7262244673841423, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.902347207069397, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7356359958648682, + "num_tokens": 406454628.0, + "step": 15719 + }, + { + "epoch": 1.726334285086756, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.023688316345215, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.73725825548172, + "num_tokens": 406484194.0, + "step": 15720 + }, + { + "epoch": 1.7264441027893698, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.415560245513916, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7172720432281494, + "num_tokens": 406508641.0, + "step": 15721 + }, + { + "epoch": 1.7265539204919833, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1840338706970215, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6962307691574097, + "num_tokens": 406535886.0, + "step": 15722 + }, + { + "epoch": 1.7266637381945968, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5535335540771484, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7189395427703857, + "num_tokens": 406557711.0, + "step": 15723 + }, + { + "epoch": 1.7267735558972106, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3869788646698, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7247589826583862, + "num_tokens": 406579582.0, + "step": 15724 + }, + { + "epoch": 1.7268833735998244, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.32625150680542, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7382786870002747, + "num_tokens": 406603065.0, + "step": 15725 + }, + { + "epoch": 1.7269931913024381, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2732667922973633, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7130522727966309, + "num_tokens": 406627249.0, + "step": 15726 + }, + { + "epoch": 1.7271030090050516, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2073333263397217, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7408645749092102, + "num_tokens": 406651523.0, + "step": 15727 + }, + { + "epoch": 1.7272128267076652, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.20379638671875, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6970230340957642, + "num_tokens": 406677929.0, + "step": 15728 + }, + { + "epoch": 1.727322644410279, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3854823112487793, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7172398567199707, + "num_tokens": 406703411.0, + "step": 15729 + }, + { + "epoch": 1.7274324621128927, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2794690132141113, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7135717868804932, + "num_tokens": 406727723.0, + "step": 15730 + }, + { + "epoch": 1.7275422798155062, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.254068613052368, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7029229402542114, + "num_tokens": 406755501.0, + "step": 15731 + }, + { + "epoch": 1.7276520975181198, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4859938621520996, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7330919504165649, + "num_tokens": 406775289.0, + "step": 15732 + }, + { + "epoch": 1.7277619152207335, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.160355567932129, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7291409373283386, + "num_tokens": 406804107.0, + "step": 15733 + }, + { + "epoch": 1.7278717329233473, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1676032543182373, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7162976264953613, + "num_tokens": 406832425.0, + "step": 15734 + }, + { + "epoch": 1.727981550625961, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3096201419830322, + "learning_rate": 1e-06, + "loss": 0.7481, + "mean_token_accuracy": 0.7604794502258301, + "num_tokens": 406855526.0, + "step": 15735 + }, + { + "epoch": 1.7280913683285746, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.536318778991699, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7215039134025574, + "num_tokens": 406876907.0, + "step": 15736 + }, + { + "epoch": 1.728201186031188, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2105491161346436, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6984531879425049, + "num_tokens": 406904760.0, + "step": 15737 + }, + { + "epoch": 1.7283110037338019, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4239864349365234, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7254641056060791, + "num_tokens": 406928549.0, + "step": 15738 + }, + { + "epoch": 1.7284208214364156, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.6110048294067383, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7138802409172058, + "num_tokens": 406949708.0, + "step": 15739 + }, + { + "epoch": 1.7285306391390292, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.073202610015869, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7119282484054565, + "num_tokens": 406977852.0, + "step": 15740 + }, + { + "epoch": 1.728640456841643, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.221123456954956, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7098479866981506, + "num_tokens": 407004183.0, + "step": 15741 + }, + { + "epoch": 1.7287502745442564, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.215338945388794, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.724919855594635, + "num_tokens": 407029641.0, + "step": 15742 + }, + { + "epoch": 1.7288600922468702, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9749919176101685, + "learning_rate": 1e-06, + "loss": 1.0199, + "mean_token_accuracy": 0.6945077180862427, + "num_tokens": 407061311.0, + "step": 15743 + }, + { + "epoch": 1.728969909949484, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.444979429244995, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7342722415924072, + "num_tokens": 407083842.0, + "step": 15744 + }, + { + "epoch": 1.7290797276520975, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4444077014923096, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7117549180984497, + "num_tokens": 407106403.0, + "step": 15745 + }, + { + "epoch": 1.729189545354711, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.539151191711426, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7358620166778564, + "num_tokens": 407127611.0, + "step": 15746 + }, + { + "epoch": 1.7292993630573248, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1527247428894043, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7073471546173096, + "num_tokens": 407155998.0, + "step": 15747 + }, + { + "epoch": 1.7294091807599385, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1900575160980225, + "learning_rate": 1e-06, + "loss": 0.7855, + "mean_token_accuracy": 0.7553902268409729, + "num_tokens": 407180192.0, + "step": 15748 + }, + { + "epoch": 1.7295189984625523, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2099130153656006, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7073317766189575, + "num_tokens": 407207098.0, + "step": 15749 + }, + { + "epoch": 1.7296288161651658, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.374309539794922, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7406548261642456, + "num_tokens": 407229106.0, + "step": 15750 + }, + { + "epoch": 1.7297386338677794, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4119820594787598, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7324679493904114, + "num_tokens": 407251733.0, + "step": 15751 + }, + { + "epoch": 1.7298484515703931, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.253749132156372, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7097723484039307, + "num_tokens": 407276925.0, + "step": 15752 + }, + { + "epoch": 1.7299582692730069, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.271695137023926, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7140142917633057, + "num_tokens": 407301318.0, + "step": 15753 + }, + { + "epoch": 1.7300680869756204, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5159900188446045, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7323140501976013, + "num_tokens": 407323652.0, + "step": 15754 + }, + { + "epoch": 1.7301779046782342, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.015132427215576, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.717584490776062, + "num_tokens": 407353689.0, + "step": 15755 + }, + { + "epoch": 1.7302877223808477, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3187525272369385, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7061564922332764, + "num_tokens": 407378942.0, + "step": 15756 + }, + { + "epoch": 1.7303975400834615, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.280113935470581, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7230055332183838, + "num_tokens": 407404492.0, + "step": 15757 + }, + { + "epoch": 1.7305073577860752, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2940564155578613, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7215856313705444, + "num_tokens": 407429503.0, + "step": 15758 + }, + { + "epoch": 1.7306171754886888, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0869452953338623, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7211440801620483, + "num_tokens": 407459862.0, + "step": 15759 + }, + { + "epoch": 1.7307269931913023, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3229379653930664, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7284446954727173, + "num_tokens": 407485344.0, + "step": 15760 + }, + { + "epoch": 1.730836810893916, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.347855806350708, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7275599837303162, + "num_tokens": 407508639.0, + "step": 15761 + }, + { + "epoch": 1.7309466285965298, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.246978759765625, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7141993045806885, + "num_tokens": 407535977.0, + "step": 15762 + }, + { + "epoch": 1.7310564462991436, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.412893533706665, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7389503121376038, + "num_tokens": 407557839.0, + "step": 15763 + }, + { + "epoch": 1.731166264001757, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.574526309967041, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7282904982566833, + "num_tokens": 407577992.0, + "step": 15764 + }, + { + "epoch": 1.7312760817043706, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.250476598739624, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7403272390365601, + "num_tokens": 407602088.0, + "step": 15765 + }, + { + "epoch": 1.7313858994069844, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5542349815368652, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7358524799346924, + "num_tokens": 407624483.0, + "step": 15766 + }, + { + "epoch": 1.7314957171095982, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.46079158782959, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.6991888284683228, + "num_tokens": 407648718.0, + "step": 15767 + }, + { + "epoch": 1.7316055348122117, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.430981159210205, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7211072444915771, + "num_tokens": 407671524.0, + "step": 15768 + }, + { + "epoch": 1.7317153525148252, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.367182970046997, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7155368328094482, + "num_tokens": 407696736.0, + "step": 15769 + }, + { + "epoch": 1.731825170217439, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3713808059692383, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.734574556350708, + "num_tokens": 407719242.0, + "step": 15770 + }, + { + "epoch": 1.7319349879200527, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.6558783054351807, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7304302453994751, + "num_tokens": 407741647.0, + "step": 15771 + }, + { + "epoch": 1.7320448056226665, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.320739269256592, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7286574840545654, + "num_tokens": 407766431.0, + "step": 15772 + }, + { + "epoch": 1.73215462332528, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1730902194976807, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7196907997131348, + "num_tokens": 407791725.0, + "step": 15773 + }, + { + "epoch": 1.7322644410278936, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.418740749359131, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.7000985145568848, + "num_tokens": 407814167.0, + "step": 15774 + }, + { + "epoch": 1.7323742587305073, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0634889602661133, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.6969131231307983, + "num_tokens": 407843181.0, + "step": 15775 + }, + { + "epoch": 1.732484076433121, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.255634307861328, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6960783004760742, + "num_tokens": 407868658.0, + "step": 15776 + }, + { + "epoch": 1.7325938941357348, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.137984275817871, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7353750467300415, + "num_tokens": 407895419.0, + "step": 15777 + }, + { + "epoch": 1.7327037118383484, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2153170108795166, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7270598411560059, + "num_tokens": 407920037.0, + "step": 15778 + }, + { + "epoch": 1.732813529540962, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.391033887863159, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7176899909973145, + "num_tokens": 407943819.0, + "step": 15779 + }, + { + "epoch": 1.7329233472435757, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.0308783054351807, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7118381261825562, + "num_tokens": 407974168.0, + "step": 15780 + }, + { + "epoch": 1.7330331649461894, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.252230644226074, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7300087213516235, + "num_tokens": 408000107.0, + "step": 15781 + }, + { + "epoch": 1.733142982648803, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2081732749938965, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6841405630111694, + "num_tokens": 408029346.0, + "step": 15782 + }, + { + "epoch": 1.7332528003514165, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3745481967926025, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7066768407821655, + "num_tokens": 408054076.0, + "step": 15783 + }, + { + "epoch": 1.7333626180540302, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.9572584629058838, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7046374082565308, + "num_tokens": 408087889.0, + "step": 15784 + }, + { + "epoch": 1.733472435756644, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1366617679595947, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7170134782791138, + "num_tokens": 408115327.0, + "step": 15785 + }, + { + "epoch": 1.7335822534592578, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.492274761199951, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.6970612406730652, + "num_tokens": 408139826.0, + "step": 15786 + }, + { + "epoch": 1.7336920711618713, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.475226879119873, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7150908708572388, + "num_tokens": 408163102.0, + "step": 15787 + }, + { + "epoch": 1.7338018888644848, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5428428649902344, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7141000628471375, + "num_tokens": 408186929.0, + "step": 15788 + }, + { + "epoch": 1.7339117065670986, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2881650924682617, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6861644983291626, + "num_tokens": 408213628.0, + "step": 15789 + }, + { + "epoch": 1.7340215242697123, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.1370139122009277, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.6982164978981018, + "num_tokens": 408244207.0, + "step": 15790 + }, + { + "epoch": 1.734131341972326, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.035186290740967, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7303321361541748, + "num_tokens": 408273952.0, + "step": 15791 + }, + { + "epoch": 1.7342411596749396, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.160266399383545, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7224709391593933, + "num_tokens": 408300411.0, + "step": 15792 + }, + { + "epoch": 1.7343509773775532, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.492392063140869, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7367439270019531, + "num_tokens": 408321815.0, + "step": 15793 + }, + { + "epoch": 1.734460795080167, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5202322006225586, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7459191083908081, + "num_tokens": 408343248.0, + "step": 15794 + }, + { + "epoch": 1.7345706127827807, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3299179077148438, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7025337815284729, + "num_tokens": 408368877.0, + "step": 15795 + }, + { + "epoch": 1.7346804304853942, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2695887088775635, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7036119699478149, + "num_tokens": 408394331.0, + "step": 15796 + }, + { + "epoch": 1.7347902481880078, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.195937156677246, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7015703916549683, + "num_tokens": 408419574.0, + "step": 15797 + }, + { + "epoch": 1.7349000658906215, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.113586902618408, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7204039692878723, + "num_tokens": 408446745.0, + "step": 15798 + }, + { + "epoch": 1.7350098835932353, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3655147552490234, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7123802900314331, + "num_tokens": 408470346.0, + "step": 15799 + }, + { + "epoch": 1.735119701295849, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.171804428100586, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7165608406066895, + "num_tokens": 408497927.0, + "step": 15800 + }, + { + "epoch": 1.7352295189984626, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2651443481445312, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7076386213302612, + "num_tokens": 408523528.0, + "step": 15801 + }, + { + "epoch": 1.735339336701076, + "ewc_loss": 0.000133514404296875, + "grad_norm": 1.8400969505310059, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7103010416030884, + "num_tokens": 408557516.0, + "step": 15802 + }, + { + "epoch": 1.7354491544036899, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.2883169651031494, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7103265523910522, + "num_tokens": 408582193.0, + "step": 15803 + }, + { + "epoch": 1.7355589721063036, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.5997159481048584, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7310537099838257, + "num_tokens": 408602622.0, + "step": 15804 + }, + { + "epoch": 1.7356687898089171, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.4847569465637207, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7253363132476807, + "num_tokens": 408624316.0, + "step": 15805 + }, + { + "epoch": 1.735778607511531, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.122750759124756, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7266888618469238, + "num_tokens": 408652235.0, + "step": 15806 + }, + { + "epoch": 1.7358884252141444, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.262458324432373, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7153297662734985, + "num_tokens": 408677377.0, + "step": 15807 + }, + { + "epoch": 1.7359982429167582, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.240626573562622, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7139024138450623, + "num_tokens": 408703453.0, + "step": 15808 + }, + { + "epoch": 1.736108060619372, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3331997394561768, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7441230416297913, + "num_tokens": 408727630.0, + "step": 15809 + }, + { + "epoch": 1.7362178783219855, + "ewc_loss": 0.000133514404296875, + "grad_norm": 2.3642563819885254, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7127293348312378, + "num_tokens": 408750923.0, + "step": 15810 + }, + { + "epoch": 1.736327696024599, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2222819328308105, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7109106779098511, + "num_tokens": 408778970.0, + "step": 15811 + }, + { + "epoch": 1.7364375137272128, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2872297763824463, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.743166446685791, + "num_tokens": 408803405.0, + "step": 15812 + }, + { + "epoch": 1.7365473314298265, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.236886501312256, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.6991080641746521, + "num_tokens": 408828951.0, + "step": 15813 + }, + { + "epoch": 1.7366571491324403, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9200927019119263, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6797261238098145, + "num_tokens": 408864782.0, + "step": 15814 + }, + { + "epoch": 1.7367669668350538, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.6290178298950195, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7380071878433228, + "num_tokens": 408883987.0, + "step": 15815 + }, + { + "epoch": 1.7368767845376674, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.190673351287842, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7066969871520996, + "num_tokens": 408911114.0, + "step": 15816 + }, + { + "epoch": 1.7369866022402811, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5102696418762207, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.735873281955719, + "num_tokens": 408932732.0, + "step": 15817 + }, + { + "epoch": 1.7370964199428949, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4393441677093506, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.740544319152832, + "num_tokens": 408956139.0, + "step": 15818 + }, + { + "epoch": 1.7372062376455084, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4785308837890625, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7287372946739197, + "num_tokens": 408979450.0, + "step": 15819 + }, + { + "epoch": 1.7373160553481222, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3816945552825928, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7376875877380371, + "num_tokens": 409002261.0, + "step": 15820 + }, + { + "epoch": 1.7374258730507357, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.250603675842285, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7112826108932495, + "num_tokens": 409031473.0, + "step": 15821 + }, + { + "epoch": 1.7375356907533495, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4220986366271973, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7213004231452942, + "num_tokens": 409053824.0, + "step": 15822 + }, + { + "epoch": 1.7376455084559632, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.298281669616699, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7069187760353088, + "num_tokens": 409081507.0, + "step": 15823 + }, + { + "epoch": 1.7377553261585768, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4566309452056885, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7444552183151245, + "num_tokens": 409104613.0, + "step": 15824 + }, + { + "epoch": 1.7378651438611903, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2143445014953613, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7144972085952759, + "num_tokens": 409131953.0, + "step": 15825 + }, + { + "epoch": 1.737974961563804, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.354809045791626, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7117717862129211, + "num_tokens": 409156621.0, + "step": 15826 + }, + { + "epoch": 1.7380847792664178, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.292846918106079, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7103211879730225, + "num_tokens": 409183001.0, + "step": 15827 + }, + { + "epoch": 1.7381945969690316, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4396793842315674, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7328181266784668, + "num_tokens": 409203539.0, + "step": 15828 + }, + { + "epoch": 1.738304414671645, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2200937271118164, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7240084409713745, + "num_tokens": 409230142.0, + "step": 15829 + }, + { + "epoch": 1.7384142323742586, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.232240915298462, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.700255811214447, + "num_tokens": 409255909.0, + "step": 15830 + }, + { + "epoch": 1.7385240500768724, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.115206241607666, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7345485687255859, + "num_tokens": 409282733.0, + "step": 15831 + }, + { + "epoch": 1.7386338677794861, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1492931842803955, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7312784194946289, + "num_tokens": 409309821.0, + "step": 15832 + }, + { + "epoch": 1.7387436854820997, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2395102977752686, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7456737756729126, + "num_tokens": 409334055.0, + "step": 15833 + }, + { + "epoch": 1.7388535031847132, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5778658390045166, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7159296274185181, + "num_tokens": 409354915.0, + "step": 15834 + }, + { + "epoch": 1.738963320887327, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.8333394527435303, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7490639686584473, + "num_tokens": 409372334.0, + "step": 15835 + }, + { + "epoch": 1.7390731385899407, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.08353328704834, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6931612491607666, + "num_tokens": 409401713.0, + "step": 15836 + }, + { + "epoch": 1.7391829562925545, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.088486671447754, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.696088433265686, + "num_tokens": 409431433.0, + "step": 15837 + }, + { + "epoch": 1.739292773995168, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.292203903198242, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.719918966293335, + "num_tokens": 409454487.0, + "step": 15838 + }, + { + "epoch": 1.7394025916977816, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1814587116241455, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7249659299850464, + "num_tokens": 409479958.0, + "step": 15839 + }, + { + "epoch": 1.7395124094003953, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.067596912384033, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7241531610488892, + "num_tokens": 409509442.0, + "step": 15840 + }, + { + "epoch": 1.739622227103009, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.523254871368408, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7086049914360046, + "num_tokens": 409530593.0, + "step": 15841 + }, + { + "epoch": 1.7397320448056228, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9865473508834839, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7153953313827515, + "num_tokens": 409563789.0, + "step": 15842 + }, + { + "epoch": 1.7398418625082364, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0580224990844727, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7080661654472351, + "num_tokens": 409592783.0, + "step": 15843 + }, + { + "epoch": 1.73995168021085, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2791812419891357, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7186226844787598, + "num_tokens": 409617309.0, + "step": 15844 + }, + { + "epoch": 1.7400614979134637, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3146486282348633, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7119876146316528, + "num_tokens": 409642941.0, + "step": 15845 + }, + { + "epoch": 1.7401713156160774, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.116830587387085, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7280206084251404, + "num_tokens": 409669861.0, + "step": 15846 + }, + { + "epoch": 1.740281133318691, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.449160099029541, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7089726328849792, + "num_tokens": 409691814.0, + "step": 15847 + }, + { + "epoch": 1.7403909510213045, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2769839763641357, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7122285962104797, + "num_tokens": 409717653.0, + "step": 15848 + }, + { + "epoch": 1.7405007687239182, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4553744792938232, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7261927723884583, + "num_tokens": 409738514.0, + "step": 15849 + }, + { + "epoch": 1.740610586426532, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3199806213378906, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7040300965309143, + "num_tokens": 409762917.0, + "step": 15850 + }, + { + "epoch": 1.7407204041291457, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3680579662323, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7275918126106262, + "num_tokens": 409786764.0, + "step": 15851 + }, + { + "epoch": 1.7408302218317593, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3218579292297363, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7118998169898987, + "num_tokens": 409814012.0, + "step": 15852 + }, + { + "epoch": 1.7409400395343728, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.273559093475342, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7138099074363708, + "num_tokens": 409841290.0, + "step": 15853 + }, + { + "epoch": 1.7410498572369866, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2923641204833984, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.71431565284729, + "num_tokens": 409865997.0, + "step": 15854 + }, + { + "epoch": 1.7411596749396003, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3654158115386963, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7063090205192566, + "num_tokens": 409888915.0, + "step": 15855 + }, + { + "epoch": 1.741269492642214, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.869956374168396, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.700311541557312, + "num_tokens": 409922378.0, + "step": 15856 + }, + { + "epoch": 1.7413793103448276, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2135887145996094, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7535368800163269, + "num_tokens": 409945678.0, + "step": 15857 + }, + { + "epoch": 1.7414891280474412, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2932348251342773, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7127863168716431, + "num_tokens": 409968835.0, + "step": 15858 + }, + { + "epoch": 1.741598945750055, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1196770668029785, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7124254107475281, + "num_tokens": 409997469.0, + "step": 15859 + }, + { + "epoch": 1.7417087634526687, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.280914783477783, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7049037218093872, + "num_tokens": 410022375.0, + "step": 15860 + }, + { + "epoch": 1.7418185811552822, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.068932294845581, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6921952962875366, + "num_tokens": 410054020.0, + "step": 15861 + }, + { + "epoch": 1.7419283988578957, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1977598667144775, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7038341164588928, + "num_tokens": 410079572.0, + "step": 15862 + }, + { + "epoch": 1.7420382165605095, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1094448566436768, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7294953465461731, + "num_tokens": 410106514.0, + "step": 15863 + }, + { + "epoch": 1.7421480342631233, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.199636697769165, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7170369625091553, + "num_tokens": 410132343.0, + "step": 15864 + }, + { + "epoch": 1.742257851965737, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.129274845123291, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7180610299110413, + "num_tokens": 410161866.0, + "step": 15865 + }, + { + "epoch": 1.7423676696683505, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2857141494750977, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.6946430206298828, + "num_tokens": 410187236.0, + "step": 15866 + }, + { + "epoch": 1.742477487370964, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1270487308502197, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.718289315700531, + "num_tokens": 410216432.0, + "step": 15867 + }, + { + "epoch": 1.7425873050735778, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0730767250061035, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7246161699295044, + "num_tokens": 410244252.0, + "step": 15868 + }, + { + "epoch": 1.7426971227761916, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1744322776794434, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7017139196395874, + "num_tokens": 410271135.0, + "step": 15869 + }, + { + "epoch": 1.7428069404788051, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.23628306388855, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7195142507553101, + "num_tokens": 410296705.0, + "step": 15870 + }, + { + "epoch": 1.742916758181419, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2480928897857666, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.686869740486145, + "num_tokens": 410324785.0, + "step": 15871 + }, + { + "epoch": 1.7430265758840324, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1853325366973877, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7274985909461975, + "num_tokens": 410350793.0, + "step": 15872 + }, + { + "epoch": 1.7431363935866462, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1594889163970947, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7216540575027466, + "num_tokens": 410380193.0, + "step": 15873 + }, + { + "epoch": 1.74324621128926, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.395536422729492, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7409908175468445, + "num_tokens": 410403374.0, + "step": 15874 + }, + { + "epoch": 1.7433560289918735, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1170523166656494, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7076056003570557, + "num_tokens": 410431523.0, + "step": 15875 + }, + { + "epoch": 1.743465846694487, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2460665702819824, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7341362237930298, + "num_tokens": 410455780.0, + "step": 15876 + }, + { + "epoch": 1.7435756643971008, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.169410228729248, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.698676347732544, + "num_tokens": 410485812.0, + "step": 15877 + }, + { + "epoch": 1.7436854820997145, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.6149215698242188, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7114973664283752, + "num_tokens": 410507164.0, + "step": 15878 + }, + { + "epoch": 1.7437952998023283, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.13136887550354, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7288991212844849, + "num_tokens": 410534324.0, + "step": 15879 + }, + { + "epoch": 1.7439051175049418, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1736137866973877, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.6859843730926514, + "num_tokens": 410564035.0, + "step": 15880 + }, + { + "epoch": 1.7440149352075554, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.7345082759857178, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.729266881942749, + "num_tokens": 410582722.0, + "step": 15881 + }, + { + "epoch": 1.744124752910169, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3517158031463623, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7342995405197144, + "num_tokens": 410606232.0, + "step": 15882 + }, + { + "epoch": 1.7442345706127829, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.15849232673645, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7084755897521973, + "num_tokens": 410631951.0, + "step": 15883 + }, + { + "epoch": 1.7443443883153964, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0211527347564697, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7489259243011475, + "num_tokens": 410660421.0, + "step": 15884 + }, + { + "epoch": 1.7444542060180102, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.276958465576172, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7019296884536743, + "num_tokens": 410688148.0, + "step": 15885 + }, + { + "epoch": 1.7445640237206237, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.999374270439148, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7038233876228333, + "num_tokens": 410721790.0, + "step": 15886 + }, + { + "epoch": 1.7446738414232374, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.768190860748291, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7314672470092773, + "num_tokens": 410740257.0, + "step": 15887 + }, + { + "epoch": 1.7447836591258512, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.233262062072754, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7335890531539917, + "num_tokens": 410763734.0, + "step": 15888 + }, + { + "epoch": 1.7448934768284647, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4865670204162598, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7299579381942749, + "num_tokens": 410785032.0, + "step": 15889 + }, + { + "epoch": 1.7450032945310783, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.239414930343628, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6926215291023254, + "num_tokens": 410810970.0, + "step": 15890 + }, + { + "epoch": 1.745113112233692, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.159741163253784, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.6926778554916382, + "num_tokens": 410841451.0, + "step": 15891 + }, + { + "epoch": 1.7452229299363058, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.336507558822632, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7198099493980408, + "num_tokens": 410865315.0, + "step": 15892 + }, + { + "epoch": 1.7453327476389195, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3691892623901367, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.710989236831665, + "num_tokens": 410889008.0, + "step": 15893 + }, + { + "epoch": 1.745442565341533, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.499872922897339, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.701543927192688, + "num_tokens": 410912808.0, + "step": 15894 + }, + { + "epoch": 1.7455523830441466, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.232295036315918, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7178570032119751, + "num_tokens": 410940907.0, + "step": 15895 + }, + { + "epoch": 1.7456622007467604, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3463780879974365, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7174418568611145, + "num_tokens": 410964706.0, + "step": 15896 + }, + { + "epoch": 1.7457720184493741, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1987290382385254, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7095004916191101, + "num_tokens": 410992747.0, + "step": 15897 + }, + { + "epoch": 1.7458818361519877, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.275211811065674, + "learning_rate": 1e-06, + "loss": 0.7772, + "mean_token_accuracy": 0.7563591599464417, + "num_tokens": 411016194.0, + "step": 15898 + }, + { + "epoch": 1.7459916538546012, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.530311107635498, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7051191329956055, + "num_tokens": 411037537.0, + "step": 15899 + }, + { + "epoch": 1.746101471557215, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.346776008605957, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7293097376823425, + "num_tokens": 411061701.0, + "step": 15900 + }, + { + "epoch": 1.7462112892598287, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2959234714508057, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7246713042259216, + "num_tokens": 411086094.0, + "step": 15901 + }, + { + "epoch": 1.7463211069624425, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4728126525878906, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7144639492034912, + "num_tokens": 411107452.0, + "step": 15902 + }, + { + "epoch": 1.746430924665056, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.411066770553589, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7370273470878601, + "num_tokens": 411129754.0, + "step": 15903 + }, + { + "epoch": 1.7465407423676695, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3297431468963623, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.710909366607666, + "num_tokens": 411155350.0, + "step": 15904 + }, + { + "epoch": 1.7466505600702833, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0486371517181396, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7208899259567261, + "num_tokens": 411183490.0, + "step": 15905 + }, + { + "epoch": 1.746760377772897, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.364051580429077, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7096408009529114, + "num_tokens": 411210915.0, + "step": 15906 + }, + { + "epoch": 1.7468701954755108, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2206056118011475, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7251235246658325, + "num_tokens": 411237205.0, + "step": 15907 + }, + { + "epoch": 1.7469800131781243, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2863352298736572, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7060152292251587, + "num_tokens": 411264664.0, + "step": 15908 + }, + { + "epoch": 1.7470898308807379, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.31549072265625, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7192078828811646, + "num_tokens": 411291043.0, + "step": 15909 + }, + { + "epoch": 1.7471996485833516, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4490408897399902, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7077762484550476, + "num_tokens": 411313227.0, + "step": 15910 + }, + { + "epoch": 1.7473094662859654, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3088393211364746, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6957834959030151, + "num_tokens": 411339882.0, + "step": 15911 + }, + { + "epoch": 1.747419283988579, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2369120121002197, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7160570621490479, + "num_tokens": 411366020.0, + "step": 15912 + }, + { + "epoch": 1.7475291016911925, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.234457492828369, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7272868156433105, + "num_tokens": 411388987.0, + "step": 15913 + }, + { + "epoch": 1.7476389193938062, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2003014087677, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.734673023223877, + "num_tokens": 411417028.0, + "step": 15914 + }, + { + "epoch": 1.74774873709642, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4806032180786133, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7090674042701721, + "num_tokens": 411439589.0, + "step": 15915 + }, + { + "epoch": 1.7478585547990337, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5533835887908936, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7003217935562134, + "num_tokens": 411461548.0, + "step": 15916 + }, + { + "epoch": 1.7479683725016473, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0349972248077393, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7299466133117676, + "num_tokens": 411489361.0, + "step": 15917 + }, + { + "epoch": 1.7480781902042608, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.01546573638916, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7109729051589966, + "num_tokens": 411519407.0, + "step": 15918 + }, + { + "epoch": 1.7481880079068746, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0163004398345947, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7243608236312866, + "num_tokens": 411547927.0, + "step": 15919 + }, + { + "epoch": 1.7482978256094883, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.129883289337158, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.711949348449707, + "num_tokens": 411575453.0, + "step": 15920 + }, + { + "epoch": 1.7484076433121019, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1288504600524902, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7278507351875305, + "num_tokens": 411603169.0, + "step": 15921 + }, + { + "epoch": 1.7485174610147156, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9863147735595703, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7029017210006714, + "num_tokens": 411635310.0, + "step": 15922 + }, + { + "epoch": 1.7486272787173291, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9209582805633545, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7354031801223755, + "num_tokens": 411666693.0, + "step": 15923 + }, + { + "epoch": 1.748737096419943, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1977651119232178, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7100836038589478, + "num_tokens": 411694091.0, + "step": 15924 + }, + { + "epoch": 1.7488469141225567, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.6539037227630615, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7200063467025757, + "num_tokens": 411715521.0, + "step": 15925 + }, + { + "epoch": 1.7489567318251702, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0776617527008057, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7224871516227722, + "num_tokens": 411744107.0, + "step": 15926 + }, + { + "epoch": 1.7490665495277837, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9765503406524658, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7118232250213623, + "num_tokens": 411774228.0, + "step": 15927 + }, + { + "epoch": 1.7491763672303975, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.7522428035736084, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7290195226669312, + "num_tokens": 411791279.0, + "step": 15928 + }, + { + "epoch": 1.7492861849330112, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3782055377960205, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7318045496940613, + "num_tokens": 411813971.0, + "step": 15929 + }, + { + "epoch": 1.749396002635625, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3937346935272217, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7226975560188293, + "num_tokens": 411836624.0, + "step": 15930 + }, + { + "epoch": 1.7495058203382385, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.277113437652588, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7163777351379395, + "num_tokens": 411862879.0, + "step": 15931 + }, + { + "epoch": 1.749615638040852, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.143941879272461, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7198875546455383, + "num_tokens": 411890777.0, + "step": 15932 + }, + { + "epoch": 1.7497254557434658, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.8906525373458862, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7061970233917236, + "num_tokens": 411929648.0, + "step": 15933 + }, + { + "epoch": 1.7498352734460796, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9604640007019043, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7399559617042542, + "num_tokens": 411959716.0, + "step": 15934 + }, + { + "epoch": 1.7499450911486931, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2708990573883057, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7226594090461731, + "num_tokens": 411984294.0, + "step": 15935 + }, + { + "epoch": 1.7500549088513069, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.255425453186035, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7166568040847778, + "num_tokens": 412008278.0, + "step": 15936 + }, + { + "epoch": 1.7501647265539204, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.237476110458374, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.730291485786438, + "num_tokens": 412035117.0, + "step": 15937 + }, + { + "epoch": 1.7502745442565342, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2523610591888428, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.722373366355896, + "num_tokens": 412059205.0, + "step": 15938 + }, + { + "epoch": 1.750384361959148, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.433530330657959, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.707206130027771, + "num_tokens": 412082773.0, + "step": 15939 + }, + { + "epoch": 1.7504941796617615, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.107851505279541, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7168096303939819, + "num_tokens": 412111838.0, + "step": 15940 + }, + { + "epoch": 1.750603997364375, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.7227611541748047, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7338285446166992, + "num_tokens": 412131623.0, + "step": 15941 + }, + { + "epoch": 1.7507138150669888, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0720648765563965, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7183630466461182, + "num_tokens": 412161242.0, + "step": 15942 + }, + { + "epoch": 1.7508236327696025, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.484869956970215, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7367942333221436, + "num_tokens": 412182544.0, + "step": 15943 + }, + { + "epoch": 1.7509334504722163, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2961432933807373, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7114118337631226, + "num_tokens": 412207490.0, + "step": 15944 + }, + { + "epoch": 1.7510432681748298, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.141343593597412, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7084115743637085, + "num_tokens": 412235966.0, + "step": 15945 + }, + { + "epoch": 1.7511530858774433, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4719653129577637, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7289731502532959, + "num_tokens": 412259956.0, + "step": 15946 + }, + { + "epoch": 1.751262903580057, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1577746868133545, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.697515606880188, + "num_tokens": 412289410.0, + "step": 15947 + }, + { + "epoch": 1.7513727212826709, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.210134744644165, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.716003954410553, + "num_tokens": 412316136.0, + "step": 15948 + }, + { + "epoch": 1.7514825389852844, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9814486503601074, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7275081276893616, + "num_tokens": 412344879.0, + "step": 15949 + }, + { + "epoch": 1.7515923566878981, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1220085620880127, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7212086915969849, + "num_tokens": 412371632.0, + "step": 15950 + }, + { + "epoch": 1.7517021743905117, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9702686071395874, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7286173701286316, + "num_tokens": 412400581.0, + "step": 15951 + }, + { + "epoch": 1.7518119920931254, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3283259868621826, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.6986702680587769, + "num_tokens": 412425480.0, + "step": 15952 + }, + { + "epoch": 1.7519218097957392, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3936967849731445, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7233157753944397, + "num_tokens": 412448255.0, + "step": 15953 + }, + { + "epoch": 1.7520316274983527, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3421168327331543, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7187260389328003, + "num_tokens": 412471255.0, + "step": 15954 + }, + { + "epoch": 1.7521414452009663, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5146806240081787, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7411289811134338, + "num_tokens": 412491773.0, + "step": 15955 + }, + { + "epoch": 1.75225126290358, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1243622303009033, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7201505303382874, + "num_tokens": 412521308.0, + "step": 15956 + }, + { + "epoch": 1.7523610806061938, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.230529308319092, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7112468481063843, + "num_tokens": 412547588.0, + "step": 15957 + }, + { + "epoch": 1.7524708983088075, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1665737628936768, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7047853469848633, + "num_tokens": 412576696.0, + "step": 15958 + }, + { + "epoch": 1.752580716011421, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4784164428710938, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7100180387496948, + "num_tokens": 412598512.0, + "step": 15959 + }, + { + "epoch": 1.7526905337140346, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.313678026199341, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7187014818191528, + "num_tokens": 412624152.0, + "step": 15960 + }, + { + "epoch": 1.7528003514166484, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.183941125869751, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7394645810127258, + "num_tokens": 412649925.0, + "step": 15961 + }, + { + "epoch": 1.7529101691192621, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.8233357667922974, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7034029364585876, + "num_tokens": 412689788.0, + "step": 15962 + }, + { + "epoch": 1.7530199868218757, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.431183338165283, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.741245687007904, + "num_tokens": 412712857.0, + "step": 15963 + }, + { + "epoch": 1.7531298045244892, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.299437999725342, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7069164514541626, + "num_tokens": 412739070.0, + "step": 15964 + }, + { + "epoch": 1.753239622227103, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0703036785125732, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.72414630651474, + "num_tokens": 412768299.0, + "step": 15965 + }, + { + "epoch": 1.7533494399297167, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.440692186355591, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7149860262870789, + "num_tokens": 412790910.0, + "step": 15966 + }, + { + "epoch": 1.7534592576323305, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.213444232940674, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.724578857421875, + "num_tokens": 412816326.0, + "step": 15967 + }, + { + "epoch": 1.753569075334944, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.244107723236084, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7366770505905151, + "num_tokens": 412842083.0, + "step": 15968 + }, + { + "epoch": 1.7536788930375575, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5361573696136475, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7154108285903931, + "num_tokens": 412862592.0, + "step": 15969 + }, + { + "epoch": 1.7537887107401713, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.30033802986145, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7196931838989258, + "num_tokens": 412886549.0, + "step": 15970 + }, + { + "epoch": 1.753898528442785, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.188616991043091, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6967947483062744, + "num_tokens": 412912929.0, + "step": 15971 + }, + { + "epoch": 1.7540083461453988, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4561927318573, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7233992218971252, + "num_tokens": 412934306.0, + "step": 15972 + }, + { + "epoch": 1.7541181638480123, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1602611541748047, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7186226844787598, + "num_tokens": 412963595.0, + "step": 15973 + }, + { + "epoch": 1.7542279815506259, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.929267168045044, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.711673378944397, + "num_tokens": 412999685.0, + "step": 15974 + }, + { + "epoch": 1.7543377992532396, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1099658012390137, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7121845483779907, + "num_tokens": 413030652.0, + "step": 15975 + }, + { + "epoch": 1.7544476169558534, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.204219341278076, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7059286832809448, + "num_tokens": 413058995.0, + "step": 15976 + }, + { + "epoch": 1.754557434658467, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.449995517730713, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7105236053466797, + "num_tokens": 413081066.0, + "step": 15977 + }, + { + "epoch": 1.7546672523610805, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1724417209625244, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7217553853988647, + "num_tokens": 413109710.0, + "step": 15978 + }, + { + "epoch": 1.7547770700636942, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.285705089569092, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7289773225784302, + "num_tokens": 413133397.0, + "step": 15979 + }, + { + "epoch": 1.754886887766308, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.01127552986145, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7097907066345215, + "num_tokens": 413165579.0, + "step": 15980 + }, + { + "epoch": 1.7549967054689217, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.371121406555176, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7345020771026611, + "num_tokens": 413190182.0, + "step": 15981 + }, + { + "epoch": 1.7551065231715353, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3184499740600586, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7178537249565125, + "num_tokens": 413214874.0, + "step": 15982 + }, + { + "epoch": 1.7552163408741488, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0400054454803467, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.715418815612793, + "num_tokens": 413244392.0, + "step": 15983 + }, + { + "epoch": 1.7553261585767626, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.999083399772644, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7163678407669067, + "num_tokens": 413276386.0, + "step": 15984 + }, + { + "epoch": 1.7554359762793763, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1596736907958984, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7235531806945801, + "num_tokens": 413304406.0, + "step": 15985 + }, + { + "epoch": 1.7555457939819898, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.210991859436035, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7265510559082031, + "num_tokens": 413329957.0, + "step": 15986 + }, + { + "epoch": 1.7556556116846036, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.161878824234009, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7185212969779968, + "num_tokens": 413355676.0, + "step": 15987 + }, + { + "epoch": 1.7557654293872171, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.364783525466919, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7237377166748047, + "num_tokens": 413377392.0, + "step": 15988 + }, + { + "epoch": 1.755875247089831, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3424084186553955, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7424708604812622, + "num_tokens": 413398830.0, + "step": 15989 + }, + { + "epoch": 1.7559850647924446, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.011747121810913, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7037712335586548, + "num_tokens": 413429855.0, + "step": 15990 + }, + { + "epoch": 1.7560948824950582, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1478190422058105, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.723007321357727, + "num_tokens": 413456455.0, + "step": 15991 + }, + { + "epoch": 1.7562047001976717, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.405817985534668, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.71822589635849, + "num_tokens": 413479740.0, + "step": 15992 + }, + { + "epoch": 1.7563145179002855, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.288238763809204, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7016354203224182, + "num_tokens": 413505148.0, + "step": 15993 + }, + { + "epoch": 1.7564243356028992, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.245363473892212, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7209891080856323, + "num_tokens": 413530087.0, + "step": 15994 + }, + { + "epoch": 1.756534153305513, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3086001873016357, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7361369132995605, + "num_tokens": 413556001.0, + "step": 15995 + }, + { + "epoch": 1.7566439710081265, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.388336658477783, + "learning_rate": 1e-06, + "loss": 0.796, + "mean_token_accuracy": 0.7482773661613464, + "num_tokens": 413578697.0, + "step": 15996 + }, + { + "epoch": 1.75675378871074, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.190180778503418, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.7006194591522217, + "num_tokens": 413607179.0, + "step": 15997 + }, + { + "epoch": 1.7568636064133538, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0296144485473633, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7121562957763672, + "num_tokens": 413635705.0, + "step": 15998 + }, + { + "epoch": 1.7569734241159676, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.343942165374756, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7458584308624268, + "num_tokens": 413658062.0, + "step": 15999 + }, + { + "epoch": 1.757083241818581, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3341257572174072, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7366954684257507, + "num_tokens": 413679936.0, + "step": 16000 + }, + { + "epoch": 1.7571930595211949, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5026180744171143, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7157561779022217, + "num_tokens": 413700924.0, + "step": 16001 + }, + { + "epoch": 1.7573028772238084, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.522876024246216, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7302336692810059, + "num_tokens": 413720728.0, + "step": 16002 + }, + { + "epoch": 1.7574126949264222, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2114293575286865, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7349848747253418, + "num_tokens": 413747592.0, + "step": 16003 + }, + { + "epoch": 1.757522512629036, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9576754570007324, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7188163995742798, + "num_tokens": 413779379.0, + "step": 16004 + }, + { + "epoch": 1.7576323303316495, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4797885417938232, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7131132483482361, + "num_tokens": 413801675.0, + "step": 16005 + }, + { + "epoch": 1.757742148034263, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1946065425872803, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7064206600189209, + "num_tokens": 413828275.0, + "step": 16006 + }, + { + "epoch": 1.7578519657368767, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0711286067962646, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7229960560798645, + "num_tokens": 413859114.0, + "step": 16007 + }, + { + "epoch": 1.7579617834394905, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0388736724853516, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7339907884597778, + "num_tokens": 413886825.0, + "step": 16008 + }, + { + "epoch": 1.7580716011421043, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.8737287521362305, + "learning_rate": 1e-06, + "loss": 1.0656, + "mean_token_accuracy": 0.6854271292686462, + "num_tokens": 413925084.0, + "step": 16009 + }, + { + "epoch": 1.7581814188447178, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3828043937683105, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.694307267665863, + "num_tokens": 413952206.0, + "step": 16010 + }, + { + "epoch": 1.7582912365473313, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.361694812774658, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7038959264755249, + "num_tokens": 413975722.0, + "step": 16011 + }, + { + "epoch": 1.758401054249945, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1682941913604736, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7379003763198853, + "num_tokens": 414000120.0, + "step": 16012 + }, + { + "epoch": 1.7585108719525588, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4150633811950684, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.72450852394104, + "num_tokens": 414022521.0, + "step": 16013 + }, + { + "epoch": 1.7586206896551724, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2391841411590576, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.705605149269104, + "num_tokens": 414047699.0, + "step": 16014 + }, + { + "epoch": 1.758730507357786, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4233875274658203, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7159045338630676, + "num_tokens": 414071095.0, + "step": 16015 + }, + { + "epoch": 1.7588403250603997, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.185366630554199, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7327331900596619, + "num_tokens": 414096434.0, + "step": 16016 + }, + { + "epoch": 1.7589501427630134, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.268176317214966, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7208307981491089, + "num_tokens": 414120608.0, + "step": 16017 + }, + { + "epoch": 1.7590599604656272, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2777750492095947, + "learning_rate": 1e-06, + "loss": 0.7784, + "mean_token_accuracy": 0.7594669461250305, + "num_tokens": 414142862.0, + "step": 16018 + }, + { + "epoch": 1.7591697781682407, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0313286781311035, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.6982725262641907, + "num_tokens": 414173434.0, + "step": 16019 + }, + { + "epoch": 1.7592795958708543, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.310451030731201, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7331565618515015, + "num_tokens": 414199609.0, + "step": 16020 + }, + { + "epoch": 1.759389413573468, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.347402334213257, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7141579389572144, + "num_tokens": 414225107.0, + "step": 16021 + }, + { + "epoch": 1.7594992312760818, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0979578495025635, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7363998889923096, + "num_tokens": 414251708.0, + "step": 16022 + }, + { + "epoch": 1.7596090489786955, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1233468055725098, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7494516372680664, + "num_tokens": 414279664.0, + "step": 16023 + }, + { + "epoch": 1.759718866681309, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0749964714050293, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7009512186050415, + "num_tokens": 414309844.0, + "step": 16024 + }, + { + "epoch": 1.7598286843839226, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2202506065368652, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7201112508773804, + "num_tokens": 414333773.0, + "step": 16025 + }, + { + "epoch": 1.7599385020865363, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.226449489593506, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7137295603752136, + "num_tokens": 414360364.0, + "step": 16026 + }, + { + "epoch": 1.76004831978915, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.151986837387085, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7434409260749817, + "num_tokens": 414387365.0, + "step": 16027 + }, + { + "epoch": 1.7601581374917636, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2055938243865967, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7300467491149902, + "num_tokens": 414413067.0, + "step": 16028 + }, + { + "epoch": 1.7602679551943772, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1135733127593994, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7086028456687927, + "num_tokens": 414440864.0, + "step": 16029 + }, + { + "epoch": 1.760377772896991, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.178471565246582, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7114219665527344, + "num_tokens": 414467877.0, + "step": 16030 + }, + { + "epoch": 1.7604875905996047, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.227651357650757, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.736822247505188, + "num_tokens": 414492795.0, + "step": 16031 + }, + { + "epoch": 1.7605974083022184, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 3.066357135772705, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7297463417053223, + "num_tokens": 414519885.0, + "step": 16032 + }, + { + "epoch": 1.760707226004832, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0083656311035156, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.725981593132019, + "num_tokens": 414549805.0, + "step": 16033 + }, + { + "epoch": 1.7608170437074455, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.166455030441284, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7291711568832397, + "num_tokens": 414576865.0, + "step": 16034 + }, + { + "epoch": 1.7609268614100593, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1064181327819824, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7099074125289917, + "num_tokens": 414605167.0, + "step": 16035 + }, + { + "epoch": 1.761036679112673, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4827558994293213, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7239737510681152, + "num_tokens": 414626986.0, + "step": 16036 + }, + { + "epoch": 1.7611464968152868, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.205841302871704, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7073696255683899, + "num_tokens": 414655905.0, + "step": 16037 + }, + { + "epoch": 1.7612563145179003, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.209108829498291, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.710959792137146, + "num_tokens": 414682696.0, + "step": 16038 + }, + { + "epoch": 1.7613661322205139, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.218860149383545, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.723085343837738, + "num_tokens": 414709072.0, + "step": 16039 + }, + { + "epoch": 1.7614759499231276, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1172943115234375, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7384753227233887, + "num_tokens": 414734163.0, + "step": 16040 + }, + { + "epoch": 1.7615857676257414, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2863993644714355, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7430939674377441, + "num_tokens": 414757528.0, + "step": 16041 + }, + { + "epoch": 1.761695585328355, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3168084621429443, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7148920297622681, + "num_tokens": 414781026.0, + "step": 16042 + }, + { + "epoch": 1.7618054030309684, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9779809713363647, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.720143735408783, + "num_tokens": 414811374.0, + "step": 16043 + }, + { + "epoch": 1.7619152207335822, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3607404232025146, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.723992109298706, + "num_tokens": 414834105.0, + "step": 16044 + }, + { + "epoch": 1.762025038436196, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1357717514038086, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7076531648635864, + "num_tokens": 414863277.0, + "step": 16045 + }, + { + "epoch": 1.7621348561388097, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2874648571014404, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7305542230606079, + "num_tokens": 414888365.0, + "step": 16046 + }, + { + "epoch": 1.7622446738414232, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.159590482711792, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7299014925956726, + "num_tokens": 414914796.0, + "step": 16047 + }, + { + "epoch": 1.7623544915440368, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9774091243743896, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7077988386154175, + "num_tokens": 414946541.0, + "step": 16048 + }, + { + "epoch": 1.7624643092466505, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3803069591522217, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7207558155059814, + "num_tokens": 414970385.0, + "step": 16049 + }, + { + "epoch": 1.7625741269492643, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1368207931518555, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.716576099395752, + "num_tokens": 414997241.0, + "step": 16050 + }, + { + "epoch": 1.7626839446518778, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.25608229637146, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7198326587677002, + "num_tokens": 415025244.0, + "step": 16051 + }, + { + "epoch": 1.7627937623544916, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1306710243225098, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7458451986312866, + "num_tokens": 415051255.0, + "step": 16052 + }, + { + "epoch": 1.7629035800571051, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0636146068573, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7338528633117676, + "num_tokens": 415080247.0, + "step": 16053 + }, + { + "epoch": 1.7630133977597189, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0497469902038574, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7110849618911743, + "num_tokens": 415109155.0, + "step": 16054 + }, + { + "epoch": 1.7631232154623326, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3353476524353027, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7242050170898438, + "num_tokens": 415133469.0, + "step": 16055 + }, + { + "epoch": 1.7632330331649462, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.50893235206604, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7086050510406494, + "num_tokens": 415155715.0, + "step": 16056 + }, + { + "epoch": 1.7633428508675597, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.7924420833587646, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7094144821166992, + "num_tokens": 415192246.0, + "step": 16057 + }, + { + "epoch": 1.7634526685701735, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.3473711013793945, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.720192551612854, + "num_tokens": 415219216.0, + "step": 16058 + }, + { + "epoch": 1.7635624862727872, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2667620182037354, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.726118803024292, + "num_tokens": 415245103.0, + "step": 16059 + }, + { + "epoch": 1.763672303975401, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.184793710708618, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.7000941038131714, + "num_tokens": 415276220.0, + "step": 16060 + }, + { + "epoch": 1.7637821216780145, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2102773189544678, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7303296327590942, + "num_tokens": 415303037.0, + "step": 16061 + }, + { + "epoch": 1.763891939380628, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.488663911819458, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7183285355567932, + "num_tokens": 415324617.0, + "step": 16062 + }, + { + "epoch": 1.7640017570832418, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.293633222579956, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7183240652084351, + "num_tokens": 415349364.0, + "step": 16063 + }, + { + "epoch": 1.7641115747858556, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1028246879577637, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7074078321456909, + "num_tokens": 415376114.0, + "step": 16064 + }, + { + "epoch": 1.764221392488469, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1165668964385986, + "learning_rate": 1e-06, + "loss": 1.0571, + "mean_token_accuracy": 0.6832400560379028, + "num_tokens": 415406244.0, + "step": 16065 + }, + { + "epoch": 1.7643312101910829, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.5010242462158203, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7035385370254517, + "num_tokens": 415428018.0, + "step": 16066 + }, + { + "epoch": 1.7644410278936964, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2486634254455566, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7026650905609131, + "num_tokens": 415455075.0, + "step": 16067 + }, + { + "epoch": 1.7645508455963101, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.0334503650665283, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7011152505874634, + "num_tokens": 415485792.0, + "step": 16068 + }, + { + "epoch": 1.764660663298924, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1113104820251465, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7069355249404907, + "num_tokens": 415514560.0, + "step": 16069 + }, + { + "epoch": 1.7647704810015374, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.332587957382202, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7215317487716675, + "num_tokens": 415537026.0, + "step": 16070 + }, + { + "epoch": 1.764880298704151, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.120619535446167, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6981538534164429, + "num_tokens": 415565677.0, + "step": 16071 + }, + { + "epoch": 1.7649901164067647, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1791722774505615, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7107924222946167, + "num_tokens": 415594587.0, + "step": 16072 + }, + { + "epoch": 1.7650999341093785, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.1262357234954834, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7142672538757324, + "num_tokens": 415622755.0, + "step": 16073 + }, + { + "epoch": 1.7652097518119922, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 1.9843155145645142, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7215774655342102, + "num_tokens": 415652436.0, + "step": 16074 + }, + { + "epoch": 1.7653195695146058, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.4921536445617676, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.71340411901474, + "num_tokens": 415676475.0, + "step": 16075 + }, + { + "epoch": 1.7654293872172193, + "ewc_loss": 0.0001354217529296875, + "grad_norm": 2.2054502964019775, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7102559804916382, + "num_tokens": 415703020.0, + "step": 16076 + }, + { + "epoch": 1.765539204919833, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.19970703125, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7088074684143066, + "num_tokens": 415729866.0, + "step": 16077 + }, + { + "epoch": 1.7656490226224468, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.494966983795166, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7260207533836365, + "num_tokens": 415752142.0, + "step": 16078 + }, + { + "epoch": 1.7657588403250604, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.136721611022949, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7114078998565674, + "num_tokens": 415780469.0, + "step": 16079 + }, + { + "epoch": 1.765868658027674, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1147844791412354, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7063243389129639, + "num_tokens": 415809635.0, + "step": 16080 + }, + { + "epoch": 1.7659784757302877, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4190175533294678, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7140359282493591, + "num_tokens": 415832723.0, + "step": 16081 + }, + { + "epoch": 1.7660882934329014, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3006961345672607, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7023658752441406, + "num_tokens": 415857939.0, + "step": 16082 + }, + { + "epoch": 1.7661981111355152, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3427672386169434, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7174750566482544, + "num_tokens": 415881968.0, + "step": 16083 + }, + { + "epoch": 1.7663079288381287, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1231093406677246, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7138595581054688, + "num_tokens": 415911357.0, + "step": 16084 + }, + { + "epoch": 1.7664177465407422, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1801598072052, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7309246063232422, + "num_tokens": 415936779.0, + "step": 16085 + }, + { + "epoch": 1.766527564243356, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1680967807769775, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.702506422996521, + "num_tokens": 415964168.0, + "step": 16086 + }, + { + "epoch": 1.7666373819459698, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.293267011642456, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.729668140411377, + "num_tokens": 415988411.0, + "step": 16087 + }, + { + "epoch": 1.7667471996485835, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.044680595397949, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.6974491477012634, + "num_tokens": 416018363.0, + "step": 16088 + }, + { + "epoch": 1.766857017351197, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1684539318084717, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7317121624946594, + "num_tokens": 416046232.0, + "step": 16089 + }, + { + "epoch": 1.7669668350538106, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1632213592529297, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7373239994049072, + "num_tokens": 416073759.0, + "step": 16090 + }, + { + "epoch": 1.7670766527564243, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2585620880126953, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7169640064239502, + "num_tokens": 416098163.0, + "step": 16091 + }, + { + "epoch": 1.767186470459038, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.106942892074585, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7155367136001587, + "num_tokens": 416126423.0, + "step": 16092 + }, + { + "epoch": 1.7672962881616516, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.579970598220825, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7381466627120972, + "num_tokens": 416147882.0, + "step": 16093 + }, + { + "epoch": 1.7674061058642652, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.426206350326538, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7195848822593689, + "num_tokens": 416169273.0, + "step": 16094 + }, + { + "epoch": 1.767515923566879, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.355512857437134, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7017191648483276, + "num_tokens": 416192765.0, + "step": 16095 + }, + { + "epoch": 1.7676257412694927, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3204829692840576, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7262189388275146, + "num_tokens": 416216952.0, + "step": 16096 + }, + { + "epoch": 1.7677355589721064, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1431777477264404, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.728476881980896, + "num_tokens": 416243681.0, + "step": 16097 + }, + { + "epoch": 1.76784537667472, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2102785110473633, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7301548719406128, + "num_tokens": 416270792.0, + "step": 16098 + }, + { + "epoch": 1.7679551943773335, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1928584575653076, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.711144208908081, + "num_tokens": 416298995.0, + "step": 16099 + }, + { + "epoch": 1.7680650120799473, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1678097248077393, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7261520624160767, + "num_tokens": 416324419.0, + "step": 16100 + }, + { + "epoch": 1.768174829782561, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4489474296569824, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7292665243148804, + "num_tokens": 416347089.0, + "step": 16101 + }, + { + "epoch": 1.7682846474851748, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3014352321624756, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6912742853164673, + "num_tokens": 416372595.0, + "step": 16102 + }, + { + "epoch": 1.7683944651877883, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4530978202819824, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7211035490036011, + "num_tokens": 416393401.0, + "step": 16103 + }, + { + "epoch": 1.7685042828904018, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.455487012863159, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7267992496490479, + "num_tokens": 416416802.0, + "step": 16104 + }, + { + "epoch": 1.7686141005930156, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2540698051452637, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7044320106506348, + "num_tokens": 416446841.0, + "step": 16105 + }, + { + "epoch": 1.7687239182956294, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4856791496276855, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7381811141967773, + "num_tokens": 416471615.0, + "step": 16106 + }, + { + "epoch": 1.768833735998243, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1373517513275146, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.708469033241272, + "num_tokens": 416500878.0, + "step": 16107 + }, + { + "epoch": 1.7689435537008564, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3593943119049072, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.73199063539505, + "num_tokens": 416524480.0, + "step": 16108 + }, + { + "epoch": 1.7690533714034702, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2407164573669434, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7317408323287964, + "num_tokens": 416549422.0, + "step": 16109 + }, + { + "epoch": 1.769163189106084, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.315577507019043, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7337627410888672, + "num_tokens": 416574267.0, + "step": 16110 + }, + { + "epoch": 1.7692730068086977, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1237454414367676, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7065262198448181, + "num_tokens": 416602495.0, + "step": 16111 + }, + { + "epoch": 1.7693828245113112, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1130638122558594, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7061752676963806, + "num_tokens": 416631999.0, + "step": 16112 + }, + { + "epoch": 1.7694926422139248, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0619211196899414, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7084552049636841, + "num_tokens": 416664289.0, + "step": 16113 + }, + { + "epoch": 1.7696024599165385, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4599180221557617, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7199488878250122, + "num_tokens": 416686554.0, + "step": 16114 + }, + { + "epoch": 1.7697122776191523, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1389269828796387, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7115297317504883, + "num_tokens": 416715570.0, + "step": 16115 + }, + { + "epoch": 1.7698220953217658, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.172553062438965, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7088540196418762, + "num_tokens": 416743489.0, + "step": 16116 + }, + { + "epoch": 1.7699319130243796, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0530989170074463, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7195868492126465, + "num_tokens": 416771487.0, + "step": 16117 + }, + { + "epoch": 1.7700417307269931, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.102375030517578, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7169314026832581, + "num_tokens": 416801416.0, + "step": 16118 + }, + { + "epoch": 1.7701515484296069, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4511733055114746, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7303668260574341, + "num_tokens": 416823188.0, + "step": 16119 + }, + { + "epoch": 1.7702613661322206, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2386739253997803, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7176657915115356, + "num_tokens": 416849104.0, + "step": 16120 + }, + { + "epoch": 1.7703711838348342, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2266342639923096, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7042261958122253, + "num_tokens": 416876435.0, + "step": 16121 + }, + { + "epoch": 1.7704810015374477, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9403643608093262, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7165192365646362, + "num_tokens": 416910832.0, + "step": 16122 + }, + { + "epoch": 1.7705908192400615, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2378907203674316, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7063479423522949, + "num_tokens": 416939844.0, + "step": 16123 + }, + { + "epoch": 1.7707006369426752, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1479828357696533, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7331453561782837, + "num_tokens": 416966003.0, + "step": 16124 + }, + { + "epoch": 1.770810454645289, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.605888843536377, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7244042158126831, + "num_tokens": 416985736.0, + "step": 16125 + }, + { + "epoch": 1.7709202723479025, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.427394390106201, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7194445133209229, + "num_tokens": 417007408.0, + "step": 16126 + }, + { + "epoch": 1.771030090050516, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.917370319366455, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7339207530021667, + "num_tokens": 417039428.0, + "step": 16127 + }, + { + "epoch": 1.7711399077531298, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2634551525115967, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7437334060668945, + "num_tokens": 417062279.0, + "step": 16128 + }, + { + "epoch": 1.7712497254557436, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2869701385498047, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7301102876663208, + "num_tokens": 417086262.0, + "step": 16129 + }, + { + "epoch": 1.771359543158357, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.139134168624878, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7062274217605591, + "num_tokens": 417113676.0, + "step": 16130 + }, + { + "epoch": 1.7714693608609708, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3610787391662598, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7233102321624756, + "num_tokens": 417137542.0, + "step": 16131 + }, + { + "epoch": 1.7715791785635844, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.391589403152466, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7194673418998718, + "num_tokens": 417160678.0, + "step": 16132 + }, + { + "epoch": 1.7716889962661981, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4680604934692383, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7112603187561035, + "num_tokens": 417183766.0, + "step": 16133 + }, + { + "epoch": 1.771798813968812, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3860347270965576, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6982483267784119, + "num_tokens": 417207715.0, + "step": 16134 + }, + { + "epoch": 1.7719086316714254, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3687477111816406, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7268044948577881, + "num_tokens": 417232832.0, + "step": 16135 + }, + { + "epoch": 1.772018449374039, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.06630277633667, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7189357280731201, + "num_tokens": 417262091.0, + "step": 16136 + }, + { + "epoch": 1.7721282670766527, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0622434616088867, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.709556519985199, + "num_tokens": 417293057.0, + "step": 16137 + }, + { + "epoch": 1.7722380847792665, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2858822345733643, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7250649929046631, + "num_tokens": 417317893.0, + "step": 16138 + }, + { + "epoch": 1.7723479024818802, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.12577223777771, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7303245663642883, + "num_tokens": 417345068.0, + "step": 16139 + }, + { + "epoch": 1.7724577201844938, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1166796684265137, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7311080098152161, + "num_tokens": 417372931.0, + "step": 16140 + }, + { + "epoch": 1.7725675378871073, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.32832670211792, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7154395580291748, + "num_tokens": 417395646.0, + "step": 16141 + }, + { + "epoch": 1.772677355589721, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1250789165496826, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7091617584228516, + "num_tokens": 417426697.0, + "step": 16142 + }, + { + "epoch": 1.7727871732923348, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3558645248413086, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7473505735397339, + "num_tokens": 417448738.0, + "step": 16143 + }, + { + "epoch": 1.7728969909949484, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.321934700012207, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7303641438484192, + "num_tokens": 417472669.0, + "step": 16144 + }, + { + "epoch": 1.7730068086975619, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1600182056427, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7273706793785095, + "num_tokens": 417500389.0, + "step": 16145 + }, + { + "epoch": 1.7731166264001756, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1316349506378174, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.7047386169433594, + "num_tokens": 417529568.0, + "step": 16146 + }, + { + "epoch": 1.7732264441027894, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3811004161834717, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7208671569824219, + "num_tokens": 417554658.0, + "step": 16147 + }, + { + "epoch": 1.7733362618054032, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.98483145236969, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7127541303634644, + "num_tokens": 417587294.0, + "step": 16148 + }, + { + "epoch": 1.7734460795080167, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2761082649230957, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7217519283294678, + "num_tokens": 417612958.0, + "step": 16149 + }, + { + "epoch": 1.7735558972106302, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.5033867359161377, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7294903993606567, + "num_tokens": 417634148.0, + "step": 16150 + }, + { + "epoch": 1.773665714913244, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.213982582092285, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7122068405151367, + "num_tokens": 417659456.0, + "step": 16151 + }, + { + "epoch": 1.7737755326158577, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.131540536880493, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7038364410400391, + "num_tokens": 417687557.0, + "step": 16152 + }, + { + "epoch": 1.7738853503184715, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2712700366973877, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7125410437583923, + "num_tokens": 417712667.0, + "step": 16153 + }, + { + "epoch": 1.773995168021085, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3118321895599365, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7406514286994934, + "num_tokens": 417735849.0, + "step": 16154 + }, + { + "epoch": 1.7741049857236986, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.094348430633545, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7009444832801819, + "num_tokens": 417765359.0, + "step": 16155 + }, + { + "epoch": 1.7742148034263123, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9511843919754028, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7026899456977844, + "num_tokens": 417797089.0, + "step": 16156 + }, + { + "epoch": 1.774324621128926, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.316324472427368, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7234342098236084, + "num_tokens": 417820062.0, + "step": 16157 + }, + { + "epoch": 1.7744344388315396, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1813976764678955, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6983178853988647, + "num_tokens": 417848752.0, + "step": 16158 + }, + { + "epoch": 1.7745442565341532, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.991220474243164, + "learning_rate": 1e-06, + "loss": 1.0399, + "mean_token_accuracy": 0.6859275102615356, + "num_tokens": 417881739.0, + "step": 16159 + }, + { + "epoch": 1.774654074236767, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0719215869903564, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7181563973426819, + "num_tokens": 417913173.0, + "step": 16160 + }, + { + "epoch": 1.7747638919393807, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.307507276535034, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7189728021621704, + "num_tokens": 417939617.0, + "step": 16161 + }, + { + "epoch": 1.7748737096419944, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2391278743743896, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.724912166595459, + "num_tokens": 417963686.0, + "step": 16162 + }, + { + "epoch": 1.774983527344608, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.125349283218384, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7071735858917236, + "num_tokens": 417992361.0, + "step": 16163 + }, + { + "epoch": 1.7750933450472215, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2068228721618652, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7432072162628174, + "num_tokens": 418017981.0, + "step": 16164 + }, + { + "epoch": 1.7752031627498353, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1995506286621094, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7372653484344482, + "num_tokens": 418043341.0, + "step": 16165 + }, + { + "epoch": 1.775312980452449, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2340118885040283, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.696401834487915, + "num_tokens": 418069128.0, + "step": 16166 + }, + { + "epoch": 1.7754227981550625, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.477365255355835, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.725382924079895, + "num_tokens": 418090271.0, + "step": 16167 + }, + { + "epoch": 1.7755326158576763, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3934242725372314, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7344782948493958, + "num_tokens": 418113119.0, + "step": 16168 + }, + { + "epoch": 1.7756424335602898, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2152254581451416, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7229967713356018, + "num_tokens": 418138183.0, + "step": 16169 + }, + { + "epoch": 1.7757522512629036, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.021252155303955, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6961995363235474, + "num_tokens": 418171349.0, + "step": 16170 + }, + { + "epoch": 1.7758620689655173, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9206700325012207, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7216031551361084, + "num_tokens": 418201351.0, + "step": 16171 + }, + { + "epoch": 1.7759718866681309, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2990643978118896, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7367256879806519, + "num_tokens": 418226757.0, + "step": 16172 + }, + { + "epoch": 1.7760817043707444, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.360244035720825, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7249794602394104, + "num_tokens": 418250369.0, + "step": 16173 + }, + { + "epoch": 1.7761915220733582, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.50364351272583, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7517580389976501, + "num_tokens": 418271077.0, + "step": 16174 + }, + { + "epoch": 1.776301339775972, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.417285442352295, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7344392538070679, + "num_tokens": 418292164.0, + "step": 16175 + }, + { + "epoch": 1.7764111574785857, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.443345546722412, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7276238203048706, + "num_tokens": 418315319.0, + "step": 16176 + }, + { + "epoch": 1.7765209751811992, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4008681774139404, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7033733129501343, + "num_tokens": 418339650.0, + "step": 16177 + }, + { + "epoch": 1.7766307928838128, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9729902744293213, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7185803055763245, + "num_tokens": 418370376.0, + "step": 16178 + }, + { + "epoch": 1.7767406105864265, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.319240093231201, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7002751231193542, + "num_tokens": 418395633.0, + "step": 16179 + }, + { + "epoch": 1.7768504282890403, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4215269088745117, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7237781882286072, + "num_tokens": 418417308.0, + "step": 16180 + }, + { + "epoch": 1.7769602459916538, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3026673793792725, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7239517569541931, + "num_tokens": 418441138.0, + "step": 16181 + }, + { + "epoch": 1.7770700636942676, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0941412448883057, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.7011204957962036, + "num_tokens": 418470919.0, + "step": 16182 + }, + { + "epoch": 1.777179881396881, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.550950050354004, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7164857387542725, + "num_tokens": 418493758.0, + "step": 16183 + }, + { + "epoch": 1.7772896990994949, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3255555629730225, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7035220265388489, + "num_tokens": 418520258.0, + "step": 16184 + }, + { + "epoch": 1.7773995168021086, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.289106607437134, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7379432916641235, + "num_tokens": 418544191.0, + "step": 16185 + }, + { + "epoch": 1.7775093345047221, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.515760898590088, + "learning_rate": 1e-06, + "loss": 0.748, + "mean_token_accuracy": 0.7613600492477417, + "num_tokens": 418564807.0, + "step": 16186 + }, + { + "epoch": 1.7776191522073357, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1414105892181396, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7388867735862732, + "num_tokens": 418590708.0, + "step": 16187 + }, + { + "epoch": 1.7777289699099494, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2705771923065186, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7098439335823059, + "num_tokens": 418617068.0, + "step": 16188 + }, + { + "epoch": 1.7778387876125632, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2483129501342773, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7230533957481384, + "num_tokens": 418640811.0, + "step": 16189 + }, + { + "epoch": 1.777948605315177, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.143663167953491, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7198171615600586, + "num_tokens": 418667841.0, + "step": 16190 + }, + { + "epoch": 1.7780584230177905, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.079711675643921, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7175478935241699, + "num_tokens": 418696645.0, + "step": 16191 + }, + { + "epoch": 1.778168240720404, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0395150184631348, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7288168668746948, + "num_tokens": 418726463.0, + "step": 16192 + }, + { + "epoch": 1.7782780584230178, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0927155017852783, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.711394190788269, + "num_tokens": 418755149.0, + "step": 16193 + }, + { + "epoch": 1.7783878761256315, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3833541870117188, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.727717399597168, + "num_tokens": 418778806.0, + "step": 16194 + }, + { + "epoch": 1.778497693828245, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4309942722320557, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7379204034805298, + "num_tokens": 418799077.0, + "step": 16195 + }, + { + "epoch": 1.7786075115308586, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3341548442840576, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7338088750839233, + "num_tokens": 418823886.0, + "step": 16196 + }, + { + "epoch": 1.7787173292334724, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.20548152923584, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7141210436820984, + "num_tokens": 418850668.0, + "step": 16197 + }, + { + "epoch": 1.7788271469360861, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.164416790008545, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7229427099227905, + "num_tokens": 418876540.0, + "step": 16198 + }, + { + "epoch": 1.7789369646386999, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.700650215148926, + "learning_rate": 1e-06, + "loss": 0.7535, + "mean_token_accuracy": 0.7560555934906006, + "num_tokens": 418894376.0, + "step": 16199 + }, + { + "epoch": 1.7790467823413134, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2373151779174805, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.705607533454895, + "num_tokens": 418919237.0, + "step": 16200 + }, + { + "epoch": 1.779156600043927, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9319407939910889, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6864737272262573, + "num_tokens": 418954808.0, + "step": 16201 + }, + { + "epoch": 1.7792664177465407, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.414646863937378, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7119474411010742, + "num_tokens": 418979222.0, + "step": 16202 + }, + { + "epoch": 1.7793762354491545, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2591235637664795, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7327832579612732, + "num_tokens": 419005298.0, + "step": 16203 + }, + { + "epoch": 1.7794860531517682, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.360041856765747, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7131252288818359, + "num_tokens": 419031203.0, + "step": 16204 + }, + { + "epoch": 1.7795958708543818, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2572898864746094, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7372678518295288, + "num_tokens": 419056050.0, + "step": 16205 + }, + { + "epoch": 1.7797056885569953, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2318618297576904, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7113747596740723, + "num_tokens": 419085664.0, + "step": 16206 + }, + { + "epoch": 1.779815506259609, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.5480124950408936, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.71905916929245, + "num_tokens": 419107035.0, + "step": 16207 + }, + { + "epoch": 1.7799253239622228, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.12707781791687, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7200407981872559, + "num_tokens": 419134962.0, + "step": 16208 + }, + { + "epoch": 1.7800351416648363, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.505262851715088, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7219408750534058, + "num_tokens": 419156469.0, + "step": 16209 + }, + { + "epoch": 1.7801449593674499, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4140987396240234, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7115414142608643, + "num_tokens": 419180412.0, + "step": 16210 + }, + { + "epoch": 1.7802547770700636, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2060446739196777, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7162928581237793, + "num_tokens": 419209102.0, + "step": 16211 + }, + { + "epoch": 1.7803645947726774, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.530125141143799, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7321038246154785, + "num_tokens": 419230360.0, + "step": 16212 + }, + { + "epoch": 1.7804744124752911, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9927780628204346, + "learning_rate": 1e-06, + "loss": 1.0635, + "mean_token_accuracy": 0.6774455308914185, + "num_tokens": 419263585.0, + "step": 16213 + }, + { + "epoch": 1.7805842301779047, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.318991184234619, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7111389636993408, + "num_tokens": 419290289.0, + "step": 16214 + }, + { + "epoch": 1.7806940478805182, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1830830574035645, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7068834900856018, + "num_tokens": 419315191.0, + "step": 16215 + }, + { + "epoch": 1.780803865583132, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.02976131439209, + "learning_rate": 1e-06, + "loss": 1.0375, + "mean_token_accuracy": 0.682903528213501, + "num_tokens": 419345979.0, + "step": 16216 + }, + { + "epoch": 1.7809136832857457, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3324382305145264, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7450516223907471, + "num_tokens": 419369204.0, + "step": 16217 + }, + { + "epoch": 1.7810235009883595, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.332275867462158, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7408437132835388, + "num_tokens": 419393122.0, + "step": 16218 + }, + { + "epoch": 1.781133318690973, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.177234649658203, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7116692066192627, + "num_tokens": 419420255.0, + "step": 16219 + }, + { + "epoch": 1.7812431363935866, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2703113555908203, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7181669473648071, + "num_tokens": 419445421.0, + "step": 16220 + }, + { + "epoch": 1.7813529540962003, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1808581352233887, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6988126039505005, + "num_tokens": 419473130.0, + "step": 16221 + }, + { + "epoch": 1.781462771798814, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4045591354370117, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7137304544448853, + "num_tokens": 419494795.0, + "step": 16222 + }, + { + "epoch": 1.7815725895014276, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.436112880706787, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7089212536811829, + "num_tokens": 419519442.0, + "step": 16223 + }, + { + "epoch": 1.7816824072040411, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3293089866638184, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7108886241912842, + "num_tokens": 419543835.0, + "step": 16224 + }, + { + "epoch": 1.781792224906655, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.582967519760132, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7192899584770203, + "num_tokens": 419565539.0, + "step": 16225 + }, + { + "epoch": 1.7819020426092687, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2943201065063477, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7224141955375671, + "num_tokens": 419589785.0, + "step": 16226 + }, + { + "epoch": 1.7820118603118824, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1382017135620117, + "learning_rate": 1e-06, + "loss": 0.7797, + "mean_token_accuracy": 0.7543767690658569, + "num_tokens": 419614859.0, + "step": 16227 + }, + { + "epoch": 1.782121678014496, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.125779628753662, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7326076626777649, + "num_tokens": 419645316.0, + "step": 16228 + }, + { + "epoch": 1.7822314957171095, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.330444812774658, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.733911395072937, + "num_tokens": 419669390.0, + "step": 16229 + }, + { + "epoch": 1.7823413134197232, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4590890407562256, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7044228315353394, + "num_tokens": 419692156.0, + "step": 16230 + }, + { + "epoch": 1.782451131122337, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.209245443344116, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7231295108795166, + "num_tokens": 419718398.0, + "step": 16231 + }, + { + "epoch": 1.7825609488249505, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1908509731292725, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7250542640686035, + "num_tokens": 419743950.0, + "step": 16232 + }, + { + "epoch": 1.7826707665275643, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.503570795059204, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7418498992919922, + "num_tokens": 419764640.0, + "step": 16233 + }, + { + "epoch": 1.7827805842301778, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1900179386138916, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7222286462783813, + "num_tokens": 419792119.0, + "step": 16234 + }, + { + "epoch": 1.7828904019327916, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0850093364715576, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7247662544250488, + "num_tokens": 419820532.0, + "step": 16235 + }, + { + "epoch": 1.7830002196354053, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.207744598388672, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7433351874351501, + "num_tokens": 419845548.0, + "step": 16236 + }, + { + "epoch": 1.7831100373380189, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3951401710510254, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7233200073242188, + "num_tokens": 419867079.0, + "step": 16237 + }, + { + "epoch": 1.7832198550406324, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3486855030059814, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7257524728775024, + "num_tokens": 419891368.0, + "step": 16238 + }, + { + "epoch": 1.7833296727432462, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.3844523429870605, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7131801843643188, + "num_tokens": 419915401.0, + "step": 16239 + }, + { + "epoch": 1.78343949044586, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.168220043182373, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7196118831634521, + "num_tokens": 419941102.0, + "step": 16240 + }, + { + "epoch": 1.7835493081484737, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.700697898864746, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7281784415245056, + "num_tokens": 419960710.0, + "step": 16241 + }, + { + "epoch": 1.7836591258510872, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.573012351989746, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7311064004898071, + "num_tokens": 419979796.0, + "step": 16242 + }, + { + "epoch": 1.7837689435537007, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.6148576736450195, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7263245582580566, + "num_tokens": 419999905.0, + "step": 16243 + }, + { + "epoch": 1.7838787612563145, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.049694538116455, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7128398418426514, + "num_tokens": 420029631.0, + "step": 16244 + }, + { + "epoch": 1.7839885789589283, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4380106925964355, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7291996479034424, + "num_tokens": 420051474.0, + "step": 16245 + }, + { + "epoch": 1.7840983966615418, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.5419838428497314, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7257140278816223, + "num_tokens": 420071650.0, + "step": 16246 + }, + { + "epoch": 1.7842082143641556, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.491305112838745, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7194993495941162, + "num_tokens": 420094751.0, + "step": 16247 + }, + { + "epoch": 1.784318032066769, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2830495834350586, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7245610356330872, + "num_tokens": 420121656.0, + "step": 16248 + }, + { + "epoch": 1.7844278497693828, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.450997829437256, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7345364689826965, + "num_tokens": 420146275.0, + "step": 16249 + }, + { + "epoch": 1.7845376674719966, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.235478162765503, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7048869729042053, + "num_tokens": 420173068.0, + "step": 16250 + }, + { + "epoch": 1.7846474851746101, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0859146118164062, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7140454053878784, + "num_tokens": 420202313.0, + "step": 16251 + }, + { + "epoch": 1.7847573028772237, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.4165594577789307, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7488303184509277, + "num_tokens": 420225258.0, + "step": 16252 + }, + { + "epoch": 1.7848671205798374, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.0924413204193115, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.719537615776062, + "num_tokens": 420254837.0, + "step": 16253 + }, + { + "epoch": 1.7849769382824512, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.053917169570923, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7032259106636047, + "num_tokens": 420284882.0, + "step": 16254 + }, + { + "epoch": 1.785086755985065, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2705984115600586, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7164232730865479, + "num_tokens": 420312526.0, + "step": 16255 + }, + { + "epoch": 1.7851965736876785, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2087841033935547, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7113362550735474, + "num_tokens": 420341823.0, + "step": 16256 + }, + { + "epoch": 1.785306391390292, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.9106045961380005, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7243208885192871, + "num_tokens": 420372663.0, + "step": 16257 + }, + { + "epoch": 1.7854162090929058, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2766430377960205, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7125498056411743, + "num_tokens": 420396983.0, + "step": 16258 + }, + { + "epoch": 1.7855260267955195, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.236071825027466, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7219442129135132, + "num_tokens": 420421503.0, + "step": 16259 + }, + { + "epoch": 1.785635844498133, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.180485725402832, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7006121277809143, + "num_tokens": 420451332.0, + "step": 16260 + }, + { + "epoch": 1.7857456622007466, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1044318675994873, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7144061326980591, + "num_tokens": 420481896.0, + "step": 16261 + }, + { + "epoch": 1.7858554799033604, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.2530431747436523, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7285460233688354, + "num_tokens": 420507325.0, + "step": 16262 + }, + { + "epoch": 1.7859652976059741, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1732115745544434, + "learning_rate": 1e-06, + "loss": 0.773, + "mean_token_accuracy": 0.7544934749603271, + "num_tokens": 420532423.0, + "step": 16263 + }, + { + "epoch": 1.7860751153085879, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.161522150039673, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7195858955383301, + "num_tokens": 420558188.0, + "step": 16264 + }, + { + "epoch": 1.7861849330112014, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1841533184051514, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7291423678398132, + "num_tokens": 420583440.0, + "step": 16265 + }, + { + "epoch": 1.786294750713815, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1791059970855713, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7299013733863831, + "num_tokens": 420608649.0, + "step": 16266 + }, + { + "epoch": 1.7864045684164287, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.743104934692383, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7272255420684814, + "num_tokens": 420627120.0, + "step": 16267 + }, + { + "epoch": 1.7865143861190425, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.081360340118408, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7090511918067932, + "num_tokens": 420656474.0, + "step": 16268 + }, + { + "epoch": 1.7866242038216562, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1497552394866943, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7311104536056519, + "num_tokens": 420683699.0, + "step": 16269 + }, + { + "epoch": 1.7867340215242697, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.1839001178741455, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7115656137466431, + "num_tokens": 420708766.0, + "step": 16270 + }, + { + "epoch": 1.7868438392268833, + "ewc_loss": 0.0001373291015625, + "grad_norm": 1.950548529624939, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.6976776123046875, + "num_tokens": 420741260.0, + "step": 16271 + }, + { + "epoch": 1.786953656929497, + "ewc_loss": 0.0001373291015625, + "grad_norm": 2.163572072982788, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7171170711517334, + "num_tokens": 420768544.0, + "step": 16272 + }, + { + "epoch": 1.7870634746321108, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.368096113204956, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7439831495285034, + "num_tokens": 420790272.0, + "step": 16273 + }, + { + "epoch": 1.7871732923347243, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.071521043777466, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7094956040382385, + "num_tokens": 420818546.0, + "step": 16274 + }, + { + "epoch": 1.7872831100373379, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0739848613739014, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6989213228225708, + "num_tokens": 420848155.0, + "step": 16275 + }, + { + "epoch": 1.7873929277399516, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.180622100830078, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7269313335418701, + "num_tokens": 420873991.0, + "step": 16276 + }, + { + "epoch": 1.7875027454425654, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.263113021850586, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7220280766487122, + "num_tokens": 420899750.0, + "step": 16277 + }, + { + "epoch": 1.7876125631451791, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.29504132270813, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7162420749664307, + "num_tokens": 420926433.0, + "step": 16278 + }, + { + "epoch": 1.7877223808477927, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1113409996032715, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.731520414352417, + "num_tokens": 420951268.0, + "step": 16279 + }, + { + "epoch": 1.7878321985504062, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1093106269836426, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7145867347717285, + "num_tokens": 420981931.0, + "step": 16280 + }, + { + "epoch": 1.78794201625302, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.44059419631958, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.6904441118240356, + "num_tokens": 421007442.0, + "step": 16281 + }, + { + "epoch": 1.7880518339556337, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1841423511505127, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.710671067237854, + "num_tokens": 421034920.0, + "step": 16282 + }, + { + "epoch": 1.7881616516582475, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9813965559005737, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.7039598226547241, + "num_tokens": 421066890.0, + "step": 16283 + }, + { + "epoch": 1.788271469360861, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3481743335723877, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.6992923021316528, + "num_tokens": 421090461.0, + "step": 16284 + }, + { + "epoch": 1.7883812870634745, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4413042068481445, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7224366068840027, + "num_tokens": 421112929.0, + "step": 16285 + }, + { + "epoch": 1.7884911047660883, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.267073392868042, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7269543409347534, + "num_tokens": 421137933.0, + "step": 16286 + }, + { + "epoch": 1.788600922468702, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.192340850830078, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7209407091140747, + "num_tokens": 421165253.0, + "step": 16287 + }, + { + "epoch": 1.7887107401713156, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2784018516540527, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7490348219871521, + "num_tokens": 421187552.0, + "step": 16288 + }, + { + "epoch": 1.7888205578739291, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.222405433654785, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7292441129684448, + "num_tokens": 421213194.0, + "step": 16289 + }, + { + "epoch": 1.7889303755765429, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.294696092605591, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7172950506210327, + "num_tokens": 421236247.0, + "step": 16290 + }, + { + "epoch": 1.7890401932791566, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.910839557647705, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7244837284088135, + "num_tokens": 421257100.0, + "step": 16291 + }, + { + "epoch": 1.7891500109817704, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2047030925750732, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7327929139137268, + "num_tokens": 421280175.0, + "step": 16292 + }, + { + "epoch": 1.789259828684384, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.476369619369507, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7498075366020203, + "num_tokens": 421300730.0, + "step": 16293 + }, + { + "epoch": 1.7893696463869975, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.330631732940674, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7415431141853333, + "num_tokens": 421323715.0, + "step": 16294 + }, + { + "epoch": 1.7894794640896112, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4747934341430664, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7095159888267517, + "num_tokens": 421346426.0, + "step": 16295 + }, + { + "epoch": 1.789589281792225, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2993223667144775, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.692220151424408, + "num_tokens": 421371371.0, + "step": 16296 + }, + { + "epoch": 1.7896990994948385, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3183374404907227, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.7034989595413208, + "num_tokens": 421396099.0, + "step": 16297 + }, + { + "epoch": 1.7898089171974523, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1930606365203857, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7214913368225098, + "num_tokens": 421424550.0, + "step": 16298 + }, + { + "epoch": 1.7899187349000658, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3795928955078125, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7401025295257568, + "num_tokens": 421446253.0, + "step": 16299 + }, + { + "epoch": 1.7900285526026796, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9292881488800049, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7128524780273438, + "num_tokens": 421480936.0, + "step": 16300 + }, + { + "epoch": 1.7901383703052933, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.424489974975586, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7212080955505371, + "num_tokens": 421501546.0, + "step": 16301 + }, + { + "epoch": 1.7902481880079069, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4544148445129395, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7159442901611328, + "num_tokens": 421523198.0, + "step": 16302 + }, + { + "epoch": 1.7903580057105204, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2043402194976807, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7191564440727234, + "num_tokens": 421549623.0, + "step": 16303 + }, + { + "epoch": 1.7904678234131342, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2791805267333984, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.6976205110549927, + "num_tokens": 421574848.0, + "step": 16304 + }, + { + "epoch": 1.790577641115748, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.381925582885742, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7224907875061035, + "num_tokens": 421596570.0, + "step": 16305 + }, + { + "epoch": 1.7906874588183617, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0628674030303955, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7124729156494141, + "num_tokens": 421626890.0, + "step": 16306 + }, + { + "epoch": 1.7907972765209752, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.288132667541504, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7230055928230286, + "num_tokens": 421652410.0, + "step": 16307 + }, + { + "epoch": 1.7909070942235887, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.270933151245117, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.710705041885376, + "num_tokens": 421677817.0, + "step": 16308 + }, + { + "epoch": 1.7910169119262025, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.293351411819458, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7316429615020752, + "num_tokens": 421701895.0, + "step": 16309 + }, + { + "epoch": 1.7911267296288162, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.181425094604492, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7224024534225464, + "num_tokens": 421727887.0, + "step": 16310 + }, + { + "epoch": 1.7912365473314298, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.195993661880493, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7208674550056458, + "num_tokens": 421754232.0, + "step": 16311 + }, + { + "epoch": 1.7913463650340435, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.632721424102783, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7135539054870605, + "num_tokens": 421774594.0, + "step": 16312 + }, + { + "epoch": 1.791456182736657, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3714945316314697, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7045258283615112, + "num_tokens": 421799392.0, + "step": 16313 + }, + { + "epoch": 1.7915660004392708, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.248471975326538, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7041218280792236, + "num_tokens": 421827295.0, + "step": 16314 + }, + { + "epoch": 1.7916758181418846, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.262648105621338, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.71889328956604, + "num_tokens": 421853525.0, + "step": 16315 + }, + { + "epoch": 1.7917856358444981, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1841068267822266, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7348067760467529, + "num_tokens": 421880922.0, + "step": 16316 + }, + { + "epoch": 1.7918954535471117, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3202064037323, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7320734262466431, + "num_tokens": 421905972.0, + "step": 16317 + }, + { + "epoch": 1.7920052712497254, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.374119520187378, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.715216875076294, + "num_tokens": 421928724.0, + "step": 16318 + }, + { + "epoch": 1.7921150889523392, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0968828201293945, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7139655947685242, + "num_tokens": 421956659.0, + "step": 16319 + }, + { + "epoch": 1.792224906654953, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.312497854232788, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7101596593856812, + "num_tokens": 421979564.0, + "step": 16320 + }, + { + "epoch": 1.7923347243575665, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3279430866241455, + "learning_rate": 1e-06, + "loss": 0.7735, + "mean_token_accuracy": 0.7554895281791687, + "num_tokens": 422000886.0, + "step": 16321 + }, + { + "epoch": 1.79244454206018, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2023277282714844, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7067766189575195, + "num_tokens": 422027894.0, + "step": 16322 + }, + { + "epoch": 1.7925543597627938, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.341987371444702, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7179213762283325, + "num_tokens": 422050366.0, + "step": 16323 + }, + { + "epoch": 1.7926641774654075, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0447371006011963, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.722396969795227, + "num_tokens": 422078367.0, + "step": 16324 + }, + { + "epoch": 1.792773995168021, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.418550729751587, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7319357991218567, + "num_tokens": 422102979.0, + "step": 16325 + }, + { + "epoch": 1.7928838128706346, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.335601568222046, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7242417335510254, + "num_tokens": 422127190.0, + "step": 16326 + }, + { + "epoch": 1.7929936305732483, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3335795402526855, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7064025402069092, + "num_tokens": 422151054.0, + "step": 16327 + }, + { + "epoch": 1.793103448275862, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.198833703994751, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7394618988037109, + "num_tokens": 422176163.0, + "step": 16328 + }, + { + "epoch": 1.7932132659784759, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.203883409500122, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7174190282821655, + "num_tokens": 422200960.0, + "step": 16329 + }, + { + "epoch": 1.7933230836810894, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9742006063461304, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.730303943157196, + "num_tokens": 422229630.0, + "step": 16330 + }, + { + "epoch": 1.793432901383703, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0683157444000244, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7101555466651917, + "num_tokens": 422259552.0, + "step": 16331 + }, + { + "epoch": 1.7935427190863167, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.144378423690796, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7146871089935303, + "num_tokens": 422289751.0, + "step": 16332 + }, + { + "epoch": 1.7936525367889304, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3257298469543457, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7285740375518799, + "num_tokens": 422314633.0, + "step": 16333 + }, + { + "epoch": 1.7937623544915442, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0932111740112305, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7238138914108276, + "num_tokens": 422341077.0, + "step": 16334 + }, + { + "epoch": 1.7938721721941577, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.248777151107788, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6950145959854126, + "num_tokens": 422369625.0, + "step": 16335 + }, + { + "epoch": 1.7939819898967713, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.337341547012329, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7236948609352112, + "num_tokens": 422394613.0, + "step": 16336 + }, + { + "epoch": 1.794091807599385, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4238224029541016, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7241506576538086, + "num_tokens": 422417982.0, + "step": 16337 + }, + { + "epoch": 1.7942016253019988, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.208523750305176, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7352536916732788, + "num_tokens": 422443533.0, + "step": 16338 + }, + { + "epoch": 1.7943114430046123, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3231124877929688, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7087857723236084, + "num_tokens": 422468024.0, + "step": 16339 + }, + { + "epoch": 1.7944212607072259, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4497053623199463, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7119422554969788, + "num_tokens": 422490614.0, + "step": 16340 + }, + { + "epoch": 1.7945310784098396, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4858272075653076, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7292625904083252, + "num_tokens": 422512313.0, + "step": 16341 + }, + { + "epoch": 1.7946408961124534, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.150800943374634, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.729210376739502, + "num_tokens": 422537722.0, + "step": 16342 + }, + { + "epoch": 1.7947507138150671, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9612997770309448, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7246637940406799, + "num_tokens": 422567382.0, + "step": 16343 + }, + { + "epoch": 1.7948605315176807, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1433780193328857, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7040678858757019, + "num_tokens": 422595383.0, + "step": 16344 + }, + { + "epoch": 1.7949703492202942, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2111144065856934, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7181788086891174, + "num_tokens": 422623755.0, + "step": 16345 + }, + { + "epoch": 1.795080166922908, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2314705848693848, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7115248441696167, + "num_tokens": 422649779.0, + "step": 16346 + }, + { + "epoch": 1.7951899846255217, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4188671112060547, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7308990955352783, + "num_tokens": 422671435.0, + "step": 16347 + }, + { + "epoch": 1.7952998023281352, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1737802028656006, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.740272045135498, + "num_tokens": 422696878.0, + "step": 16348 + }, + { + "epoch": 1.795409620030749, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.310403823852539, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7433469891548157, + "num_tokens": 422720433.0, + "step": 16349 + }, + { + "epoch": 1.7955194377333625, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4511332511901855, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7013125419616699, + "num_tokens": 422743653.0, + "step": 16350 + }, + { + "epoch": 1.7956292554359763, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2115139961242676, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7144834995269775, + "num_tokens": 422771067.0, + "step": 16351 + }, + { + "epoch": 1.79573907313859, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1648807525634766, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.6979584693908691, + "num_tokens": 422798121.0, + "step": 16352 + }, + { + "epoch": 1.7958488908412036, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4504802227020264, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7147974967956543, + "num_tokens": 422821499.0, + "step": 16353 + }, + { + "epoch": 1.7959587085438171, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3278582096099854, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7333581447601318, + "num_tokens": 422843980.0, + "step": 16354 + }, + { + "epoch": 1.7960685262464309, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3633627891540527, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7246044278144836, + "num_tokens": 422866031.0, + "step": 16355 + }, + { + "epoch": 1.7961783439490446, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9651211500167847, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7205947637557983, + "num_tokens": 422900563.0, + "step": 16356 + }, + { + "epoch": 1.7962881616516584, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.195561647415161, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7147831916809082, + "num_tokens": 422926656.0, + "step": 16357 + }, + { + "epoch": 1.796397979354272, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.240417003631592, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7234320044517517, + "num_tokens": 422952485.0, + "step": 16358 + }, + { + "epoch": 1.7965077970568855, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0095367431640625, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.698107123374939, + "num_tokens": 422984735.0, + "step": 16359 + }, + { + "epoch": 1.7966176147594992, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2701587677001953, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7324190735816956, + "num_tokens": 423008382.0, + "step": 16360 + }, + { + "epoch": 1.796727432462113, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.580626964569092, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7368853092193604, + "num_tokens": 423028419.0, + "step": 16361 + }, + { + "epoch": 1.7968372501647265, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2756505012512207, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7219198942184448, + "num_tokens": 423053152.0, + "step": 16362 + }, + { + "epoch": 1.7969470678673403, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.162421464920044, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7360246181488037, + "num_tokens": 423079331.0, + "step": 16363 + }, + { + "epoch": 1.7970568855699538, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4430837631225586, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7148107886314392, + "num_tokens": 423102215.0, + "step": 16364 + }, + { + "epoch": 1.7971667032725676, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9946943521499634, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7483588457107544, + "num_tokens": 423131041.0, + "step": 16365 + }, + { + "epoch": 1.7972765209751813, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.151496410369873, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7288954854011536, + "num_tokens": 423158313.0, + "step": 16366 + }, + { + "epoch": 1.7973863386777948, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.628767490386963, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7226752042770386, + "num_tokens": 423177086.0, + "step": 16367 + }, + { + "epoch": 1.7974961563804084, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.228543519973755, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7205097079277039, + "num_tokens": 423202341.0, + "step": 16368 + }, + { + "epoch": 1.7976059740830221, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.012936592102051, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.702935516834259, + "num_tokens": 423232975.0, + "step": 16369 + }, + { + "epoch": 1.797715791785636, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4588255882263184, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7125263214111328, + "num_tokens": 423255271.0, + "step": 16370 + }, + { + "epoch": 1.7978256094882497, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2904486656188965, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7092052698135376, + "num_tokens": 423279840.0, + "step": 16371 + }, + { + "epoch": 1.7979354271908632, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.393378734588623, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7327927350997925, + "num_tokens": 423301839.0, + "step": 16372 + }, + { + "epoch": 1.7980452448934767, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.291165590286255, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7093600034713745, + "num_tokens": 423325332.0, + "step": 16373 + }, + { + "epoch": 1.7981550625960905, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0789802074432373, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7304368019104004, + "num_tokens": 423354366.0, + "step": 16374 + }, + { + "epoch": 1.7982648802987042, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3527445793151855, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7053841352462769, + "num_tokens": 423378460.0, + "step": 16375 + }, + { + "epoch": 1.7983746980013178, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.066577196121216, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.7064120173454285, + "num_tokens": 423409450.0, + "step": 16376 + }, + { + "epoch": 1.7984845157039313, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3497536182403564, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.720687747001648, + "num_tokens": 423432078.0, + "step": 16377 + }, + { + "epoch": 1.798594333406545, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9968791007995605, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.6976020336151123, + "num_tokens": 423463737.0, + "step": 16378 + }, + { + "epoch": 1.7987041511091588, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.202239990234375, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7364437580108643, + "num_tokens": 423490202.0, + "step": 16379 + }, + { + "epoch": 1.7988139688117726, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.083800792694092, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7277054786682129, + "num_tokens": 423518171.0, + "step": 16380 + }, + { + "epoch": 1.7989237865143861, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0167295932769775, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7040389776229858, + "num_tokens": 423549214.0, + "step": 16381 + }, + { + "epoch": 1.7990336042169996, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9796537160873413, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.6921924352645874, + "num_tokens": 423583822.0, + "step": 16382 + }, + { + "epoch": 1.7991434219196134, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0656445026397705, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7178524732589722, + "num_tokens": 423611450.0, + "step": 16383 + }, + { + "epoch": 1.7992532396222272, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.148685932159424, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.727053165435791, + "num_tokens": 423639246.0, + "step": 16384 + }, + { + "epoch": 1.799363057324841, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3605008125305176, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7167537212371826, + "num_tokens": 423663237.0, + "step": 16385 + }, + { + "epoch": 1.7994728750274545, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.9148247241973877, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7194613814353943, + "num_tokens": 423682575.0, + "step": 16386 + }, + { + "epoch": 1.799582692730068, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2019569873809814, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7123847007751465, + "num_tokens": 423709095.0, + "step": 16387 + }, + { + "epoch": 1.7996925104326817, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.5376574993133545, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7436631321907043, + "num_tokens": 423729133.0, + "step": 16388 + }, + { + "epoch": 1.7998023281352955, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.062269687652588, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7366674542427063, + "num_tokens": 423756251.0, + "step": 16389 + }, + { + "epoch": 1.799912145837909, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3105223178863525, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7050365805625916, + "num_tokens": 423780483.0, + "step": 16390 + }, + { + "epoch": 1.8000219635405226, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.128141164779663, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7208913564682007, + "num_tokens": 423807886.0, + "step": 16391 + }, + { + "epoch": 1.8001317812431363, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3956120014190674, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7360367774963379, + "num_tokens": 423830403.0, + "step": 16392 + }, + { + "epoch": 1.80024159894575, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.005488872528076, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7035007476806641, + "num_tokens": 423862554.0, + "step": 16393 + }, + { + "epoch": 1.8003514166483638, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2458436489105225, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7301639318466187, + "num_tokens": 423889404.0, + "step": 16394 + }, + { + "epoch": 1.8004612343509774, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9899804592132568, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7137499451637268, + "num_tokens": 423920984.0, + "step": 16395 + }, + { + "epoch": 1.800571052053591, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4638559818267822, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7037312984466553, + "num_tokens": 423944673.0, + "step": 16396 + }, + { + "epoch": 1.8006808697562047, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0934696197509766, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7110239267349243, + "num_tokens": 423971785.0, + "step": 16397 + }, + { + "epoch": 1.8007906874588184, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.565676212310791, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7360575199127197, + "num_tokens": 423993324.0, + "step": 16398 + }, + { + "epoch": 1.8009005051614322, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.093705654144287, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7212334871292114, + "num_tokens": 424022594.0, + "step": 16399 + }, + { + "epoch": 1.8010103228640457, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1406126022338867, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7126639485359192, + "num_tokens": 424050715.0, + "step": 16400 + }, + { + "epoch": 1.8011201405666593, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3734114170074463, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.710154116153717, + "num_tokens": 424075604.0, + "step": 16401 + }, + { + "epoch": 1.801229958269273, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.758174180984497, + "learning_rate": 1e-06, + "loss": 0.7675, + "mean_token_accuracy": 0.7523652911186218, + "num_tokens": 424093160.0, + "step": 16402 + }, + { + "epoch": 1.8013397759718868, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.100447416305542, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.6992714405059814, + "num_tokens": 424121916.0, + "step": 16403 + }, + { + "epoch": 1.8014495936745003, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.157763957977295, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7262852191925049, + "num_tokens": 424146915.0, + "step": 16404 + }, + { + "epoch": 1.8015594113771138, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2295379638671875, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7023850679397583, + "num_tokens": 424174791.0, + "step": 16405 + }, + { + "epoch": 1.8016692290797276, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.929837942123413, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7133170366287231, + "num_tokens": 424206433.0, + "step": 16406 + }, + { + "epoch": 1.8017790467823414, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.333188533782959, + "learning_rate": 1e-06, + "loss": 0.7442, + "mean_token_accuracy": 0.7625699639320374, + "num_tokens": 424229767.0, + "step": 16407 + }, + { + "epoch": 1.801888864484955, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2186601161956787, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7203179597854614, + "num_tokens": 424258047.0, + "step": 16408 + }, + { + "epoch": 1.8019986821875686, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3321475982666016, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.703687846660614, + "num_tokens": 424281964.0, + "step": 16409 + }, + { + "epoch": 1.8021084998901822, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1037862300872803, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7177384495735168, + "num_tokens": 424309618.0, + "step": 16410 + }, + { + "epoch": 1.802218317592796, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.223903179168701, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7249947786331177, + "num_tokens": 424336400.0, + "step": 16411 + }, + { + "epoch": 1.8023281352954097, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4603793621063232, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7240286469459534, + "num_tokens": 424358957.0, + "step": 16412 + }, + { + "epoch": 1.8024379529980232, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.938126564025879, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7111166715621948, + "num_tokens": 424390698.0, + "step": 16413 + }, + { + "epoch": 1.802547770700637, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.728132724761963, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7312933206558228, + "num_tokens": 424408802.0, + "step": 16414 + }, + { + "epoch": 1.8026575884032505, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4375393390655518, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7206957936286926, + "num_tokens": 424432093.0, + "step": 16415 + }, + { + "epoch": 1.8027674061058643, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2575488090515137, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7206878662109375, + "num_tokens": 424457333.0, + "step": 16416 + }, + { + "epoch": 1.802877223808478, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4529316425323486, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7272156476974487, + "num_tokens": 424479372.0, + "step": 16417 + }, + { + "epoch": 1.8029870415110916, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2348148822784424, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6986120939254761, + "num_tokens": 424506414.0, + "step": 16418 + }, + { + "epoch": 1.803096859213705, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4771485328674316, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6991429328918457, + "num_tokens": 424530747.0, + "step": 16419 + }, + { + "epoch": 1.8032066769163189, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3248565196990967, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7193353772163391, + "num_tokens": 424554944.0, + "step": 16420 + }, + { + "epoch": 1.8033164946189326, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2778499126434326, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7486006021499634, + "num_tokens": 424577934.0, + "step": 16421 + }, + { + "epoch": 1.8034263123215464, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.302403688430786, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7112785577774048, + "num_tokens": 424600956.0, + "step": 16422 + }, + { + "epoch": 1.80353613002416, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1884090900421143, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7269197106361389, + "num_tokens": 424625842.0, + "step": 16423 + }, + { + "epoch": 1.8036459477267734, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2349069118499756, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7242281436920166, + "num_tokens": 424652210.0, + "step": 16424 + }, + { + "epoch": 1.8037557654293872, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.360358476638794, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7290748357772827, + "num_tokens": 424676026.0, + "step": 16425 + }, + { + "epoch": 1.803865583132001, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.53780460357666, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7121758460998535, + "num_tokens": 424697535.0, + "step": 16426 + }, + { + "epoch": 1.8039754008346145, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.279709815979004, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7301955223083496, + "num_tokens": 424725156.0, + "step": 16427 + }, + { + "epoch": 1.8040852185372283, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.5393686294555664, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7142945528030396, + "num_tokens": 424746333.0, + "step": 16428 + }, + { + "epoch": 1.8041950362398418, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2894465923309326, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7354033589363098, + "num_tokens": 424771557.0, + "step": 16429 + }, + { + "epoch": 1.8043048539424555, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4773058891296387, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7132598757743835, + "num_tokens": 424792970.0, + "step": 16430 + }, + { + "epoch": 1.8044146716450693, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.5511958599090576, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7580853700637817, + "num_tokens": 424812348.0, + "step": 16431 + }, + { + "epoch": 1.8045244893476828, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1275172233581543, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7071269750595093, + "num_tokens": 424840825.0, + "step": 16432 + }, + { + "epoch": 1.8046343070502964, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.346020460128784, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7387292385101318, + "num_tokens": 424864673.0, + "step": 16433 + }, + { + "epoch": 1.8047441247529101, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4492971897125244, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7296417355537415, + "num_tokens": 424885996.0, + "step": 16434 + }, + { + "epoch": 1.8048539424555239, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.264359474182129, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7140229940414429, + "num_tokens": 424913553.0, + "step": 16435 + }, + { + "epoch": 1.8049637601581376, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4465384483337402, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.709858775138855, + "num_tokens": 424936463.0, + "step": 16436 + }, + { + "epoch": 1.8050735778607512, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1077301502227783, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.7032610774040222, + "num_tokens": 424967622.0, + "step": 16437 + }, + { + "epoch": 1.8051833955633647, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.150937795639038, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7094327807426453, + "num_tokens": 424996346.0, + "step": 16438 + }, + { + "epoch": 1.8052932132659785, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.541461706161499, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7301392555236816, + "num_tokens": 425016697.0, + "step": 16439 + }, + { + "epoch": 1.8054030309685922, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3842313289642334, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7212808728218079, + "num_tokens": 425040663.0, + "step": 16440 + }, + { + "epoch": 1.8055128486712058, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2178566455841064, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7343989610671997, + "num_tokens": 425066619.0, + "step": 16441 + }, + { + "epoch": 1.8056226663738193, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.09952974319458, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7139437198638916, + "num_tokens": 425095651.0, + "step": 16442 + }, + { + "epoch": 1.805732484076433, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.937267780303955, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7029228210449219, + "num_tokens": 425129103.0, + "step": 16443 + }, + { + "epoch": 1.8058423017790468, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1379268169403076, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.719097375869751, + "num_tokens": 425158032.0, + "step": 16444 + }, + { + "epoch": 1.8059521194816606, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3955323696136475, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7165552377700806, + "num_tokens": 425179784.0, + "step": 16445 + }, + { + "epoch": 1.806061937184274, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.232231616973877, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7354184985160828, + "num_tokens": 425204646.0, + "step": 16446 + }, + { + "epoch": 1.8061717548868876, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4103856086730957, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7003271579742432, + "num_tokens": 425230154.0, + "step": 16447 + }, + { + "epoch": 1.8062815725895014, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2836811542510986, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7300116419792175, + "num_tokens": 425252949.0, + "step": 16448 + }, + { + "epoch": 1.8063913902921152, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2027502059936523, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7201251983642578, + "num_tokens": 425279039.0, + "step": 16449 + }, + { + "epoch": 1.806501207994729, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.8076893091201782, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7130336761474609, + "num_tokens": 425314120.0, + "step": 16450 + }, + { + "epoch": 1.8066110256973424, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.180997610092163, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7246900796890259, + "num_tokens": 425341235.0, + "step": 16451 + }, + { + "epoch": 1.806720843399956, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.358912944793701, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7457517981529236, + "num_tokens": 425362936.0, + "step": 16452 + }, + { + "epoch": 1.8068306611025697, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2720630168914795, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7131766080856323, + "num_tokens": 425387980.0, + "step": 16453 + }, + { + "epoch": 1.8069404788051835, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.20912504196167, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7286320924758911, + "num_tokens": 425415000.0, + "step": 16454 + }, + { + "epoch": 1.807050296507797, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.417250633239746, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.721917986869812, + "num_tokens": 425438014.0, + "step": 16455 + }, + { + "epoch": 1.8071601142104106, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.105027437210083, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7093229293823242, + "num_tokens": 425465809.0, + "step": 16456 + }, + { + "epoch": 1.8072699319130243, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.12131929397583, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7066183090209961, + "num_tokens": 425494739.0, + "step": 16457 + }, + { + "epoch": 1.807379749615638, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2207021713256836, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7170183658599854, + "num_tokens": 425519222.0, + "step": 16458 + }, + { + "epoch": 1.8074895673182518, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1991755962371826, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7011520266532898, + "num_tokens": 425546860.0, + "step": 16459 + }, + { + "epoch": 1.8075993850208654, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.076284646987915, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6910583972930908, + "num_tokens": 425578098.0, + "step": 16460 + }, + { + "epoch": 1.807709202723479, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.236724376678467, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7108449935913086, + "num_tokens": 425603109.0, + "step": 16461 + }, + { + "epoch": 1.8078190204260927, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4972572326660156, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7014039158821106, + "num_tokens": 425624711.0, + "step": 16462 + }, + { + "epoch": 1.8079288381287064, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1566660404205322, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7138162851333618, + "num_tokens": 425651333.0, + "step": 16463 + }, + { + "epoch": 1.8080386558313202, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0860023498535156, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7191075086593628, + "num_tokens": 425681369.0, + "step": 16464 + }, + { + "epoch": 1.8081484735339337, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.298475503921509, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7509968280792236, + "num_tokens": 425705748.0, + "step": 16465 + }, + { + "epoch": 1.8082582912365472, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2023699283599854, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7500476837158203, + "num_tokens": 425730636.0, + "step": 16466 + }, + { + "epoch": 1.808368108939161, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1509475708007812, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.7094286680221558, + "num_tokens": 425758932.0, + "step": 16467 + }, + { + "epoch": 1.8084779266417748, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.5658552646636963, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7261984348297119, + "num_tokens": 425778446.0, + "step": 16468 + }, + { + "epoch": 1.8085877443443883, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.028949022293091, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.728592038154602, + "num_tokens": 425807780.0, + "step": 16469 + }, + { + "epoch": 1.8086975620470018, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2402708530426025, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.6970877647399902, + "num_tokens": 425834352.0, + "step": 16470 + }, + { + "epoch": 1.8088073797496156, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0380139350891113, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.708530604839325, + "num_tokens": 425864174.0, + "step": 16471 + }, + { + "epoch": 1.8089171974522293, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.022674322128296, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7270294427871704, + "num_tokens": 425894335.0, + "step": 16472 + }, + { + "epoch": 1.809027015154843, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.134209156036377, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7066671848297119, + "num_tokens": 425925002.0, + "step": 16473 + }, + { + "epoch": 1.8091368328574566, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2761590480804443, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7117235660552979, + "num_tokens": 425950776.0, + "step": 16474 + }, + { + "epoch": 1.8092466505600702, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.1228537559509277, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7115363478660583, + "num_tokens": 425977702.0, + "step": 16475 + }, + { + "epoch": 1.809356468262684, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3791606426239014, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7427883148193359, + "num_tokens": 425999777.0, + "step": 16476 + }, + { + "epoch": 1.8094662859652977, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.4197700023651123, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7056159973144531, + "num_tokens": 426023601.0, + "step": 16477 + }, + { + "epoch": 1.8095761036679112, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0423688888549805, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.722956657409668, + "num_tokens": 426052195.0, + "step": 16478 + }, + { + "epoch": 1.809685921370525, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.258934736251831, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7199798822402954, + "num_tokens": 426076543.0, + "step": 16479 + }, + { + "epoch": 1.8097957390731385, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2521615028381348, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7095642685890198, + "num_tokens": 426102646.0, + "step": 16480 + }, + { + "epoch": 1.8099055567757523, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2215628623962402, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.726615309715271, + "num_tokens": 426128498.0, + "step": 16481 + }, + { + "epoch": 1.810015374478366, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2952919006347656, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7219164967536926, + "num_tokens": 426152242.0, + "step": 16482 + }, + { + "epoch": 1.8101251921809796, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.166889190673828, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7173103094100952, + "num_tokens": 426179491.0, + "step": 16483 + }, + { + "epoch": 1.810235009883593, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.093920946121216, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6869299411773682, + "num_tokens": 426210144.0, + "step": 16484 + }, + { + "epoch": 1.8103448275862069, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.0535240173339844, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7067659497261047, + "num_tokens": 426240115.0, + "step": 16485 + }, + { + "epoch": 1.8104546452888206, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.3036253452301025, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7201480269432068, + "num_tokens": 426264571.0, + "step": 16486 + }, + { + "epoch": 1.8105644629914344, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 1.9137829542160034, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.6960396766662598, + "num_tokens": 426299083.0, + "step": 16487 + }, + { + "epoch": 1.810674280694048, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.179877519607544, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7161435484886169, + "num_tokens": 426324702.0, + "step": 16488 + }, + { + "epoch": 1.8107840983966614, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.383676290512085, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7317255139350891, + "num_tokens": 426346607.0, + "step": 16489 + }, + { + "epoch": 1.8108939160992752, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.2388885021209717, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7068547010421753, + "num_tokens": 426370961.0, + "step": 16490 + }, + { + "epoch": 1.811003733801889, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.114469051361084, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7216036319732666, + "num_tokens": 426401716.0, + "step": 16491 + }, + { + "epoch": 1.8111135515045025, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.5819947719573975, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7141312956809998, + "num_tokens": 426423232.0, + "step": 16492 + }, + { + "epoch": 1.8112233692071162, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.582887887954712, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7194668054580688, + "num_tokens": 426444908.0, + "step": 16493 + }, + { + "epoch": 1.8113331869097298, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.576176404953003, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7467085123062134, + "num_tokens": 426464873.0, + "step": 16494 + }, + { + "epoch": 1.8114430046123435, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.019442319869995, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7127747535705566, + "num_tokens": 426493448.0, + "step": 16495 + }, + { + "epoch": 1.8115528223149573, + "ewc_loss": 0.00013828277587890625, + "grad_norm": 2.51667857170105, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7191885113716125, + "num_tokens": 426514810.0, + "step": 16496 + }, + { + "epoch": 1.8116626400175708, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.557443141937256, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7389349937438965, + "num_tokens": 426535324.0, + "step": 16497 + }, + { + "epoch": 1.8117724577201844, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.145317792892456, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7212762832641602, + "num_tokens": 426563419.0, + "step": 16498 + }, + { + "epoch": 1.8118822754227981, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.205045700073242, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.730094313621521, + "num_tokens": 426589380.0, + "step": 16499 + }, + { + "epoch": 1.8119920931254119, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.438575267791748, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7244864702224731, + "num_tokens": 426613698.0, + "step": 16500 + }, + { + "epoch": 1.8121019108280256, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.8171266317367554, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6917554140090942, + "num_tokens": 426652593.0, + "step": 16501 + }, + { + "epoch": 1.8122117285306392, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.071258068084717, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7044659852981567, + "num_tokens": 426682565.0, + "step": 16502 + }, + { + "epoch": 1.8123215462332527, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.8035207986831665, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7144723534584045, + "num_tokens": 426718171.0, + "step": 16503 + }, + { + "epoch": 1.8124313639358665, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.8840241432189941, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7291363477706909, + "num_tokens": 426751561.0, + "step": 16504 + }, + { + "epoch": 1.8125411816384802, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.282696008682251, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7029728889465332, + "num_tokens": 426776114.0, + "step": 16505 + }, + { + "epoch": 1.8126509993410937, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1132867336273193, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7091105580329895, + "num_tokens": 426805410.0, + "step": 16506 + }, + { + "epoch": 1.8127608170437073, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2373785972595215, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7019124031066895, + "num_tokens": 426831192.0, + "step": 16507 + }, + { + "epoch": 1.812870634746321, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.154313087463379, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7025661468505859, + "num_tokens": 426858715.0, + "step": 16508 + }, + { + "epoch": 1.8129804524489348, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.137350559234619, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7000209093093872, + "num_tokens": 426888925.0, + "step": 16509 + }, + { + "epoch": 1.8130902701515486, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.354158401489258, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7241460680961609, + "num_tokens": 426911423.0, + "step": 16510 + }, + { + "epoch": 1.813200087854162, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.455306053161621, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7047092914581299, + "num_tokens": 426935761.0, + "step": 16511 + }, + { + "epoch": 1.8133099055567756, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.189081907272339, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7265657782554626, + "num_tokens": 426963565.0, + "step": 16512 + }, + { + "epoch": 1.8134197232593894, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.241389274597168, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7356691956520081, + "num_tokens": 426988193.0, + "step": 16513 + }, + { + "epoch": 1.8135295409620031, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.8673385381698608, + "learning_rate": 1e-06, + "loss": 1.055, + "mean_token_accuracy": 0.6866638660430908, + "num_tokens": 427022948.0, + "step": 16514 + }, + { + "epoch": 1.813639358664617, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.0695436000823975, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7197278738021851, + "num_tokens": 427052540.0, + "step": 16515 + }, + { + "epoch": 1.8137491763672304, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3278558254241943, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7195194959640503, + "num_tokens": 427075944.0, + "step": 16516 + }, + { + "epoch": 1.813858994069844, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.8981186151504517, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7240583896636963, + "num_tokens": 427108854.0, + "step": 16517 + }, + { + "epoch": 1.8139688117724577, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.4288740158081055, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7182562947273254, + "num_tokens": 427131274.0, + "step": 16518 + }, + { + "epoch": 1.8140786294750715, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3752450942993164, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7141783833503723, + "num_tokens": 427155856.0, + "step": 16519 + }, + { + "epoch": 1.814188447177685, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.164553642272949, + "learning_rate": 1e-06, + "loss": 1.0526, + "mean_token_accuracy": 0.6835628747940063, + "num_tokens": 427185417.0, + "step": 16520 + }, + { + "epoch": 1.8142982648802986, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.202354669570923, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6932862997055054, + "num_tokens": 427212866.0, + "step": 16521 + }, + { + "epoch": 1.8144080825829123, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1741034984588623, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7086825370788574, + "num_tokens": 427238942.0, + "step": 16522 + }, + { + "epoch": 1.814517900285526, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1810569763183594, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.722119152545929, + "num_tokens": 427264887.0, + "step": 16523 + }, + { + "epoch": 1.8146277179881398, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.0643818378448486, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7336665987968445, + "num_tokens": 427292572.0, + "step": 16524 + }, + { + "epoch": 1.8147375356907534, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.388328790664673, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7079495191574097, + "num_tokens": 427316787.0, + "step": 16525 + }, + { + "epoch": 1.814847353393367, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.08959698677063, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6917833089828491, + "num_tokens": 427348035.0, + "step": 16526 + }, + { + "epoch": 1.8149571710959806, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.28631329536438, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7215700149536133, + "num_tokens": 427373501.0, + "step": 16527 + }, + { + "epoch": 1.8150669887985944, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.6007769107818604, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7336934208869934, + "num_tokens": 427392578.0, + "step": 16528 + }, + { + "epoch": 1.815176806501208, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.134723424911499, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7212398648262024, + "num_tokens": 427420076.0, + "step": 16529 + }, + { + "epoch": 1.8152866242038217, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.286100387573242, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7061361074447632, + "num_tokens": 427447000.0, + "step": 16530 + }, + { + "epoch": 1.8153964419064352, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.5530807971954346, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.713498592376709, + "num_tokens": 427467724.0, + "step": 16531 + }, + { + "epoch": 1.815506259609049, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.050295114517212, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7287222146987915, + "num_tokens": 427498695.0, + "step": 16532 + }, + { + "epoch": 1.8156160773116627, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2598679065704346, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7206002473831177, + "num_tokens": 427524312.0, + "step": 16533 + }, + { + "epoch": 1.8157258950142763, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2647624015808105, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6933179497718811, + "num_tokens": 427552420.0, + "step": 16534 + }, + { + "epoch": 1.8158357127168898, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2909693717956543, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7234631180763245, + "num_tokens": 427577265.0, + "step": 16535 + }, + { + "epoch": 1.8159455304195036, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.0055394172668457, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7289614677429199, + "num_tokens": 427607241.0, + "step": 16536 + }, + { + "epoch": 1.8160553481221173, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.6444272994995117, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7266590595245361, + "num_tokens": 427626819.0, + "step": 16537 + }, + { + "epoch": 1.816165165824731, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.042821168899536, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7293432950973511, + "num_tokens": 427656171.0, + "step": 16538 + }, + { + "epoch": 1.8162749835273446, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.139268159866333, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.6975861191749573, + "num_tokens": 427683797.0, + "step": 16539 + }, + { + "epoch": 1.8163848012299582, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.087231397628784, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7401604056358337, + "num_tokens": 427710480.0, + "step": 16540 + }, + { + "epoch": 1.816494618932572, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2114126682281494, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7229573726654053, + "num_tokens": 427736371.0, + "step": 16541 + }, + { + "epoch": 1.8166044366351857, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.4328370094299316, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7075120210647583, + "num_tokens": 427758803.0, + "step": 16542 + }, + { + "epoch": 1.8167142543377992, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.5106828212738037, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7310830950737, + "num_tokens": 427779095.0, + "step": 16543 + }, + { + "epoch": 1.816824072040413, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.5485193729400635, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7070907354354858, + "num_tokens": 427799114.0, + "step": 16544 + }, + { + "epoch": 1.8169338897430265, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1976091861724854, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6949865818023682, + "num_tokens": 427826069.0, + "step": 16545 + }, + { + "epoch": 1.8170437074456403, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3392131328582764, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7199069857597351, + "num_tokens": 427849250.0, + "step": 16546 + }, + { + "epoch": 1.817153525148254, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.096372604370117, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7055619955062866, + "num_tokens": 427880158.0, + "step": 16547 + }, + { + "epoch": 1.8172633428508675, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.9488557577133179, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7226535081863403, + "num_tokens": 427911055.0, + "step": 16548 + }, + { + "epoch": 1.817373160553481, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.0166068077087402, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.6996890306472778, + "num_tokens": 427941798.0, + "step": 16549 + }, + { + "epoch": 1.8174829782560948, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3249385356903076, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7288429141044617, + "num_tokens": 427965018.0, + "step": 16550 + }, + { + "epoch": 1.8175927959587086, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.4549784660339355, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7185644507408142, + "num_tokens": 427986831.0, + "step": 16551 + }, + { + "epoch": 1.8177026136613224, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2756080627441406, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.740627646446228, + "num_tokens": 428011835.0, + "step": 16552 + }, + { + "epoch": 1.8178124313639359, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.7351365089416504, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7279608249664307, + "num_tokens": 428030219.0, + "step": 16553 + }, + { + "epoch": 1.8179222490665494, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.568143606185913, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7234922647476196, + "num_tokens": 428050693.0, + "step": 16554 + }, + { + "epoch": 1.8180320667691632, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.226653575897217, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7257698774337769, + "num_tokens": 428075890.0, + "step": 16555 + }, + { + "epoch": 1.818141884471777, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.147554636001587, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7083820700645447, + "num_tokens": 428104492.0, + "step": 16556 + }, + { + "epoch": 1.8182517021743905, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.502821445465088, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7249723672866821, + "num_tokens": 428125142.0, + "step": 16557 + }, + { + "epoch": 1.818361519877004, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3031744956970215, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7163617014884949, + "num_tokens": 428149052.0, + "step": 16558 + }, + { + "epoch": 1.8184713375796178, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.0943796634674072, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7463281750679016, + "num_tokens": 428175138.0, + "step": 16559 + }, + { + "epoch": 1.8185811552822315, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1957931518554688, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.720954179763794, + "num_tokens": 428201616.0, + "step": 16560 + }, + { + "epoch": 1.8186909729848453, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1087427139282227, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.727421224117279, + "num_tokens": 428229204.0, + "step": 16561 + }, + { + "epoch": 1.8188007906874588, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.318612575531006, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7184813618659973, + "num_tokens": 428252398.0, + "step": 16562 + }, + { + "epoch": 1.8189106083900723, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1357321739196777, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.733485221862793, + "num_tokens": 428280091.0, + "step": 16563 + }, + { + "epoch": 1.819020426092686, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.5683066844940186, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7424353361129761, + "num_tokens": 428303463.0, + "step": 16564 + }, + { + "epoch": 1.8191302437952999, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.353238582611084, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.72174471616745, + "num_tokens": 428326495.0, + "step": 16565 + }, + { + "epoch": 1.8192400614979136, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.05952787399292, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.6988382339477539, + "num_tokens": 428356987.0, + "step": 16566 + }, + { + "epoch": 1.8193498792005272, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.3768882751464844, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7121875286102295, + "num_tokens": 428382039.0, + "step": 16567 + }, + { + "epoch": 1.8194596969031407, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1708734035491943, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7229701280593872, + "num_tokens": 428407816.0, + "step": 16568 + }, + { + "epoch": 1.8195695146057544, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.15158748626709, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7185251116752625, + "num_tokens": 428434878.0, + "step": 16569 + }, + { + "epoch": 1.8196793323083682, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.1517558097839355, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7131122946739197, + "num_tokens": 428461640.0, + "step": 16570 + }, + { + "epoch": 1.8197891500109817, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2882425785064697, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7159594893455505, + "num_tokens": 428486922.0, + "step": 16571 + }, + { + "epoch": 1.8198989677135953, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.482224225997925, + "learning_rate": 1e-06, + "loss": 0.7861, + "mean_token_accuracy": 0.7496576905250549, + "num_tokens": 428509178.0, + "step": 16572 + }, + { + "epoch": 1.820008785416209, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.812185049057007, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7313258647918701, + "num_tokens": 428526688.0, + "step": 16573 + }, + { + "epoch": 1.8201186031188228, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.273733377456665, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7060997486114502, + "num_tokens": 428552595.0, + "step": 16574 + }, + { + "epoch": 1.8202284208214365, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.195828437805176, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7298858165740967, + "num_tokens": 428580590.0, + "step": 16575 + }, + { + "epoch": 1.82033823852405, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.9732781648635864, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7295424938201904, + "num_tokens": 428610668.0, + "step": 16576 + }, + { + "epoch": 1.8204480562266636, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.503150224685669, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.731174886226654, + "num_tokens": 428630137.0, + "step": 16577 + }, + { + "epoch": 1.8205578739292774, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2155704498291016, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7354573011398315, + "num_tokens": 428656268.0, + "step": 16578 + }, + { + "epoch": 1.8206676916318911, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.212203025817871, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7268171310424805, + "num_tokens": 428684183.0, + "step": 16579 + }, + { + "epoch": 1.8207775093345049, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.2542564868927, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7383272051811218, + "num_tokens": 428708305.0, + "step": 16580 + }, + { + "epoch": 1.8208873270371184, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.283268451690674, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.72786945104599, + "num_tokens": 428732776.0, + "step": 16581 + }, + { + "epoch": 1.820997144739732, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.9775835275650024, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7110958099365234, + "num_tokens": 428764808.0, + "step": 16582 + }, + { + "epoch": 1.8211069624423457, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.557605266571045, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7156515121459961, + "num_tokens": 428784878.0, + "step": 16583 + }, + { + "epoch": 1.8212167801449595, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 1.994605302810669, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7380150556564331, + "num_tokens": 428815897.0, + "step": 16584 + }, + { + "epoch": 1.821326597847573, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.46475887298584, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7337114810943604, + "num_tokens": 428837386.0, + "step": 16585 + }, + { + "epoch": 1.8214364155501865, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.053706407546997, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7293537855148315, + "num_tokens": 428863800.0, + "step": 16586 + }, + { + "epoch": 1.8215462332528003, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.4384231567382812, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7504696846008301, + "num_tokens": 428884082.0, + "step": 16587 + }, + { + "epoch": 1.821656050955414, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.5470292568206787, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7425198554992676, + "num_tokens": 428904607.0, + "step": 16588 + }, + { + "epoch": 1.8217658686580278, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.203630208969116, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.716202974319458, + "num_tokens": 428930390.0, + "step": 16589 + }, + { + "epoch": 1.8218756863606413, + "ewc_loss": 0.0001392364501953125, + "grad_norm": 2.354269504547119, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7203026413917542, + "num_tokens": 428955621.0, + "step": 16590 + }, + { + "epoch": 1.8219855040632549, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0677602291107178, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7388111352920532, + "num_tokens": 428986531.0, + "step": 16591 + }, + { + "epoch": 1.8220953217658686, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1648905277252197, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7347453832626343, + "num_tokens": 429011547.0, + "step": 16592 + }, + { + "epoch": 1.8222051394684824, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.240433692932129, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7185441255569458, + "num_tokens": 429037503.0, + "step": 16593 + }, + { + "epoch": 1.822314957171096, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2931172847747803, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7118312120437622, + "num_tokens": 429063438.0, + "step": 16594 + }, + { + "epoch": 1.8224247748737097, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1841952800750732, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7257814407348633, + "num_tokens": 429089960.0, + "step": 16595 + }, + { + "epoch": 1.8225345925763232, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3920674324035645, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7198168039321899, + "num_tokens": 429112940.0, + "step": 16596 + }, + { + "epoch": 1.822644410278937, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.18111515045166, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7410683631896973, + "num_tokens": 429139904.0, + "step": 16597 + }, + { + "epoch": 1.8227542279815507, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.9477589130401611, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7118601202964783, + "num_tokens": 429172666.0, + "step": 16598 + }, + { + "epoch": 1.8228640456841643, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3994927406311035, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7260057926177979, + "num_tokens": 429195114.0, + "step": 16599 + }, + { + "epoch": 1.8229738633867778, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1128785610198975, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7009680271148682, + "num_tokens": 429222275.0, + "step": 16600 + }, + { + "epoch": 1.8230836810893916, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2125327587127686, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7036558389663696, + "num_tokens": 429248468.0, + "step": 16601 + }, + { + "epoch": 1.8231934987920053, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0947725772857666, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7118499279022217, + "num_tokens": 429276517.0, + "step": 16602 + }, + { + "epoch": 1.823303316494619, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.041463851928711, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7102953195571899, + "num_tokens": 429305965.0, + "step": 16603 + }, + { + "epoch": 1.8234131341972326, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2883927822113037, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7120530605316162, + "num_tokens": 429329134.0, + "step": 16604 + }, + { + "epoch": 1.8235229518998461, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.144094944000244, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7099609375, + "num_tokens": 429358836.0, + "step": 16605 + }, + { + "epoch": 1.82363276960246, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.5036165714263916, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7376673221588135, + "num_tokens": 429380611.0, + "step": 16606 + }, + { + "epoch": 1.8237425873050737, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2171106338500977, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.7039422988891602, + "num_tokens": 429407180.0, + "step": 16607 + }, + { + "epoch": 1.8238524050076872, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0984864234924316, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6955298781394958, + "num_tokens": 429435290.0, + "step": 16608 + }, + { + "epoch": 1.823962222710301, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3523449897766113, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7195168733596802, + "num_tokens": 429460877.0, + "step": 16609 + }, + { + "epoch": 1.8240720404129145, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.076408624649048, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7260463237762451, + "num_tokens": 429490257.0, + "step": 16610 + }, + { + "epoch": 1.8241818581155282, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1529242992401123, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6994013786315918, + "num_tokens": 429518656.0, + "step": 16611 + }, + { + "epoch": 1.824291675818142, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.959517240524292, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6917946338653564, + "num_tokens": 429551292.0, + "step": 16612 + }, + { + "epoch": 1.8244014935207555, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.9982842206954956, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6994754076004028, + "num_tokens": 429581395.0, + "step": 16613 + }, + { + "epoch": 1.824511311223369, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0666885375976562, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.723183274269104, + "num_tokens": 429610545.0, + "step": 16614 + }, + { + "epoch": 1.8246211289259828, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0492746829986572, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7043222188949585, + "num_tokens": 429640094.0, + "step": 16615 + }, + { + "epoch": 1.8247309466285966, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.5698869228363037, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.6947241425514221, + "num_tokens": 429664718.0, + "step": 16616 + }, + { + "epoch": 1.8248407643312103, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0202219486236572, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.724602222442627, + "num_tokens": 429693462.0, + "step": 16617 + }, + { + "epoch": 1.8249505820338239, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.146284341812134, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6933164000511169, + "num_tokens": 429722076.0, + "step": 16618 + }, + { + "epoch": 1.8250603997364374, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0715012550354004, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7036702632904053, + "num_tokens": 429751375.0, + "step": 16619 + }, + { + "epoch": 1.8251702174390512, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.4143388271331787, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7213483452796936, + "num_tokens": 429773201.0, + "step": 16620 + }, + { + "epoch": 1.825280035141665, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.082115650177002, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7431573867797852, + "num_tokens": 429801196.0, + "step": 16621 + }, + { + "epoch": 1.8253898528442785, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.366532802581787, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7244750261306763, + "num_tokens": 429823930.0, + "step": 16622 + }, + { + "epoch": 1.825499670546892, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.330496072769165, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.714957594871521, + "num_tokens": 429852022.0, + "step": 16623 + }, + { + "epoch": 1.8256094882495058, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0722522735595703, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.6991065740585327, + "num_tokens": 429881626.0, + "step": 16624 + }, + { + "epoch": 1.8257193059521195, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0758142471313477, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7177587747573853, + "num_tokens": 429910784.0, + "step": 16625 + }, + { + "epoch": 1.8258291236547333, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0263657569885254, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7223140001296997, + "num_tokens": 429940358.0, + "step": 16626 + }, + { + "epoch": 1.8259389413573468, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0685818195343018, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.6970787644386292, + "num_tokens": 429969579.0, + "step": 16627 + }, + { + "epoch": 1.8260487590599603, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.322442054748535, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7065110802650452, + "num_tokens": 429996881.0, + "step": 16628 + }, + { + "epoch": 1.826158576762574, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.253774404525757, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7256017327308655, + "num_tokens": 430022276.0, + "step": 16629 + }, + { + "epoch": 1.8262683944651878, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2366607189178467, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7249321937561035, + "num_tokens": 430047514.0, + "step": 16630 + }, + { + "epoch": 1.8263782121678016, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3767874240875244, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7195602655410767, + "num_tokens": 430068455.0, + "step": 16631 + }, + { + "epoch": 1.8264880298704151, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3994531631469727, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7185924053192139, + "num_tokens": 430091890.0, + "step": 16632 + }, + { + "epoch": 1.8265978475730287, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.0261125564575195, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6918889880180359, + "num_tokens": 430122910.0, + "step": 16633 + }, + { + "epoch": 1.8267076652756424, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3047144412994385, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7090311050415039, + "num_tokens": 430148544.0, + "step": 16634 + }, + { + "epoch": 1.8268174829782562, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.128207206726074, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6953697800636292, + "num_tokens": 430177519.0, + "step": 16635 + }, + { + "epoch": 1.8269273006808697, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3348491191864014, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.716816782951355, + "num_tokens": 430202491.0, + "step": 16636 + }, + { + "epoch": 1.8270371183834833, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2568602561950684, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7255409955978394, + "num_tokens": 430228359.0, + "step": 16637 + }, + { + "epoch": 1.827146936086097, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.667572259902954, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7235643863677979, + "num_tokens": 430246665.0, + "step": 16638 + }, + { + "epoch": 1.8272567537887108, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.248159408569336, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7260400056838989, + "num_tokens": 430271257.0, + "step": 16639 + }, + { + "epoch": 1.8273665714913245, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.177063226699829, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7453063130378723, + "num_tokens": 430295702.0, + "step": 16640 + }, + { + "epoch": 1.827476389193938, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.317209482192993, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7266798615455627, + "num_tokens": 430319395.0, + "step": 16641 + }, + { + "epoch": 1.8275862068965516, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2922470569610596, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7148159742355347, + "num_tokens": 430344876.0, + "step": 16642 + }, + { + "epoch": 1.8276960245991654, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1748080253601074, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7367398738861084, + "num_tokens": 430369863.0, + "step": 16643 + }, + { + "epoch": 1.8278058423017791, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3841731548309326, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7114366292953491, + "num_tokens": 430394954.0, + "step": 16644 + }, + { + "epoch": 1.8279156600043929, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.9738566875457764, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.6986520290374756, + "num_tokens": 430426202.0, + "step": 16645 + }, + { + "epoch": 1.8280254777070064, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2522435188293457, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7080309987068176, + "num_tokens": 430452728.0, + "step": 16646 + }, + { + "epoch": 1.82813529540962, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3223319053649902, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7208248376846313, + "num_tokens": 430476884.0, + "step": 16647 + }, + { + "epoch": 1.8282451131122337, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2129218578338623, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7216062545776367, + "num_tokens": 430501811.0, + "step": 16648 + }, + { + "epoch": 1.8283549308148475, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1922237873077393, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7327044606208801, + "num_tokens": 430525781.0, + "step": 16649 + }, + { + "epoch": 1.828464748517461, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.9902375936508179, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7259631156921387, + "num_tokens": 430556236.0, + "step": 16650 + }, + { + "epoch": 1.8285745662200745, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.699861526489258, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7337344288825989, + "num_tokens": 430575557.0, + "step": 16651 + }, + { + "epoch": 1.8286843839226883, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.042170763015747, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7383874654769897, + "num_tokens": 430602998.0, + "step": 16652 + }, + { + "epoch": 1.828794201625302, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2987136840820312, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7227329015731812, + "num_tokens": 430626485.0, + "step": 16653 + }, + { + "epoch": 1.8289040193279158, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.169940233230591, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7228425741195679, + "num_tokens": 430651993.0, + "step": 16654 + }, + { + "epoch": 1.8290138370305293, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.153015613555908, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7088221907615662, + "num_tokens": 430677563.0, + "step": 16655 + }, + { + "epoch": 1.8291236547331429, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1008946895599365, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6987737417221069, + "num_tokens": 430705592.0, + "step": 16656 + }, + { + "epoch": 1.8292334724357566, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.7964632511138916, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7203969955444336, + "num_tokens": 430725170.0, + "step": 16657 + }, + { + "epoch": 1.8293432901383704, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1923680305480957, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7062052488327026, + "num_tokens": 430751860.0, + "step": 16658 + }, + { + "epoch": 1.829453107840984, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1454765796661377, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7069494128227234, + "num_tokens": 430781725.0, + "step": 16659 + }, + { + "epoch": 1.8295629255435977, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.168301820755005, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7391747236251831, + "num_tokens": 430807699.0, + "step": 16660 + }, + { + "epoch": 1.8296727432462112, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2905075550079346, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7279022336006165, + "num_tokens": 430832221.0, + "step": 16661 + }, + { + "epoch": 1.829782560948825, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2865569591522217, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7218309640884399, + "num_tokens": 430856288.0, + "step": 16662 + }, + { + "epoch": 1.8298923786514387, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.397453784942627, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7162730693817139, + "num_tokens": 430878692.0, + "step": 16663 + }, + { + "epoch": 1.8300021963540523, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.228163957595825, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7327858805656433, + "num_tokens": 430902390.0, + "step": 16664 + }, + { + "epoch": 1.8301120140566658, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2178573608398438, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7138866186141968, + "num_tokens": 430927656.0, + "step": 16665 + }, + { + "epoch": 1.8302218317592795, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.153456449508667, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7206649780273438, + "num_tokens": 430955307.0, + "step": 16666 + }, + { + "epoch": 1.8303316494618933, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.092883348464966, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7175143957138062, + "num_tokens": 430984328.0, + "step": 16667 + }, + { + "epoch": 1.830441467164507, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 1.8692656755447388, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.706123948097229, + "num_tokens": 431019994.0, + "step": 16668 + }, + { + "epoch": 1.8305512848671206, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.342337131500244, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7368913888931274, + "num_tokens": 431042177.0, + "step": 16669 + }, + { + "epoch": 1.8306611025697341, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.1324055194854736, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7166387438774109, + "num_tokens": 431068939.0, + "step": 16670 + }, + { + "epoch": 1.830770920272348, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.4678070545196533, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7053649425506592, + "num_tokens": 431089876.0, + "step": 16671 + }, + { + "epoch": 1.8308807379749616, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.288876533508301, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6973437070846558, + "num_tokens": 431114712.0, + "step": 16672 + }, + { + "epoch": 1.8309905556775752, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3500571250915527, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7336684465408325, + "num_tokens": 431136945.0, + "step": 16673 + }, + { + "epoch": 1.831100373380189, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.4504165649414062, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7196208238601685, + "num_tokens": 431159197.0, + "step": 16674 + }, + { + "epoch": 1.8312101910828025, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.3249363899230957, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7092103958129883, + "num_tokens": 431183436.0, + "step": 16675 + }, + { + "epoch": 1.8313200087854162, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.2212815284729004, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7173140048980713, + "num_tokens": 431208000.0, + "step": 16676 + }, + { + "epoch": 1.83142982648803, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.4968788623809814, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7162190675735474, + "num_tokens": 431229066.0, + "step": 16677 + }, + { + "epoch": 1.8315396441906435, + "ewc_loss": 0.00014019012451171875, + "grad_norm": 2.336000442504883, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7315974831581116, + "num_tokens": 431253228.0, + "step": 16678 + }, + { + "epoch": 1.831649461893257, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.290865659713745, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7146658897399902, + "num_tokens": 431277606.0, + "step": 16679 + }, + { + "epoch": 1.8317592795958708, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.0684821605682373, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.696557343006134, + "num_tokens": 431306796.0, + "step": 16680 + }, + { + "epoch": 1.8318690972984846, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2221899032592773, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.712342381477356, + "num_tokens": 431332376.0, + "step": 16681 + }, + { + "epoch": 1.8319789150010983, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.957039713859558, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7234127521514893, + "num_tokens": 431361858.0, + "step": 16682 + }, + { + "epoch": 1.8320887327037119, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.309966564178467, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7220858335494995, + "num_tokens": 431385909.0, + "step": 16683 + }, + { + "epoch": 1.8321985504063254, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2135989665985107, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.6984202861785889, + "num_tokens": 431411587.0, + "step": 16684 + }, + { + "epoch": 1.8323083681089392, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2552225589752197, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7166720628738403, + "num_tokens": 431438484.0, + "step": 16685 + }, + { + "epoch": 1.832418185811553, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3021416664123535, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7027108073234558, + "num_tokens": 431463481.0, + "step": 16686 + }, + { + "epoch": 1.8325280035141664, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.0969481468200684, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7201430797576904, + "num_tokens": 431491588.0, + "step": 16687 + }, + { + "epoch": 1.83263782121678, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2510218620300293, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7453588247299194, + "num_tokens": 431517339.0, + "step": 16688 + }, + { + "epoch": 1.8327476389193937, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1785404682159424, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7494909763336182, + "num_tokens": 431543952.0, + "step": 16689 + }, + { + "epoch": 1.8328574566220075, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.454699754714966, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7139084935188293, + "num_tokens": 431566628.0, + "step": 16690 + }, + { + "epoch": 1.8329672743246213, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2305562496185303, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7095555663108826, + "num_tokens": 431595692.0, + "step": 16691 + }, + { + "epoch": 1.8330770920272348, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4956369400024414, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7141163349151611, + "num_tokens": 431617405.0, + "step": 16692 + }, + { + "epoch": 1.8331869097298483, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.555699348449707, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7240820527076721, + "num_tokens": 431639808.0, + "step": 16693 + }, + { + "epoch": 1.833296727432462, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.227377414703369, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7249084711074829, + "num_tokens": 431665541.0, + "step": 16694 + }, + { + "epoch": 1.8334065451350758, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4234368801116943, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7456362247467041, + "num_tokens": 431686833.0, + "step": 16695 + }, + { + "epoch": 1.8335163628376896, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.5498452186584473, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7149111032485962, + "num_tokens": 431707512.0, + "step": 16696 + }, + { + "epoch": 1.8336261805403031, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.215141773223877, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7220789194107056, + "num_tokens": 431734053.0, + "step": 16697 + }, + { + "epoch": 1.8337359982429167, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.191864490509033, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7095129489898682, + "num_tokens": 431762374.0, + "step": 16698 + }, + { + "epoch": 1.8338458159455304, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.200009822845459, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7091375589370728, + "num_tokens": 431791432.0, + "step": 16699 + }, + { + "epoch": 1.8339556336481442, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.201160192489624, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7192689180374146, + "num_tokens": 431815212.0, + "step": 16700 + }, + { + "epoch": 1.8340654513507577, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1771492958068848, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7394211888313293, + "num_tokens": 431840246.0, + "step": 16701 + }, + { + "epoch": 1.8341752690533712, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2620856761932373, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.711764395236969, + "num_tokens": 431865239.0, + "step": 16702 + }, + { + "epoch": 1.834285086755985, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.285264492034912, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7148442268371582, + "num_tokens": 431889686.0, + "step": 16703 + }, + { + "epoch": 1.8343949044585988, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.071406126022339, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7117789387702942, + "num_tokens": 431916534.0, + "step": 16704 + }, + { + "epoch": 1.8345047221612125, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4070303440093994, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.715869665145874, + "num_tokens": 431939591.0, + "step": 16705 + }, + { + "epoch": 1.834614539863826, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.348386764526367, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7350873947143555, + "num_tokens": 431962665.0, + "step": 16706 + }, + { + "epoch": 1.8347243575664396, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.299696207046509, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7254713177680969, + "num_tokens": 431986591.0, + "step": 16707 + }, + { + "epoch": 1.8348341752690533, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.184361219406128, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7312737703323364, + "num_tokens": 432012179.0, + "step": 16708 + }, + { + "epoch": 1.834943992971667, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3690574169158936, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7218358516693115, + "num_tokens": 432035559.0, + "step": 16709 + }, + { + "epoch": 1.8350538106742806, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.375023603439331, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.734339714050293, + "num_tokens": 432058218.0, + "step": 16710 + }, + { + "epoch": 1.8351636283768944, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2169501781463623, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7292864918708801, + "num_tokens": 432087561.0, + "step": 16711 + }, + { + "epoch": 1.835273446079508, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.0688540935516357, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7329527139663696, + "num_tokens": 432115675.0, + "step": 16712 + }, + { + "epoch": 1.8353832637821217, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.176893711090088, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7052099704742432, + "num_tokens": 432143933.0, + "step": 16713 + }, + { + "epoch": 1.8354930814847354, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9531233310699463, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7058672904968262, + "num_tokens": 432174272.0, + "step": 16714 + }, + { + "epoch": 1.835602899187349, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4627482891082764, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7038331627845764, + "num_tokens": 432196508.0, + "step": 16715 + }, + { + "epoch": 1.8357127168899625, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4696483612060547, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7157220244407654, + "num_tokens": 432219663.0, + "step": 16716 + }, + { + "epoch": 1.8358225345925763, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.0360865592956543, + "learning_rate": 1e-06, + "loss": 1.0785, + "mean_token_accuracy": 0.681003749370575, + "num_tokens": 432251645.0, + "step": 16717 + }, + { + "epoch": 1.83593235229519, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.023677349090576, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7174919247627258, + "num_tokens": 432279998.0, + "step": 16718 + }, + { + "epoch": 1.8360421699978038, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.159514904022217, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7079545855522156, + "num_tokens": 432309369.0, + "step": 16719 + }, + { + "epoch": 1.8361519877004173, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.6672000885009766, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7406454086303711, + "num_tokens": 432335263.0, + "step": 16720 + }, + { + "epoch": 1.8362618054030309, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3072917461395264, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7133873701095581, + "num_tokens": 432362386.0, + "step": 16721 + }, + { + "epoch": 1.8363716231056446, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.296884536743164, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6937006115913391, + "num_tokens": 432388428.0, + "step": 16722 + }, + { + "epoch": 1.8364814408082584, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1454153060913086, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7056554555892944, + "num_tokens": 432415030.0, + "step": 16723 + }, + { + "epoch": 1.836591258510872, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2367215156555176, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7198227643966675, + "num_tokens": 432438623.0, + "step": 16724 + }, + { + "epoch": 1.8367010762134857, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.552457094192505, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7367829084396362, + "num_tokens": 432458752.0, + "step": 16725 + }, + { + "epoch": 1.8368108939160992, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.294912338256836, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7117739319801331, + "num_tokens": 432482944.0, + "step": 16726 + }, + { + "epoch": 1.836920711618713, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1104042530059814, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7164744138717651, + "num_tokens": 432510600.0, + "step": 16727 + }, + { + "epoch": 1.8370305293213267, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4233036041259766, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7439783811569214, + "num_tokens": 432532220.0, + "step": 16728 + }, + { + "epoch": 1.8371403470239402, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3628149032592773, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7273197174072266, + "num_tokens": 432556467.0, + "step": 16729 + }, + { + "epoch": 1.8372501647265538, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9948633909225464, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7071722149848938, + "num_tokens": 432588323.0, + "step": 16730 + }, + { + "epoch": 1.8373599824291675, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3815650939941406, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7059577107429504, + "num_tokens": 432610998.0, + "step": 16731 + }, + { + "epoch": 1.8374698001317813, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2105259895324707, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7077930569648743, + "num_tokens": 432639635.0, + "step": 16732 + }, + { + "epoch": 1.837579617834395, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2724828720092773, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7400518655776978, + "num_tokens": 432665412.0, + "step": 16733 + }, + { + "epoch": 1.8376894355370086, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.6543796062469482, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.73072350025177, + "num_tokens": 432684610.0, + "step": 16734 + }, + { + "epoch": 1.8377992532396221, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.452744960784912, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.716636061668396, + "num_tokens": 432707798.0, + "step": 16735 + }, + { + "epoch": 1.8379090709422359, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.398733139038086, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7289841175079346, + "num_tokens": 432731428.0, + "step": 16736 + }, + { + "epoch": 1.8380188886448496, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4189555644989014, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.713240385055542, + "num_tokens": 432755040.0, + "step": 16737 + }, + { + "epoch": 1.8381287063474632, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.5565123558044434, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7454059720039368, + "num_tokens": 432775161.0, + "step": 16738 + }, + { + "epoch": 1.838238524050077, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4279165267944336, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7252262830734253, + "num_tokens": 432797686.0, + "step": 16739 + }, + { + "epoch": 1.8383483417526905, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.948043704032898, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.70287024974823, + "num_tokens": 432831722.0, + "step": 16740 + }, + { + "epoch": 1.8384581594553042, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2289514541625977, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7227094173431396, + "num_tokens": 432857199.0, + "step": 16741 + }, + { + "epoch": 1.838567977157918, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3033254146575928, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7068593502044678, + "num_tokens": 432884640.0, + "step": 16742 + }, + { + "epoch": 1.8386777948605315, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.5855484008789062, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7178052663803101, + "num_tokens": 432906955.0, + "step": 16743 + }, + { + "epoch": 1.838787612563145, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.031621217727661, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7223109006881714, + "num_tokens": 432936136.0, + "step": 16744 + }, + { + "epoch": 1.8388974302657588, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.147169589996338, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7159876823425293, + "num_tokens": 432962710.0, + "step": 16745 + }, + { + "epoch": 1.8390072479683726, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3827390670776367, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7244751453399658, + "num_tokens": 432985909.0, + "step": 16746 + }, + { + "epoch": 1.8391170656709863, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4230759143829346, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7307459115982056, + "num_tokens": 433005395.0, + "step": 16747 + }, + { + "epoch": 1.8392268833735999, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1780905723571777, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7357895374298096, + "num_tokens": 433031676.0, + "step": 16748 + }, + { + "epoch": 1.8393367010762134, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.205502510070801, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7192654609680176, + "num_tokens": 433058830.0, + "step": 16749 + }, + { + "epoch": 1.8394465187788271, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3207826614379883, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7294594645500183, + "num_tokens": 433082843.0, + "step": 16750 + }, + { + "epoch": 1.839556336481441, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.5312345027923584, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7279210090637207, + "num_tokens": 433102919.0, + "step": 16751 + }, + { + "epoch": 1.8396661541840544, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.7844431400299072, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7388195395469666, + "num_tokens": 433120615.0, + "step": 16752 + }, + { + "epoch": 1.839775971886668, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4301939010620117, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7173760533332825, + "num_tokens": 433144967.0, + "step": 16753 + }, + { + "epoch": 1.8398857895892817, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.477745532989502, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7166950106620789, + "num_tokens": 433167433.0, + "step": 16754 + }, + { + "epoch": 1.8399956072918955, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.166301727294922, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7147910594940186, + "num_tokens": 433194566.0, + "step": 16755 + }, + { + "epoch": 1.8401054249945092, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.077677011489868, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7270926237106323, + "num_tokens": 433222826.0, + "step": 16756 + }, + { + "epoch": 1.8402152426971228, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2903828620910645, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7207857370376587, + "num_tokens": 433246856.0, + "step": 16757 + }, + { + "epoch": 1.8403250603997363, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.27237606048584, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.732216477394104, + "num_tokens": 433270364.0, + "step": 16758 + }, + { + "epoch": 1.84043487810235, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2387077808380127, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7025452256202698, + "num_tokens": 433298091.0, + "step": 16759 + }, + { + "epoch": 1.8405446958049638, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.066035509109497, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7431785464286804, + "num_tokens": 433326190.0, + "step": 16760 + }, + { + "epoch": 1.8406545135075776, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9255157709121704, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7206845283508301, + "num_tokens": 433356604.0, + "step": 16761 + }, + { + "epoch": 1.8407643312101911, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.16914701461792, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7110010981559753, + "num_tokens": 433383447.0, + "step": 16762 + }, + { + "epoch": 1.8408741489128047, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.180556297302246, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6956005692481995, + "num_tokens": 433411016.0, + "step": 16763 + }, + { + "epoch": 1.8409839666154184, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.191979169845581, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7108684778213501, + "num_tokens": 433437278.0, + "step": 16764 + }, + { + "epoch": 1.8410937843180322, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.427536725997925, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.733758807182312, + "num_tokens": 433458780.0, + "step": 16765 + }, + { + "epoch": 1.8412036020206457, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.163015365600586, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7045668363571167, + "num_tokens": 433483918.0, + "step": 16766 + }, + { + "epoch": 1.8413134197232592, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.450509548187256, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7239423990249634, + "num_tokens": 433505207.0, + "step": 16767 + }, + { + "epoch": 1.841423237425873, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4227404594421387, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7452623248100281, + "num_tokens": 433527139.0, + "step": 16768 + }, + { + "epoch": 1.8415330551284868, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2754316329956055, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7172459959983826, + "num_tokens": 433552161.0, + "step": 16769 + }, + { + "epoch": 1.8416428728311005, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.3094024658203125, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6924504637718201, + "num_tokens": 433578670.0, + "step": 16770 + }, + { + "epoch": 1.841752690533714, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.4184606075286865, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7150745987892151, + "num_tokens": 433599905.0, + "step": 16771 + }, + { + "epoch": 1.8418625082363276, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.187284469604492, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7037388682365417, + "num_tokens": 433628044.0, + "step": 16772 + }, + { + "epoch": 1.8419723259389413, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.1866984367370605, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7184323668479919, + "num_tokens": 433654532.0, + "step": 16773 + }, + { + "epoch": 1.842082143641555, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.958217978477478, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7245937585830688, + "num_tokens": 433685377.0, + "step": 16774 + }, + { + "epoch": 1.8421919613441686, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.27719783782959, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7175750732421875, + "num_tokens": 433708832.0, + "step": 16775 + }, + { + "epoch": 1.8423017790467824, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.159424066543579, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7477370500564575, + "num_tokens": 433733598.0, + "step": 16776 + }, + { + "epoch": 1.842411596749396, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.270296573638916, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7400966882705688, + "num_tokens": 433757291.0, + "step": 16777 + }, + { + "epoch": 1.8425214144520097, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.224771499633789, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7305570840835571, + "num_tokens": 433781578.0, + "step": 16778 + }, + { + "epoch": 1.8426312321546234, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.2176551818847656, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7065210938453674, + "num_tokens": 433807409.0, + "step": 16779 + }, + { + "epoch": 1.842741049857237, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.0016822814941406, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7200827598571777, + "num_tokens": 433836391.0, + "step": 16780 + }, + { + "epoch": 1.8428508675598505, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9774010181427002, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6949120163917542, + "num_tokens": 433869430.0, + "step": 16781 + }, + { + "epoch": 1.8429606852624643, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9960198402404785, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7202010154724121, + "num_tokens": 433899256.0, + "step": 16782 + }, + { + "epoch": 1.843070502965078, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.330390214920044, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7146474719047546, + "num_tokens": 433922331.0, + "step": 16783 + }, + { + "epoch": 1.8431803206676918, + "ewc_loss": 0.000141143798828125, + "grad_norm": 1.9946106672286987, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7185256481170654, + "num_tokens": 433952389.0, + "step": 16784 + }, + { + "epoch": 1.8432901383703053, + "ewc_loss": 0.000141143798828125, + "grad_norm": 2.5593960285186768, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.732896089553833, + "num_tokens": 433972878.0, + "step": 16785 + }, + { + "epoch": 1.8433999560729188, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.3631951808929443, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7284793257713318, + "num_tokens": 433996169.0, + "step": 16786 + }, + { + "epoch": 1.8435097737755326, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1198627948760986, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7425416707992554, + "num_tokens": 434023402.0, + "step": 16787 + }, + { + "epoch": 1.8436195914781464, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.237602472305298, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7251717448234558, + "num_tokens": 434048443.0, + "step": 16788 + }, + { + "epoch": 1.84372940918076, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.241013526916504, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7334409952163696, + "num_tokens": 434074861.0, + "step": 16789 + }, + { + "epoch": 1.8438392268833736, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1506145000457764, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.707813024520874, + "num_tokens": 434105980.0, + "step": 16790 + }, + { + "epoch": 1.8439490445859872, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.3377227783203125, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7202958464622498, + "num_tokens": 434129951.0, + "step": 16791 + }, + { + "epoch": 1.844058862288601, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.4897148609161377, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7391998171806335, + "num_tokens": 434151178.0, + "step": 16792 + }, + { + "epoch": 1.8441686799912147, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.140265464782715, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7318713068962097, + "num_tokens": 434179016.0, + "step": 16793 + }, + { + "epoch": 1.8442784976938282, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.654961585998535, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7127492427825928, + "num_tokens": 434198588.0, + "step": 16794 + }, + { + "epoch": 1.8443883153964418, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.345473051071167, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7393748760223389, + "num_tokens": 434220242.0, + "step": 16795 + }, + { + "epoch": 1.8444981330990555, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1625921726226807, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7140890955924988, + "num_tokens": 434246939.0, + "step": 16796 + }, + { + "epoch": 1.8446079508016693, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.065858840942383, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6958760023117065, + "num_tokens": 434275946.0, + "step": 16797 + }, + { + "epoch": 1.844717768504283, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.3989064693450928, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7127334475517273, + "num_tokens": 434299740.0, + "step": 16798 + }, + { + "epoch": 1.8448275862068966, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.0042243003845215, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.6995787620544434, + "num_tokens": 434330720.0, + "step": 16799 + }, + { + "epoch": 1.84493740390951, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1186270713806152, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7171252965927124, + "num_tokens": 434359474.0, + "step": 16800 + }, + { + "epoch": 1.8450472216121239, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.241274118423462, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7304713726043701, + "num_tokens": 434383772.0, + "step": 16801 + }, + { + "epoch": 1.8451570393147376, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.042459726333618, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7202326655387878, + "num_tokens": 434412512.0, + "step": 16802 + }, + { + "epoch": 1.8452668570173512, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.295633554458618, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7193917036056519, + "num_tokens": 434435345.0, + "step": 16803 + }, + { + "epoch": 1.8453766747199647, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.334994316101074, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7144823670387268, + "num_tokens": 434459575.0, + "step": 16804 + }, + { + "epoch": 1.8454864924225785, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.200924873352051, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7102799415588379, + "num_tokens": 434487834.0, + "step": 16805 + }, + { + "epoch": 1.8455963101251922, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.617064952850342, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7232908010482788, + "num_tokens": 434508043.0, + "step": 16806 + }, + { + "epoch": 1.845706127827806, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.2469208240509033, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7363784313201904, + "num_tokens": 434531909.0, + "step": 16807 + }, + { + "epoch": 1.8458159455304195, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.2670226097106934, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7197608351707458, + "num_tokens": 434556812.0, + "step": 16808 + }, + { + "epoch": 1.845925763233033, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1190905570983887, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7286297082901001, + "num_tokens": 434585807.0, + "step": 16809 + }, + { + "epoch": 1.8460355809356468, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 1.9674407243728638, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7123163938522339, + "num_tokens": 434619430.0, + "step": 16810 + }, + { + "epoch": 1.8461453986382605, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.3418080806732178, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7129247188568115, + "num_tokens": 434644649.0, + "step": 16811 + }, + { + "epoch": 1.8462552163408743, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 1.9918150901794434, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7264500260353088, + "num_tokens": 434675536.0, + "step": 16812 + }, + { + "epoch": 1.8463650340434878, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.198134660720825, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7146869897842407, + "num_tokens": 434702111.0, + "step": 16813 + }, + { + "epoch": 1.8464748517461014, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1568479537963867, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7410003542900085, + "num_tokens": 434727742.0, + "step": 16814 + }, + { + "epoch": 1.8465846694487151, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1733176708221436, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7235379219055176, + "num_tokens": 434753236.0, + "step": 16815 + }, + { + "epoch": 1.846694487151329, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.1460182666778564, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7093044519424438, + "num_tokens": 434780497.0, + "step": 16816 + }, + { + "epoch": 1.8468043048539424, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.483107805252075, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7419964075088501, + "num_tokens": 434801305.0, + "step": 16817 + }, + { + "epoch": 1.846914122556556, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.345012664794922, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7358112931251526, + "num_tokens": 434825040.0, + "step": 16818 + }, + { + "epoch": 1.8470239402591697, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.503133535385132, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7118312120437622, + "num_tokens": 434848252.0, + "step": 16819 + }, + { + "epoch": 1.8471337579617835, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.2850334644317627, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7128146290779114, + "num_tokens": 434874096.0, + "step": 16820 + }, + { + "epoch": 1.8472435756643972, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.4671833515167236, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7411169409751892, + "num_tokens": 434894708.0, + "step": 16821 + }, + { + "epoch": 1.8473533933670108, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.0515401363372803, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7159018516540527, + "num_tokens": 434923205.0, + "step": 16822 + }, + { + "epoch": 1.8474632110696243, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.56895112991333, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7416449785232544, + "num_tokens": 434944794.0, + "step": 16823 + }, + { + "epoch": 1.847573028772238, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.251492500305176, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7134147882461548, + "num_tokens": 434969538.0, + "step": 16824 + }, + { + "epoch": 1.8476828464748518, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.4146502017974854, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7165575623512268, + "num_tokens": 434991412.0, + "step": 16825 + }, + { + "epoch": 1.8477926641774656, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.143585681915283, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6988399028778076, + "num_tokens": 435021016.0, + "step": 16826 + }, + { + "epoch": 1.847902481880079, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.3782145977020264, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7090295553207397, + "num_tokens": 435044767.0, + "step": 16827 + }, + { + "epoch": 1.8480122995826926, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.2679378986358643, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7260133028030396, + "num_tokens": 435068460.0, + "step": 16828 + }, + { + "epoch": 1.8481221172853064, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.232985496520996, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7229251861572266, + "num_tokens": 435092176.0, + "step": 16829 + }, + { + "epoch": 1.8482319349879202, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.057288408279419, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7091751098632812, + "num_tokens": 435122503.0, + "step": 16830 + }, + { + "epoch": 1.8483417526905337, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.3180582523345947, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7221568822860718, + "num_tokens": 435147829.0, + "step": 16831 + }, + { + "epoch": 1.8484515703931472, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.3354883193969727, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7065330743789673, + "num_tokens": 435170486.0, + "step": 16832 + }, + { + "epoch": 1.848561388095761, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.174619436264038, + "learning_rate": 1e-06, + "loss": 1.014, + "mean_token_accuracy": 0.7151154279708862, + "num_tokens": 435198505.0, + "step": 16833 + }, + { + "epoch": 1.8486712057983747, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.5619776248931885, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7178574800491333, + "num_tokens": 435220991.0, + "step": 16834 + }, + { + "epoch": 1.8487810235009885, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.3739402294158936, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7136678695678711, + "num_tokens": 435247749.0, + "step": 16835 + }, + { + "epoch": 1.848890841203602, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.682438373565674, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7159101963043213, + "num_tokens": 435266982.0, + "step": 16836 + }, + { + "epoch": 1.8490006589062156, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.5966153144836426, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7415821552276611, + "num_tokens": 435286124.0, + "step": 16837 + }, + { + "epoch": 1.8491104766088293, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.2745351791381836, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7430047392845154, + "num_tokens": 435308497.0, + "step": 16838 + }, + { + "epoch": 1.849220294311443, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.099066972732544, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6842674612998962, + "num_tokens": 435337722.0, + "step": 16839 + }, + { + "epoch": 1.8493301120140566, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.46596622467041, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7371499538421631, + "num_tokens": 435358826.0, + "step": 16840 + }, + { + "epoch": 1.8494399297166704, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.435703992843628, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7043666243553162, + "num_tokens": 435383016.0, + "step": 16841 + }, + { + "epoch": 1.849549747419284, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.215012311935425, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.7023516297340393, + "num_tokens": 435411040.0, + "step": 16842 + }, + { + "epoch": 1.8496595651218977, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.226954460144043, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7322877049446106, + "num_tokens": 435437714.0, + "step": 16843 + }, + { + "epoch": 1.8497693828245114, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.204756021499634, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7207927703857422, + "num_tokens": 435463892.0, + "step": 16844 + }, + { + "epoch": 1.849879200527125, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.040097236633301, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7025021910667419, + "num_tokens": 435494700.0, + "step": 16845 + }, + { + "epoch": 1.8499890182297385, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.6413536071777344, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7307619452476501, + "num_tokens": 435514755.0, + "step": 16846 + }, + { + "epoch": 1.8500988359323522, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.584010362625122, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.748066246509552, + "num_tokens": 435534376.0, + "step": 16847 + }, + { + "epoch": 1.850208653634966, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.5108587741851807, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7156267166137695, + "num_tokens": 435557910.0, + "step": 16848 + }, + { + "epoch": 1.8503184713375798, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 1.849041223526001, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7027788162231445, + "num_tokens": 435593600.0, + "step": 16849 + }, + { + "epoch": 1.8504282890401933, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.4582104682922363, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.6994264721870422, + "num_tokens": 435615993.0, + "step": 16850 + }, + { + "epoch": 1.8505381067428068, + "ewc_loss": 0.0001430511474609375, + "grad_norm": 2.227518081665039, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7164970636367798, + "num_tokens": 435640678.0, + "step": 16851 + }, + { + "epoch": 1.8506479244454206, + "ewc_loss": 0.00014400482177734375, + "grad_norm": 2.257408857345581, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7072361707687378, + "num_tokens": 435666368.0, + "step": 16852 + }, + { + "epoch": 1.8507577421480343, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4889144897460938, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7292935252189636, + "num_tokens": 435686611.0, + "step": 16853 + }, + { + "epoch": 1.8508675598506479, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.129196882247925, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.716519296169281, + "num_tokens": 435715453.0, + "step": 16854 + }, + { + "epoch": 1.8509773775532616, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.437922239303589, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7020753622055054, + "num_tokens": 435739985.0, + "step": 16855 + }, + { + "epoch": 1.8510871952558752, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1525003910064697, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7119417190551758, + "num_tokens": 435767221.0, + "step": 16856 + }, + { + "epoch": 1.851197012958489, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3700950145721436, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7447564601898193, + "num_tokens": 435790554.0, + "step": 16857 + }, + { + "epoch": 1.8513068306611027, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.187328338623047, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7275424003601074, + "num_tokens": 435817734.0, + "step": 16858 + }, + { + "epoch": 1.8514166483637162, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1377930641174316, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7042819857597351, + "num_tokens": 435845409.0, + "step": 16859 + }, + { + "epoch": 1.8515264660663298, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4775209426879883, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7236936092376709, + "num_tokens": 435867146.0, + "step": 16860 + }, + { + "epoch": 1.8516362837689435, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2569377422332764, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7142770886421204, + "num_tokens": 435896313.0, + "step": 16861 + }, + { + "epoch": 1.8517461014715573, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.6352996826171875, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7126464247703552, + "num_tokens": 435916873.0, + "step": 16862 + }, + { + "epoch": 1.851855919174171, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2106401920318604, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7350896596908569, + "num_tokens": 435940352.0, + "step": 16863 + }, + { + "epoch": 1.8519657368767846, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2823245525360107, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7315548658370972, + "num_tokens": 435965409.0, + "step": 16864 + }, + { + "epoch": 1.852075554579398, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.355508327484131, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7356758117675781, + "num_tokens": 435989123.0, + "step": 16865 + }, + { + "epoch": 1.8521853722820119, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.39100980758667, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7378430366516113, + "num_tokens": 436010896.0, + "step": 16866 + }, + { + "epoch": 1.8522951899846256, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.222989797592163, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7091597318649292, + "num_tokens": 436037438.0, + "step": 16867 + }, + { + "epoch": 1.8524050076872391, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.573784351348877, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7104273438453674, + "num_tokens": 436056645.0, + "step": 16868 + }, + { + "epoch": 1.8525148253898527, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.207777261734009, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7020303606987, + "num_tokens": 436084261.0, + "step": 16869 + }, + { + "epoch": 1.8526246430924664, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0641586780548096, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7098411321640015, + "num_tokens": 436115170.0, + "step": 16870 + }, + { + "epoch": 1.8527344607950802, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4209930896759033, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7231048345565796, + "num_tokens": 436138017.0, + "step": 16871 + }, + { + "epoch": 1.852844278497694, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2877304553985596, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7088958621025085, + "num_tokens": 436164826.0, + "step": 16872 + }, + { + "epoch": 1.8529540962003075, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4833295345306396, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7366352081298828, + "num_tokens": 436186950.0, + "step": 16873 + }, + { + "epoch": 1.853063913902921, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.071563243865967, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6999554634094238, + "num_tokens": 436218554.0, + "step": 16874 + }, + { + "epoch": 1.8531737316055348, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2398061752319336, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.6878116130828857, + "num_tokens": 436245732.0, + "step": 16875 + }, + { + "epoch": 1.8532835493081485, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0285050868988037, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7303215265274048, + "num_tokens": 436275907.0, + "step": 16876 + }, + { + "epoch": 1.8533933670107623, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3667984008789062, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7503930330276489, + "num_tokens": 436298559.0, + "step": 16877 + }, + { + "epoch": 1.8535031847133758, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.059234857559204, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7083790302276611, + "num_tokens": 436328711.0, + "step": 16878 + }, + { + "epoch": 1.8536130024159894, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.371696710586548, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7232468128204346, + "num_tokens": 436351002.0, + "step": 16879 + }, + { + "epoch": 1.8537228201186031, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.241020679473877, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7090773582458496, + "num_tokens": 436376844.0, + "step": 16880 + }, + { + "epoch": 1.8538326378212169, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2268521785736084, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6979003548622131, + "num_tokens": 436403455.0, + "step": 16881 + }, + { + "epoch": 1.8539424555238304, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.312580108642578, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7171611785888672, + "num_tokens": 436427524.0, + "step": 16882 + }, + { + "epoch": 1.854052273226444, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1487512588500977, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7299948930740356, + "num_tokens": 436454649.0, + "step": 16883 + }, + { + "epoch": 1.8541620909290577, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.357255220413208, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.734332799911499, + "num_tokens": 436477195.0, + "step": 16884 + }, + { + "epoch": 1.8542719086316715, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.045464277267456, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7190029621124268, + "num_tokens": 436505904.0, + "step": 16885 + }, + { + "epoch": 1.8543817263342852, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3783085346221924, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7183816432952881, + "num_tokens": 436531585.0, + "step": 16886 + }, + { + "epoch": 1.8544915440368988, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0433766841888428, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.7017744779586792, + "num_tokens": 436563798.0, + "step": 16887 + }, + { + "epoch": 1.8546013617395123, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0684521198272705, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7178256511688232, + "num_tokens": 436592580.0, + "step": 16888 + }, + { + "epoch": 1.854711179442126, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.271986961364746, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7338826060295105, + "num_tokens": 436615814.0, + "step": 16889 + }, + { + "epoch": 1.8548209971447398, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1822962760925293, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7251260876655579, + "num_tokens": 436642617.0, + "step": 16890 + }, + { + "epoch": 1.8549308148473536, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3499011993408203, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7112764120101929, + "num_tokens": 436666971.0, + "step": 16891 + }, + { + "epoch": 1.855040632549967, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.12302303314209, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7067825794219971, + "num_tokens": 436695065.0, + "step": 16892 + }, + { + "epoch": 1.8551504502525806, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0053956508636475, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7124900221824646, + "num_tokens": 436724658.0, + "step": 16893 + }, + { + "epoch": 1.8552602679551944, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2450239658355713, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7181110978126526, + "num_tokens": 436750097.0, + "step": 16894 + }, + { + "epoch": 1.8553700856578081, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2293431758880615, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7311962842941284, + "num_tokens": 436774023.0, + "step": 16895 + }, + { + "epoch": 1.8554799033604217, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.173192262649536, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7187621593475342, + "num_tokens": 436800885.0, + "step": 16896 + }, + { + "epoch": 1.8555897210630352, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.59121036529541, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7124209403991699, + "num_tokens": 436824212.0, + "step": 16897 + }, + { + "epoch": 1.855699538765649, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.6525814533233643, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7408538460731506, + "num_tokens": 436843050.0, + "step": 16898 + }, + { + "epoch": 1.8558093564682627, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9395301342010498, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7002902626991272, + "num_tokens": 436880926.0, + "step": 16899 + }, + { + "epoch": 1.8559191741708765, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1593070030212402, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7153480052947998, + "num_tokens": 436908418.0, + "step": 16900 + }, + { + "epoch": 1.85602899187349, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.283505916595459, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7394090294837952, + "num_tokens": 436933281.0, + "step": 16901 + }, + { + "epoch": 1.8561388095761036, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1178910732269287, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7188513278961182, + "num_tokens": 436959755.0, + "step": 16902 + }, + { + "epoch": 1.8562486272787173, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.237338066101074, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7486016750335693, + "num_tokens": 436983183.0, + "step": 16903 + }, + { + "epoch": 1.856358444981331, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.6437766551971436, + "learning_rate": 1e-06, + "loss": 0.7479, + "mean_token_accuracy": 0.761841893196106, + "num_tokens": 437000635.0, + "step": 16904 + }, + { + "epoch": 1.8564682626839446, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1769886016845703, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7128653526306152, + "num_tokens": 437025503.0, + "step": 16905 + }, + { + "epoch": 1.8565780803865584, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2506165504455566, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7217217683792114, + "num_tokens": 437051630.0, + "step": 16906 + }, + { + "epoch": 1.856687898089172, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4565491676330566, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7412041425704956, + "num_tokens": 437076770.0, + "step": 16907 + }, + { + "epoch": 1.8567977157917857, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.021867036819458, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7270044684410095, + "num_tokens": 437108400.0, + "step": 16908 + }, + { + "epoch": 1.8569075334943994, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0308995246887207, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7091056704521179, + "num_tokens": 437140555.0, + "step": 16909 + }, + { + "epoch": 1.857017351197013, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.469219446182251, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7231799960136414, + "num_tokens": 437162231.0, + "step": 16910 + }, + { + "epoch": 1.8571271688996265, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1410834789276123, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7129344940185547, + "num_tokens": 437188757.0, + "step": 16911 + }, + { + "epoch": 1.8572369866022402, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.089696168899536, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7449380159378052, + "num_tokens": 437215715.0, + "step": 16912 + }, + { + "epoch": 1.857346804304854, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.265103578567505, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7222315669059753, + "num_tokens": 437240293.0, + "step": 16913 + }, + { + "epoch": 1.8574566220074678, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0371415615081787, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7259578704833984, + "num_tokens": 437271806.0, + "step": 16914 + }, + { + "epoch": 1.8575664397100813, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2936694622039795, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7357346415519714, + "num_tokens": 437295395.0, + "step": 16915 + }, + { + "epoch": 1.8576762574126948, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.5506327152252197, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7187471389770508, + "num_tokens": 437316129.0, + "step": 16916 + }, + { + "epoch": 1.8577860751153086, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2237746715545654, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.700890302658081, + "num_tokens": 437342790.0, + "step": 16917 + }, + { + "epoch": 1.8578958928179223, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1089963912963867, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.6993833780288696, + "num_tokens": 437372755.0, + "step": 16918 + }, + { + "epoch": 1.8580057105205359, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9251222610473633, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7166854739189148, + "num_tokens": 437406550.0, + "step": 16919 + }, + { + "epoch": 1.8581155282231496, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3318753242492676, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7455599308013916, + "num_tokens": 437428676.0, + "step": 16920 + }, + { + "epoch": 1.8582253459257632, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.210054874420166, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7041015028953552, + "num_tokens": 437455018.0, + "step": 16921 + }, + { + "epoch": 1.858335163628377, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3820362091064453, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.7053025960922241, + "num_tokens": 437479847.0, + "step": 16922 + }, + { + "epoch": 1.8584449813309907, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2715771198272705, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7074998617172241, + "num_tokens": 437507113.0, + "step": 16923 + }, + { + "epoch": 1.8585547990336042, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.4449193477630615, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7063307166099548, + "num_tokens": 437529316.0, + "step": 16924 + }, + { + "epoch": 1.8586646167362177, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.037346363067627, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6915920376777649, + "num_tokens": 437560522.0, + "step": 16925 + }, + { + "epoch": 1.8587744344388315, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0370686054229736, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6986225843429565, + "num_tokens": 437590821.0, + "step": 16926 + }, + { + "epoch": 1.8588842521414453, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.241934299468994, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.701786994934082, + "num_tokens": 437619416.0, + "step": 16927 + }, + { + "epoch": 1.858994069844059, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3808162212371826, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7031293511390686, + "num_tokens": 437643616.0, + "step": 16928 + }, + { + "epoch": 1.8591038875466726, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.147742986679077, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7284941077232361, + "num_tokens": 437670853.0, + "step": 16929 + }, + { + "epoch": 1.859213705249286, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.465585947036743, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7268667221069336, + "num_tokens": 437692695.0, + "step": 16930 + }, + { + "epoch": 1.8593235229518998, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.126516342163086, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7249590158462524, + "num_tokens": 437719371.0, + "step": 16931 + }, + { + "epoch": 1.8594333406545136, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.222344160079956, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7334139347076416, + "num_tokens": 437743999.0, + "step": 16932 + }, + { + "epoch": 1.8595431583571271, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3415372371673584, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7119945287704468, + "num_tokens": 437766725.0, + "step": 16933 + }, + { + "epoch": 1.8596529760597407, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.309393882751465, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7463664412498474, + "num_tokens": 437790488.0, + "step": 16934 + }, + { + "epoch": 1.8597627937623544, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3047540187835693, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.713823676109314, + "num_tokens": 437814868.0, + "step": 16935 + }, + { + "epoch": 1.8598726114649682, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.490910053253174, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7288452982902527, + "num_tokens": 437834285.0, + "step": 16936 + }, + { + "epoch": 1.859982429167582, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0491602420806885, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.6971297860145569, + "num_tokens": 437862652.0, + "step": 16937 + }, + { + "epoch": 1.8600922468701955, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2335150241851807, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7169168591499329, + "num_tokens": 437886820.0, + "step": 16938 + }, + { + "epoch": 1.860202064572809, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.01828932762146, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7274459600448608, + "num_tokens": 437919256.0, + "step": 16939 + }, + { + "epoch": 1.8603118822754228, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1323647499084473, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7015633583068848, + "num_tokens": 437947549.0, + "step": 16940 + }, + { + "epoch": 1.8604216999780365, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2667946815490723, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.705974280834198, + "num_tokens": 437972843.0, + "step": 16941 + }, + { + "epoch": 1.8605315176806503, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.292658567428589, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7293798923492432, + "num_tokens": 437998600.0, + "step": 16942 + }, + { + "epoch": 1.8606413353832638, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.033482551574707, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7200574278831482, + "num_tokens": 438028102.0, + "step": 16943 + }, + { + "epoch": 1.8607511530858774, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.341931104660034, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7065184116363525, + "num_tokens": 438052822.0, + "step": 16944 + }, + { + "epoch": 1.860860970788491, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.105205535888672, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7104438543319702, + "num_tokens": 438082844.0, + "step": 16945 + }, + { + "epoch": 1.8609707884911049, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2601513862609863, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7186188101768494, + "num_tokens": 438109890.0, + "step": 16946 + }, + { + "epoch": 1.8610806061937184, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9322094917297363, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7168904542922974, + "num_tokens": 438140895.0, + "step": 16947 + }, + { + "epoch": 1.861190423896332, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2662534713745117, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7334038615226746, + "num_tokens": 438165806.0, + "step": 16948 + }, + { + "epoch": 1.8613002415989457, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3470962047576904, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7319779396057129, + "num_tokens": 438190317.0, + "step": 16949 + }, + { + "epoch": 1.8614100593015594, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1556220054626465, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.717430830001831, + "num_tokens": 438215540.0, + "step": 16950 + }, + { + "epoch": 1.8615198770041732, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2790722846984863, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7332751750946045, + "num_tokens": 438239349.0, + "step": 16951 + }, + { + "epoch": 1.8616296947067867, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9598109722137451, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.70267653465271, + "num_tokens": 438272407.0, + "step": 16952 + }, + { + "epoch": 1.8617395124094003, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.5332276821136475, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.756072998046875, + "num_tokens": 438292060.0, + "step": 16953 + }, + { + "epoch": 1.861849330112014, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9881932735443115, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7046641111373901, + "num_tokens": 438323163.0, + "step": 16954 + }, + { + "epoch": 1.8619591478146278, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.21362042427063, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6980314254760742, + "num_tokens": 438349393.0, + "step": 16955 + }, + { + "epoch": 1.8620689655172413, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2369792461395264, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7224607467651367, + "num_tokens": 438374889.0, + "step": 16956 + }, + { + "epoch": 1.862178783219855, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1379590034484863, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7057695984840393, + "num_tokens": 438404116.0, + "step": 16957 + }, + { + "epoch": 1.8622886009224686, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.295518159866333, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7174062728881836, + "num_tokens": 438428559.0, + "step": 16958 + }, + { + "epoch": 1.8623984186250824, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.401479482650757, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7068865299224854, + "num_tokens": 438452106.0, + "step": 16959 + }, + { + "epoch": 1.8625082363276961, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2844409942626953, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7150609493255615, + "num_tokens": 438476888.0, + "step": 16960 + }, + { + "epoch": 1.8626180540303097, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9769338369369507, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6951198577880859, + "num_tokens": 438509799.0, + "step": 16961 + }, + { + "epoch": 1.8627278717329232, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.269763469696045, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7221141457557678, + "num_tokens": 438533819.0, + "step": 16962 + }, + { + "epoch": 1.862837689435537, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1840147972106934, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.70778489112854, + "num_tokens": 438559396.0, + "step": 16963 + }, + { + "epoch": 1.8629475071381507, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.2864227294921875, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7524471879005432, + "num_tokens": 438581493.0, + "step": 16964 + }, + { + "epoch": 1.8630573248407645, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.3990960121154785, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7406305074691772, + "num_tokens": 438603969.0, + "step": 16965 + }, + { + "epoch": 1.863167142543378, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.1798579692840576, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.733830451965332, + "num_tokens": 438629259.0, + "step": 16966 + }, + { + "epoch": 1.8632769602459915, + "ewc_loss": 0.00014495849609375, + "grad_norm": 2.0351736545562744, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6867661476135254, + "num_tokens": 438659759.0, + "step": 16967 + }, + { + "epoch": 1.8633867779486053, + "ewc_loss": 0.00014495849609375, + "grad_norm": 1.9138691425323486, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.6944060921669006, + "num_tokens": 438693993.0, + "step": 16968 + }, + { + "epoch": 1.863496595651219, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.33786940574646, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7317893505096436, + "num_tokens": 438717877.0, + "step": 16969 + }, + { + "epoch": 1.8636064133538326, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.290663242340088, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7196341156959534, + "num_tokens": 438741555.0, + "step": 16970 + }, + { + "epoch": 1.8637162310564463, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.133171319961548, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7088630795478821, + "num_tokens": 438772483.0, + "step": 16971 + }, + { + "epoch": 1.8638260487590599, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.289292097091675, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7359304428100586, + "num_tokens": 438796946.0, + "step": 16972 + }, + { + "epoch": 1.8639358664616736, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9699331521987915, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.706575870513916, + "num_tokens": 438828630.0, + "step": 16973 + }, + { + "epoch": 1.8640456841642874, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.311446189880371, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7243533134460449, + "num_tokens": 438854638.0, + "step": 16974 + }, + { + "epoch": 1.864155501866901, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.211404323577881, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7323017716407776, + "num_tokens": 438880277.0, + "step": 16975 + }, + { + "epoch": 1.8642653195695145, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9035158157348633, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7073276042938232, + "num_tokens": 438913785.0, + "step": 16976 + }, + { + "epoch": 1.8643751372721282, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.510474920272827, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7286546230316162, + "num_tokens": 438935297.0, + "step": 16977 + }, + { + "epoch": 1.864484954974742, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1816904544830322, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7256818413734436, + "num_tokens": 438960935.0, + "step": 16978 + }, + { + "epoch": 1.8645947726773557, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2319133281707764, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7253937125205994, + "num_tokens": 438985536.0, + "step": 16979 + }, + { + "epoch": 1.8647045903799693, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.181732654571533, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6970919370651245, + "num_tokens": 439014239.0, + "step": 16980 + }, + { + "epoch": 1.8648144080825828, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2711052894592285, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7203736305236816, + "num_tokens": 439039532.0, + "step": 16981 + }, + { + "epoch": 1.8649242257851966, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0220987796783447, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.719204306602478, + "num_tokens": 439070929.0, + "step": 16982 + }, + { + "epoch": 1.8650340434878103, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3497462272644043, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7235027551651001, + "num_tokens": 439093825.0, + "step": 16983 + }, + { + "epoch": 1.8651438611904239, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.404050350189209, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7105189561843872, + "num_tokens": 439117753.0, + "step": 16984 + }, + { + "epoch": 1.8652536788930374, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1149604320526123, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7507719397544861, + "num_tokens": 439144930.0, + "step": 16985 + }, + { + "epoch": 1.8653634965956511, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.644223690032959, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7209193706512451, + "num_tokens": 439165820.0, + "step": 16986 + }, + { + "epoch": 1.865473314298265, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.376417875289917, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7410950660705566, + "num_tokens": 439188198.0, + "step": 16987 + }, + { + "epoch": 1.8655831320008787, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0766284465789795, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7315086126327515, + "num_tokens": 439215580.0, + "step": 16988 + }, + { + "epoch": 1.8656929497034922, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.238203525543213, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7299436330795288, + "num_tokens": 439238300.0, + "step": 16989 + }, + { + "epoch": 1.8658027674061057, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2585692405700684, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7095718383789062, + "num_tokens": 439263492.0, + "step": 16990 + }, + { + "epoch": 1.8659125851087195, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.312053918838501, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.715791642665863, + "num_tokens": 439290625.0, + "step": 16991 + }, + { + "epoch": 1.8660224028113332, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.346967935562134, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6941994428634644, + "num_tokens": 439316988.0, + "step": 16992 + }, + { + "epoch": 1.866132220513947, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.4365463256835938, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7282031774520874, + "num_tokens": 439338441.0, + "step": 16993 + }, + { + "epoch": 1.8662420382165605, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.6123855113983154, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7319889068603516, + "num_tokens": 439359117.0, + "step": 16994 + }, + { + "epoch": 1.866351855919174, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3140029907226562, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7114516496658325, + "num_tokens": 439383259.0, + "step": 16995 + }, + { + "epoch": 1.8664616736217878, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.237589120864868, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7149903178215027, + "num_tokens": 439407786.0, + "step": 16996 + }, + { + "epoch": 1.8665714913244016, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.026989698410034, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7118473052978516, + "num_tokens": 439437130.0, + "step": 16997 + }, + { + "epoch": 1.8666813090270151, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0636918544769287, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7349543571472168, + "num_tokens": 439465442.0, + "step": 16998 + }, + { + "epoch": 1.8667911267296287, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.126427173614502, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7141551971435547, + "num_tokens": 439494648.0, + "step": 16999 + }, + { + "epoch": 1.8669009444322424, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.212475538253784, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7250632643699646, + "num_tokens": 439520459.0, + "step": 17000 + }, + { + "epoch": 1.8670107621348562, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.294748067855835, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7097828388214111, + "num_tokens": 439546501.0, + "step": 17001 + }, + { + "epoch": 1.86712057983747, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0245423316955566, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7235458493232727, + "num_tokens": 439576103.0, + "step": 17002 + }, + { + "epoch": 1.8672303975400835, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.592813491821289, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7153395414352417, + "num_tokens": 439597541.0, + "step": 17003 + }, + { + "epoch": 1.867340215242697, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.105137348175049, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7270229458808899, + "num_tokens": 439627409.0, + "step": 17004 + }, + { + "epoch": 1.8674500329453108, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.409877300262451, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7157303690910339, + "num_tokens": 439649990.0, + "step": 17005 + }, + { + "epoch": 1.8675598506479245, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.278930187225342, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7272146344184875, + "num_tokens": 439676100.0, + "step": 17006 + }, + { + "epoch": 1.8676696683505383, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2730960845947266, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7065625190734863, + "num_tokens": 439701954.0, + "step": 17007 + }, + { + "epoch": 1.8677794860531518, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2010793685913086, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7219088077545166, + "num_tokens": 439729123.0, + "step": 17008 + }, + { + "epoch": 1.8678893037557653, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.256546974182129, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.7007327675819397, + "num_tokens": 439755615.0, + "step": 17009 + }, + { + "epoch": 1.867999121458379, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0490870475769043, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7095708847045898, + "num_tokens": 439786232.0, + "step": 17010 + }, + { + "epoch": 1.8681089391609929, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.283569097518921, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7252712249755859, + "num_tokens": 439810890.0, + "step": 17011 + }, + { + "epoch": 1.8682187568636064, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.4253268241882324, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7165800333023071, + "num_tokens": 439833482.0, + "step": 17012 + }, + { + "epoch": 1.86832857456622, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.136486530303955, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7276686429977417, + "num_tokens": 439859092.0, + "step": 17013 + }, + { + "epoch": 1.8684383922688337, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.09511661529541, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7222887277603149, + "num_tokens": 439886906.0, + "step": 17014 + }, + { + "epoch": 1.8685482099714474, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.235546588897705, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7052112817764282, + "num_tokens": 439912722.0, + "step": 17015 + }, + { + "epoch": 1.8686580276740612, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.412774085998535, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7326959371566772, + "num_tokens": 439936368.0, + "step": 17016 + }, + { + "epoch": 1.8687678453766747, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.374476909637451, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.723967969417572, + "num_tokens": 439961316.0, + "step": 17017 + }, + { + "epoch": 1.8688776630792883, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2140705585479736, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6984020471572876, + "num_tokens": 439989532.0, + "step": 17018 + }, + { + "epoch": 1.868987480781902, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.226421594619751, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7151468992233276, + "num_tokens": 440016373.0, + "step": 17019 + }, + { + "epoch": 1.8690972984845158, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1648011207580566, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.739852786064148, + "num_tokens": 440041652.0, + "step": 17020 + }, + { + "epoch": 1.8692071161871293, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.286250591278076, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7137071490287781, + "num_tokens": 440065463.0, + "step": 17021 + }, + { + "epoch": 1.869316933889743, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.183391571044922, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7040552496910095, + "num_tokens": 440092608.0, + "step": 17022 + }, + { + "epoch": 1.8694267515923566, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.009254217147827, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.716964840888977, + "num_tokens": 440122046.0, + "step": 17023 + }, + { + "epoch": 1.8695365692949704, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9765446186065674, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7284688949584961, + "num_tokens": 440152237.0, + "step": 17024 + }, + { + "epoch": 1.8696463869975841, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1288235187530518, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7066291570663452, + "num_tokens": 440179011.0, + "step": 17025 + }, + { + "epoch": 1.8697562047001977, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.247166156768799, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7303636074066162, + "num_tokens": 440202428.0, + "step": 17026 + }, + { + "epoch": 1.8698660224028112, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3110005855560303, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7016385197639465, + "num_tokens": 440228411.0, + "step": 17027 + }, + { + "epoch": 1.869975840105425, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2106196880340576, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7255983352661133, + "num_tokens": 440253868.0, + "step": 17028 + }, + { + "epoch": 1.8700856578080387, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2316739559173584, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7280267477035522, + "num_tokens": 440276775.0, + "step": 17029 + }, + { + "epoch": 1.8701954755106525, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0859460830688477, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7113374471664429, + "num_tokens": 440304245.0, + "step": 17030 + }, + { + "epoch": 1.870305293213266, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3271069526672363, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7243242859840393, + "num_tokens": 440328836.0, + "step": 17031 + }, + { + "epoch": 1.8704151109158795, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.423421859741211, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7240156531333923, + "num_tokens": 440350971.0, + "step": 17032 + }, + { + "epoch": 1.8705249286184933, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1693997383117676, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6787396669387817, + "num_tokens": 440377766.0, + "step": 17033 + }, + { + "epoch": 1.870634746321107, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9980143308639526, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.739654541015625, + "num_tokens": 440407060.0, + "step": 17034 + }, + { + "epoch": 1.8707445640237206, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0902576446533203, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7031130194664001, + "num_tokens": 440433321.0, + "step": 17035 + }, + { + "epoch": 1.8708543817263343, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.077730894088745, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7258966565132141, + "num_tokens": 440461191.0, + "step": 17036 + }, + { + "epoch": 1.8709641994289479, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2721829414367676, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6977976560592651, + "num_tokens": 440485817.0, + "step": 17037 + }, + { + "epoch": 1.8710740171315616, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0787081718444824, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7174268960952759, + "num_tokens": 440512880.0, + "step": 17038 + }, + { + "epoch": 1.8711838348341754, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.027820348739624, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7262310981750488, + "num_tokens": 440543790.0, + "step": 17039 + }, + { + "epoch": 1.871293652536789, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3000600337982178, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7103644609451294, + "num_tokens": 440568853.0, + "step": 17040 + }, + { + "epoch": 1.8714034702394025, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0202739238739014, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7036491632461548, + "num_tokens": 440599282.0, + "step": 17041 + }, + { + "epoch": 1.8715132879420162, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0294346809387207, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7483729720115662, + "num_tokens": 440629140.0, + "step": 17042 + }, + { + "epoch": 1.87162310564463, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.270555019378662, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7195336818695068, + "num_tokens": 440653455.0, + "step": 17043 + }, + { + "epoch": 1.8717329233472437, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1827621459960938, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7244636416435242, + "num_tokens": 440680585.0, + "step": 17044 + }, + { + "epoch": 1.8718427410498573, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.479306697845459, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7249210476875305, + "num_tokens": 440703091.0, + "step": 17045 + }, + { + "epoch": 1.8719525587524708, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.266916275024414, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7218945026397705, + "num_tokens": 440729516.0, + "step": 17046 + }, + { + "epoch": 1.8720623764550846, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9523109197616577, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7276052832603455, + "num_tokens": 440764333.0, + "step": 17047 + }, + { + "epoch": 1.8721721941576983, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0110301971435547, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7061401605606079, + "num_tokens": 440794112.0, + "step": 17048 + }, + { + "epoch": 1.8722820118603118, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.36146879196167, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7370844483375549, + "num_tokens": 440816068.0, + "step": 17049 + }, + { + "epoch": 1.8723918295629254, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.397674322128296, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7308074235916138, + "num_tokens": 440839606.0, + "step": 17050 + }, + { + "epoch": 1.8725016472655391, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1585423946380615, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7072904109954834, + "num_tokens": 440867155.0, + "step": 17051 + }, + { + "epoch": 1.872611464968153, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1339480876922607, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.714296817779541, + "num_tokens": 440892677.0, + "step": 17052 + }, + { + "epoch": 1.8727212826707667, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.169667959213257, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7105358242988586, + "num_tokens": 440920143.0, + "step": 17053 + }, + { + "epoch": 1.8728311003733802, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 3.005195379257202, + "learning_rate": 1e-06, + "loss": 0.7275, + "mean_token_accuracy": 0.7667833566665649, + "num_tokens": 440940554.0, + "step": 17054 + }, + { + "epoch": 1.8729409180759937, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0423662662506104, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7329549789428711, + "num_tokens": 440971347.0, + "step": 17055 + }, + { + "epoch": 1.8730507357786075, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2464420795440674, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7151328921318054, + "num_tokens": 440997428.0, + "step": 17056 + }, + { + "epoch": 1.8731605534812212, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.26362681388855, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7315118908882141, + "num_tokens": 441020541.0, + "step": 17057 + }, + { + "epoch": 1.873270371183835, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.191913604736328, + "learning_rate": 1e-06, + "loss": 0.7613, + "mean_token_accuracy": 0.7599849104881287, + "num_tokens": 441044714.0, + "step": 17058 + }, + { + "epoch": 1.8733801888864485, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.244976043701172, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7315723299980164, + "num_tokens": 441068532.0, + "step": 17059 + }, + { + "epoch": 1.873490006589062, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.479797840118408, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7476745843887329, + "num_tokens": 441091202.0, + "step": 17060 + }, + { + "epoch": 1.8735998242916758, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.5491766929626465, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7146208882331848, + "num_tokens": 441112498.0, + "step": 17061 + }, + { + "epoch": 1.8737096419942896, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.121394634246826, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7116299867630005, + "num_tokens": 441142109.0, + "step": 17062 + }, + { + "epoch": 1.8738194596969031, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.7397890090942383, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7010985612869263, + "num_tokens": 441164003.0, + "step": 17063 + }, + { + "epoch": 1.8739292773995166, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1987462043762207, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6956679821014404, + "num_tokens": 441189905.0, + "step": 17064 + }, + { + "epoch": 1.8740390951021304, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2588093280792236, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.722188413143158, + "num_tokens": 441215805.0, + "step": 17065 + }, + { + "epoch": 1.8741489128047442, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.126939535140991, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7076336145401001, + "num_tokens": 441246601.0, + "step": 17066 + }, + { + "epoch": 1.874258730507358, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.050018787384033, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7312031388282776, + "num_tokens": 441274797.0, + "step": 17067 + }, + { + "epoch": 1.8743685482099715, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2796459197998047, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7046759128570557, + "num_tokens": 441301370.0, + "step": 17068 + }, + { + "epoch": 1.874478365912585, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.48709774017334, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7242894172668457, + "num_tokens": 441323660.0, + "step": 17069 + }, + { + "epoch": 1.8745881836151987, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.106919288635254, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7111949324607849, + "num_tokens": 441352209.0, + "step": 17070 + }, + { + "epoch": 1.8746980013178125, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9477177858352661, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7178327441215515, + "num_tokens": 441383721.0, + "step": 17071 + }, + { + "epoch": 1.8748078190204263, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.201111316680908, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7101908922195435, + "num_tokens": 441409299.0, + "step": 17072 + }, + { + "epoch": 1.8749176367230398, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2450637817382812, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7190532684326172, + "num_tokens": 441434302.0, + "step": 17073 + }, + { + "epoch": 1.8750274544256533, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1349782943725586, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7208091020584106, + "num_tokens": 441461110.0, + "step": 17074 + }, + { + "epoch": 1.875137272128267, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.905326008796692, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7179484963417053, + "num_tokens": 441493487.0, + "step": 17075 + }, + { + "epoch": 1.8752470898308808, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1966450214385986, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7450429201126099, + "num_tokens": 441518779.0, + "step": 17076 + }, + { + "epoch": 1.8753569075334944, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3062899112701416, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7408185005187988, + "num_tokens": 441542646.0, + "step": 17077 + }, + { + "epoch": 1.875466725236108, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 1.9016472101211548, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7242823839187622, + "num_tokens": 441573331.0, + "step": 17078 + }, + { + "epoch": 1.8755765429387217, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0782992839813232, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7307569980621338, + "num_tokens": 441602259.0, + "step": 17079 + }, + { + "epoch": 1.8756863606413354, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.026271104812622, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6964536905288696, + "num_tokens": 441634513.0, + "step": 17080 + }, + { + "epoch": 1.8757961783439492, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.388286828994751, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7253984808921814, + "num_tokens": 441657125.0, + "step": 17081 + }, + { + "epoch": 1.8759059960465627, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.169889211654663, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7123733758926392, + "num_tokens": 441683874.0, + "step": 17082 + }, + { + "epoch": 1.8760158137491763, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2046475410461426, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.699244499206543, + "num_tokens": 441710732.0, + "step": 17083 + }, + { + "epoch": 1.87612563145179, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.592790126800537, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7323349118232727, + "num_tokens": 441736131.0, + "step": 17084 + }, + { + "epoch": 1.8762354491544038, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.450502872467041, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7147932052612305, + "num_tokens": 441758858.0, + "step": 17085 + }, + { + "epoch": 1.8763452668570173, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.3491575717926025, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7285800576210022, + "num_tokens": 441782241.0, + "step": 17086 + }, + { + "epoch": 1.876455084559631, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.29536771774292, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7315987348556519, + "num_tokens": 441805904.0, + "step": 17087 + }, + { + "epoch": 1.8765649022622446, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.571418523788452, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7292487025260925, + "num_tokens": 441826314.0, + "step": 17088 + }, + { + "epoch": 1.8766747199648584, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.311224937438965, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7189632654190063, + "num_tokens": 441852105.0, + "step": 17089 + }, + { + "epoch": 1.876784537667472, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0206379890441895, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7106528282165527, + "num_tokens": 441882518.0, + "step": 17090 + }, + { + "epoch": 1.8768943553700856, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.5384035110473633, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7070542573928833, + "num_tokens": 441905581.0, + "step": 17091 + }, + { + "epoch": 1.8770041730726992, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.800652265548706, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7497434616088867, + "num_tokens": 441921785.0, + "step": 17092 + }, + { + "epoch": 1.877113990775313, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2814202308654785, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.735343337059021, + "num_tokens": 441944956.0, + "step": 17093 + }, + { + "epoch": 1.8772238084779267, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.415494680404663, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7250326871871948, + "num_tokens": 441967212.0, + "step": 17094 + }, + { + "epoch": 1.8773336261805404, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.0065500736236572, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7151659727096558, + "num_tokens": 441997960.0, + "step": 17095 + }, + { + "epoch": 1.877443443883154, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2719950675964355, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7143182158470154, + "num_tokens": 442022187.0, + "step": 17096 + }, + { + "epoch": 1.8775532615857675, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.161339044570923, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.6996841430664062, + "num_tokens": 442050708.0, + "step": 17097 + }, + { + "epoch": 1.8776630792883813, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.2402734756469727, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7440757751464844, + "num_tokens": 442075492.0, + "step": 17098 + }, + { + "epoch": 1.877772896990995, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.166581869125366, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.701285183429718, + "num_tokens": 442102548.0, + "step": 17099 + }, + { + "epoch": 1.8778827146936086, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.1425669193267822, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7111728191375732, + "num_tokens": 442130805.0, + "step": 17100 + }, + { + "epoch": 1.8779925323962223, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4527924060821533, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7106025218963623, + "num_tokens": 442154279.0, + "step": 17101 + }, + { + "epoch": 1.8781023500988359, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1658546924591064, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.714515209197998, + "num_tokens": 442182025.0, + "step": 17102 + }, + { + "epoch": 1.8782121678014496, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3673782348632812, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7322891354560852, + "num_tokens": 442205503.0, + "step": 17103 + }, + { + "epoch": 1.8783219855040634, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.333583354949951, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7337174415588379, + "num_tokens": 442228249.0, + "step": 17104 + }, + { + "epoch": 1.878431803206677, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.436920404434204, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7203714847564697, + "num_tokens": 442249551.0, + "step": 17105 + }, + { + "epoch": 1.8785416209092904, + "ewc_loss": 0.0001468658447265625, + "grad_norm": 2.370818614959717, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7330166101455688, + "num_tokens": 442271194.0, + "step": 17106 + }, + { + "epoch": 1.8786514386119042, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1508114337921143, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7214838266372681, + "num_tokens": 442300042.0, + "step": 17107 + }, + { + "epoch": 1.878761256314518, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2624928951263428, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7064483165740967, + "num_tokens": 442328394.0, + "step": 17108 + }, + { + "epoch": 1.8788710740171317, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.03908634185791, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7003812789916992, + "num_tokens": 442360783.0, + "step": 17109 + }, + { + "epoch": 1.8789808917197452, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.314393997192383, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7184739708900452, + "num_tokens": 442385450.0, + "step": 17110 + }, + { + "epoch": 1.8790907094223588, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.252845287322998, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7125898599624634, + "num_tokens": 442410895.0, + "step": 17111 + }, + { + "epoch": 1.8792005271249725, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3045763969421387, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7392117381095886, + "num_tokens": 442433978.0, + "step": 17112 + }, + { + "epoch": 1.8793103448275863, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2907216548919678, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7156744599342346, + "num_tokens": 442459487.0, + "step": 17113 + }, + { + "epoch": 1.8794201625301998, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2091290950775146, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7168911099433899, + "num_tokens": 442488748.0, + "step": 17114 + }, + { + "epoch": 1.8795299802328134, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1453659534454346, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7378178834915161, + "num_tokens": 442513277.0, + "step": 17115 + }, + { + "epoch": 1.8796397979354271, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.361955165863037, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.742621660232544, + "num_tokens": 442535248.0, + "step": 17116 + }, + { + "epoch": 1.8797496156380409, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0173230171203613, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7192200422286987, + "num_tokens": 442565238.0, + "step": 17117 + }, + { + "epoch": 1.8798594333406546, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1539244651794434, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7305330634117126, + "num_tokens": 442592048.0, + "step": 17118 + }, + { + "epoch": 1.8799692510432682, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1745402812957764, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.7071132063865662, + "num_tokens": 442619496.0, + "step": 17119 + }, + { + "epoch": 1.8800790687458817, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.271127462387085, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7154832482337952, + "num_tokens": 442644476.0, + "step": 17120 + }, + { + "epoch": 1.8801888864484955, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.246499538421631, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7071611881256104, + "num_tokens": 442670967.0, + "step": 17121 + }, + { + "epoch": 1.8802987041511092, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3362507820129395, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7183902263641357, + "num_tokens": 442694055.0, + "step": 17122 + }, + { + "epoch": 1.880408521853723, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3054745197296143, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6980335712432861, + "num_tokens": 442719608.0, + "step": 17123 + }, + { + "epoch": 1.8805183395563365, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.214106321334839, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7183132171630859, + "num_tokens": 442745522.0, + "step": 17124 + }, + { + "epoch": 1.88062815725895, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9776109457015991, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7444092035293579, + "num_tokens": 442774483.0, + "step": 17125 + }, + { + "epoch": 1.8807379749615638, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3561959266662598, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7227063179016113, + "num_tokens": 442799053.0, + "step": 17126 + }, + { + "epoch": 1.8808477926641776, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.949796199798584, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7013226747512817, + "num_tokens": 442830422.0, + "step": 17127 + }, + { + "epoch": 1.880957610366791, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.144892930984497, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7064473628997803, + "num_tokens": 442858114.0, + "step": 17128 + }, + { + "epoch": 1.8810674280694046, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.213642120361328, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.6978158950805664, + "num_tokens": 442885656.0, + "step": 17129 + }, + { + "epoch": 1.8811772457720184, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9277297258377075, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7353270649909973, + "num_tokens": 442915149.0, + "step": 17130 + }, + { + "epoch": 1.8812870634746321, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2131600379943848, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7196713089942932, + "num_tokens": 442940307.0, + "step": 17131 + }, + { + "epoch": 1.881396881177246, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.95063316822052, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7147325873374939, + "num_tokens": 442971327.0, + "step": 17132 + }, + { + "epoch": 1.8815066988798594, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.494755506515503, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7261946201324463, + "num_tokens": 442993456.0, + "step": 17133 + }, + { + "epoch": 1.881616516582473, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.079023599624634, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7121145725250244, + "num_tokens": 443022173.0, + "step": 17134 + }, + { + "epoch": 1.8817263342850867, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.196956157684326, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7304300665855408, + "num_tokens": 443048031.0, + "step": 17135 + }, + { + "epoch": 1.8818361519877005, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5922980308532715, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7237329483032227, + "num_tokens": 443068998.0, + "step": 17136 + }, + { + "epoch": 1.881945969690314, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.43807053565979, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7047790884971619, + "num_tokens": 443090374.0, + "step": 17137 + }, + { + "epoch": 1.8820557873929278, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2211763858795166, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.72874516248703, + "num_tokens": 443116477.0, + "step": 17138 + }, + { + "epoch": 1.8821656050955413, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.235567808151245, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.744265079498291, + "num_tokens": 443142580.0, + "step": 17139 + }, + { + "epoch": 1.882275422798155, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.490511417388916, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7178460359573364, + "num_tokens": 443165982.0, + "step": 17140 + }, + { + "epoch": 1.8823852405007688, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9782764911651611, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7281538248062134, + "num_tokens": 443197490.0, + "step": 17141 + }, + { + "epoch": 1.8824950582033824, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5288517475128174, + "learning_rate": 1e-06, + "loss": 0.7906, + "mean_token_accuracy": 0.7500852346420288, + "num_tokens": 443216578.0, + "step": 17142 + }, + { + "epoch": 1.882604875905996, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.044440746307373, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7353206872940063, + "num_tokens": 443246303.0, + "step": 17143 + }, + { + "epoch": 1.8827146936086097, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2544124126434326, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7154324054718018, + "num_tokens": 443272511.0, + "step": 17144 + }, + { + "epoch": 1.8828245113112234, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2432501316070557, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7128152847290039, + "num_tokens": 443297898.0, + "step": 17145 + }, + { + "epoch": 1.8829343290138372, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3541038036346436, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7269008159637451, + "num_tokens": 443320756.0, + "step": 17146 + }, + { + "epoch": 1.8830441467164507, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.31907320022583, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7143414616584778, + "num_tokens": 443344590.0, + "step": 17147 + }, + { + "epoch": 1.8831539644190642, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.184922933578491, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.706190824508667, + "num_tokens": 443372794.0, + "step": 17148 + }, + { + "epoch": 1.883263782121678, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2307779788970947, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.702027440071106, + "num_tokens": 443401270.0, + "step": 17149 + }, + { + "epoch": 1.8833735998242918, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.973433494567871, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7106591463088989, + "num_tokens": 443432763.0, + "step": 17150 + }, + { + "epoch": 1.8834834175269053, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0692367553710938, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7031614780426025, + "num_tokens": 443462146.0, + "step": 17151 + }, + { + "epoch": 1.883593235229519, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.105234384536743, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.740739107131958, + "num_tokens": 443489265.0, + "step": 17152 + }, + { + "epoch": 1.8837030529321326, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0790863037109375, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7099610567092896, + "num_tokens": 443521043.0, + "step": 17153 + }, + { + "epoch": 1.8838128706347463, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1633572578430176, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7220412492752075, + "num_tokens": 443548868.0, + "step": 17154 + }, + { + "epoch": 1.88392268833736, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2742295265197754, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.7571924924850464, + "num_tokens": 443572380.0, + "step": 17155 + }, + { + "epoch": 1.8840325060399736, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3646864891052246, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7238125205039978, + "num_tokens": 443595791.0, + "step": 17156 + }, + { + "epoch": 1.8841423237425872, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0928385257720947, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7109134197235107, + "num_tokens": 443623415.0, + "step": 17157 + }, + { + "epoch": 1.884252141445201, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0960683822631836, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7129274010658264, + "num_tokens": 443653345.0, + "step": 17158 + }, + { + "epoch": 1.8843619591478147, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2067177295684814, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7307763695716858, + "num_tokens": 443681577.0, + "step": 17159 + }, + { + "epoch": 1.8844717768504284, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.242666006088257, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7241090536117554, + "num_tokens": 443707466.0, + "step": 17160 + }, + { + "epoch": 1.884581594553042, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.361069679260254, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7149183750152588, + "num_tokens": 443730996.0, + "step": 17161 + }, + { + "epoch": 1.8846914122556555, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4198858737945557, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7088580131530762, + "num_tokens": 443754314.0, + "step": 17162 + }, + { + "epoch": 1.8848012299582693, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.473445415496826, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7301866412162781, + "num_tokens": 443775221.0, + "step": 17163 + }, + { + "epoch": 1.884911047660883, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3333284854888916, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6984084844589233, + "num_tokens": 443799455.0, + "step": 17164 + }, + { + "epoch": 1.8850208653634966, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3070647716522217, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7065149545669556, + "num_tokens": 443823852.0, + "step": 17165 + }, + { + "epoch": 1.88513068306611, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2045557498931885, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.698095977306366, + "num_tokens": 443850175.0, + "step": 17166 + }, + { + "epoch": 1.8852405007687238, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3697738647460938, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7325719594955444, + "num_tokens": 443874583.0, + "step": 17167 + }, + { + "epoch": 1.8853503184713376, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.589282274246216, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.738274097442627, + "num_tokens": 443894730.0, + "step": 17168 + }, + { + "epoch": 1.8854601361739514, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.6350619792938232, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.744705319404602, + "num_tokens": 443914815.0, + "step": 17169 + }, + { + "epoch": 1.885569953876565, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.996686577796936, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6929250359535217, + "num_tokens": 443946817.0, + "step": 17170 + }, + { + "epoch": 1.8856797715791784, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2742300033569336, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7115144729614258, + "num_tokens": 443972576.0, + "step": 17171 + }, + { + "epoch": 1.8857895892817922, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.251258611679077, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7295697927474976, + "num_tokens": 443997637.0, + "step": 17172 + }, + { + "epoch": 1.885899406984406, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.6336147785186768, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7274017930030823, + "num_tokens": 444016602.0, + "step": 17173 + }, + { + "epoch": 1.8860092246870197, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1809582710266113, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7271609306335449, + "num_tokens": 444042596.0, + "step": 17174 + }, + { + "epoch": 1.8861190423896332, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9640158414840698, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.739310622215271, + "num_tokens": 444071027.0, + "step": 17175 + }, + { + "epoch": 1.8862288600922468, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.220841884613037, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7069994211196899, + "num_tokens": 444097511.0, + "step": 17176 + }, + { + "epoch": 1.8863386777948605, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2302019596099854, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.7098279595375061, + "num_tokens": 444123105.0, + "step": 17177 + }, + { + "epoch": 1.8864484954974743, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0007998943328857, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6972314715385437, + "num_tokens": 444152880.0, + "step": 17178 + }, + { + "epoch": 1.8865583132000878, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2432024478912354, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7119167447090149, + "num_tokens": 444177159.0, + "step": 17179 + }, + { + "epoch": 1.8866681309027014, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0177841186523438, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7299825549125671, + "num_tokens": 444204765.0, + "step": 17180 + }, + { + "epoch": 1.8867779486053151, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.8041929006576538, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7072339057922363, + "num_tokens": 444240019.0, + "step": 17181 + }, + { + "epoch": 1.8868877663079289, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.119647264480591, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7264309525489807, + "num_tokens": 444265462.0, + "step": 17182 + }, + { + "epoch": 1.8869975840105426, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2438619136810303, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7453075647354126, + "num_tokens": 444290496.0, + "step": 17183 + }, + { + "epoch": 1.8871074017131562, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1532416343688965, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7356075644493103, + "num_tokens": 444316974.0, + "step": 17184 + }, + { + "epoch": 1.8872172194157697, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.130441665649414, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7177813053131104, + "num_tokens": 444347182.0, + "step": 17185 + }, + { + "epoch": 1.8873270371183835, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.280144691467285, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7085854411125183, + "num_tokens": 444373363.0, + "step": 17186 + }, + { + "epoch": 1.8874368548209972, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5672643184661865, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7085913419723511, + "num_tokens": 444395418.0, + "step": 17187 + }, + { + "epoch": 1.887546672523611, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0127687454223633, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7447459697723389, + "num_tokens": 444423375.0, + "step": 17188 + }, + { + "epoch": 1.8876564902262245, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4208340644836426, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7290031313896179, + "num_tokens": 444444920.0, + "step": 17189 + }, + { + "epoch": 1.887766307928838, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.165722608566284, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7082458138465881, + "num_tokens": 444474529.0, + "step": 17190 + }, + { + "epoch": 1.8878761256314518, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.261348247528076, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7074439525604248, + "num_tokens": 444500093.0, + "step": 17191 + }, + { + "epoch": 1.8879859433340656, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2314419746398926, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7054303884506226, + "num_tokens": 444527583.0, + "step": 17192 + }, + { + "epoch": 1.888095761036679, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.106001853942871, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7458833456039429, + "num_tokens": 444554468.0, + "step": 17193 + }, + { + "epoch": 1.8882055787392926, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4514877796173096, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7112663388252258, + "num_tokens": 444576911.0, + "step": 17194 + }, + { + "epoch": 1.8883153964419064, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1768064498901367, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7103409767150879, + "num_tokens": 444606474.0, + "step": 17195 + }, + { + "epoch": 1.8884252141445201, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.130425453186035, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.6961476802825928, + "num_tokens": 444634777.0, + "step": 17196 + }, + { + "epoch": 1.888535031847134, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2395431995391846, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7186439037322998, + "num_tokens": 444659687.0, + "step": 17197 + }, + { + "epoch": 1.8886448495497474, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.332386016845703, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.73475182056427, + "num_tokens": 444684014.0, + "step": 17198 + }, + { + "epoch": 1.888754667252361, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.450685739517212, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7051042318344116, + "num_tokens": 444706289.0, + "step": 17199 + }, + { + "epoch": 1.8888644849549747, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.498525381088257, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7191469073295593, + "num_tokens": 444727952.0, + "step": 17200 + }, + { + "epoch": 1.8889743026575885, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0757665634155273, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7221969366073608, + "num_tokens": 444757089.0, + "step": 17201 + }, + { + "epoch": 1.889084120360202, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9977573156356812, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7029541730880737, + "num_tokens": 444787948.0, + "step": 17202 + }, + { + "epoch": 1.8891939380628158, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0349199771881104, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.693088173866272, + "num_tokens": 444821121.0, + "step": 17203 + }, + { + "epoch": 1.8893037557654293, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4492976665496826, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7219914197921753, + "num_tokens": 444843954.0, + "step": 17204 + }, + { + "epoch": 1.889413573468043, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4074833393096924, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.703736424446106, + "num_tokens": 444865200.0, + "step": 17205 + }, + { + "epoch": 1.8895233911706568, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3473904132843018, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7258082628250122, + "num_tokens": 444888083.0, + "step": 17206 + }, + { + "epoch": 1.8896332088732704, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2897043228149414, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7364775538444519, + "num_tokens": 444912017.0, + "step": 17207 + }, + { + "epoch": 1.8897430265758839, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4100499153137207, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7337040901184082, + "num_tokens": 444934303.0, + "step": 17208 + }, + { + "epoch": 1.8898528442784976, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.502777576446533, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7228503823280334, + "num_tokens": 444956792.0, + "step": 17209 + }, + { + "epoch": 1.8899626619811114, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.443796157836914, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7248747944831848, + "num_tokens": 444978512.0, + "step": 17210 + }, + { + "epoch": 1.8900724796837252, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1463029384613037, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7151639461517334, + "num_tokens": 445007847.0, + "step": 17211 + }, + { + "epoch": 1.8901822973863387, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1797635555267334, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7154541611671448, + "num_tokens": 445034565.0, + "step": 17212 + }, + { + "epoch": 1.8902921150889522, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9905635118484497, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7094069719314575, + "num_tokens": 445066671.0, + "step": 17213 + }, + { + "epoch": 1.890401932791566, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9629868268966675, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7049943208694458, + "num_tokens": 445100563.0, + "step": 17214 + }, + { + "epoch": 1.8905117504941797, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.058164596557617, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7302000522613525, + "num_tokens": 445129719.0, + "step": 17215 + }, + { + "epoch": 1.8906215681967933, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.458811044692993, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7128623723983765, + "num_tokens": 445152273.0, + "step": 17216 + }, + { + "epoch": 1.890731385899407, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2114951610565186, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7323155403137207, + "num_tokens": 445176937.0, + "step": 17217 + }, + { + "epoch": 1.8908412036020206, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0242960453033447, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.699705958366394, + "num_tokens": 445207189.0, + "step": 17218 + }, + { + "epoch": 1.8909510213046343, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4068291187286377, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7192293405532837, + "num_tokens": 445229436.0, + "step": 17219 + }, + { + "epoch": 1.891060839007248, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.235922336578369, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7202665209770203, + "num_tokens": 445257816.0, + "step": 17220 + }, + { + "epoch": 1.8911706567098616, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1951396465301514, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7125377655029297, + "num_tokens": 445282658.0, + "step": 17221 + }, + { + "epoch": 1.8912804744124752, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.225423574447632, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7110105156898499, + "num_tokens": 445309648.0, + "step": 17222 + }, + { + "epoch": 1.891390292115089, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.233513355255127, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7215926647186279, + "num_tokens": 445336103.0, + "step": 17223 + }, + { + "epoch": 1.8915001098177027, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9890464544296265, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7188402414321899, + "num_tokens": 445366873.0, + "step": 17224 + }, + { + "epoch": 1.8916099275203164, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.194572687149048, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7029411196708679, + "num_tokens": 445394047.0, + "step": 17225 + }, + { + "epoch": 1.89171974522293, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.381157636642456, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.715674638748169, + "num_tokens": 445418118.0, + "step": 17226 + }, + { + "epoch": 1.8918295629255435, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0900611877441406, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7174279093742371, + "num_tokens": 445446647.0, + "step": 17227 + }, + { + "epoch": 1.8919393806281573, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2790489196777344, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.726852297782898, + "num_tokens": 445470464.0, + "step": 17228 + }, + { + "epoch": 1.892049198330771, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4135589599609375, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7365587949752808, + "num_tokens": 445492287.0, + "step": 17229 + }, + { + "epoch": 1.8921590160333845, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3292598724365234, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7133548259735107, + "num_tokens": 445516898.0, + "step": 17230 + }, + { + "epoch": 1.892268833735998, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.275475263595581, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7103317379951477, + "num_tokens": 445541226.0, + "step": 17231 + }, + { + "epoch": 1.8923786514386118, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.362539529800415, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7303243279457092, + "num_tokens": 445564130.0, + "step": 17232 + }, + { + "epoch": 1.8924884691412256, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1731996536254883, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7275109887123108, + "num_tokens": 445589156.0, + "step": 17233 + }, + { + "epoch": 1.8925982868438394, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.273684024810791, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7184172868728638, + "num_tokens": 445613269.0, + "step": 17234 + }, + { + "epoch": 1.8927081045464529, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1422646045684814, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.706511914730072, + "num_tokens": 445642926.0, + "step": 17235 + }, + { + "epoch": 1.8928179222490664, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4274516105651855, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7485930323600769, + "num_tokens": 445663012.0, + "step": 17236 + }, + { + "epoch": 1.8929277399516802, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.160813093185425, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7253354787826538, + "num_tokens": 445690161.0, + "step": 17237 + }, + { + "epoch": 1.893037557654294, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1984035968780518, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7085498571395874, + "num_tokens": 445716987.0, + "step": 17238 + }, + { + "epoch": 1.8931473753569077, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2336556911468506, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7440038919448853, + "num_tokens": 445741219.0, + "step": 17239 + }, + { + "epoch": 1.8932571930595212, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3589859008789062, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.706762433052063, + "num_tokens": 445766150.0, + "step": 17240 + }, + { + "epoch": 1.8933670107621348, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0762534141540527, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7073748707771301, + "num_tokens": 445793915.0, + "step": 17241 + }, + { + "epoch": 1.8934768284647485, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1853623390197754, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.709938645362854, + "num_tokens": 445819999.0, + "step": 17242 + }, + { + "epoch": 1.8935866461673623, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.125380754470825, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7130231857299805, + "num_tokens": 445849330.0, + "step": 17243 + }, + { + "epoch": 1.8936964638699758, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.185859441757202, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7050014734268188, + "num_tokens": 445876150.0, + "step": 17244 + }, + { + "epoch": 1.8938062815725893, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1077492237091064, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7203835248947144, + "num_tokens": 445906024.0, + "step": 17245 + }, + { + "epoch": 1.893916099275203, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0835530757904053, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.700819730758667, + "num_tokens": 445934276.0, + "step": 17246 + }, + { + "epoch": 1.8940259169778169, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1776912212371826, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7071443200111389, + "num_tokens": 445961624.0, + "step": 17247 + }, + { + "epoch": 1.8941357346804306, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1397833824157715, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7234852313995361, + "num_tokens": 445991363.0, + "step": 17248 + }, + { + "epoch": 1.8942455523830442, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.415935516357422, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.709887683391571, + "num_tokens": 446014534.0, + "step": 17249 + }, + { + "epoch": 1.8943553700856577, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5060529708862305, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7380083203315735, + "num_tokens": 446036606.0, + "step": 17250 + }, + { + "epoch": 1.8944651877882714, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5792465209960938, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7260141968727112, + "num_tokens": 446056241.0, + "step": 17251 + }, + { + "epoch": 1.8945750054908852, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.9196596145629883, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7273921966552734, + "num_tokens": 446073179.0, + "step": 17252 + }, + { + "epoch": 1.894684823193499, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.257232904434204, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.7032839059829712, + "num_tokens": 446102627.0, + "step": 17253 + }, + { + "epoch": 1.8947946408961125, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.204442262649536, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7270938158035278, + "num_tokens": 446129432.0, + "step": 17254 + }, + { + "epoch": 1.894904458598726, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3616693019866943, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7306298017501831, + "num_tokens": 446151081.0, + "step": 17255 + }, + { + "epoch": 1.8950142763013398, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4203765392303467, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7142987251281738, + "num_tokens": 446176354.0, + "step": 17256 + }, + { + "epoch": 1.8951240940039535, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2268283367156982, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7185273170471191, + "num_tokens": 446202852.0, + "step": 17257 + }, + { + "epoch": 1.895233911706567, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.452277660369873, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6921257972717285, + "num_tokens": 446229433.0, + "step": 17258 + }, + { + "epoch": 1.8953437294091806, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4396278858184814, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.733496904373169, + "num_tokens": 446250044.0, + "step": 17259 + }, + { + "epoch": 1.8954535471117944, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.336968183517456, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7126673460006714, + "num_tokens": 446274982.0, + "step": 17260 + }, + { + "epoch": 1.8955633648144081, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.499298334121704, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7259408235549927, + "num_tokens": 446296475.0, + "step": 17261 + }, + { + "epoch": 1.8956731825170219, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9438161849975586, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7113230228424072, + "num_tokens": 446329801.0, + "step": 17262 + }, + { + "epoch": 1.8957830002196354, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2509219646453857, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7043582201004028, + "num_tokens": 446354878.0, + "step": 17263 + }, + { + "epoch": 1.895892817922249, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9814256429672241, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7205191850662231, + "num_tokens": 446388505.0, + "step": 17264 + }, + { + "epoch": 1.8960026356248627, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4892382621765137, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7105316519737244, + "num_tokens": 446409303.0, + "step": 17265 + }, + { + "epoch": 1.8961124533274765, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0863091945648193, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7136516571044922, + "num_tokens": 446437388.0, + "step": 17266 + }, + { + "epoch": 1.89622227103009, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0729758739471436, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7042095065116882, + "num_tokens": 446467633.0, + "step": 17267 + }, + { + "epoch": 1.8963320887327038, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.192854166030884, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7135565876960754, + "num_tokens": 446495235.0, + "step": 17268 + }, + { + "epoch": 1.8964419064353173, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.149675130844116, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.707481324672699, + "num_tokens": 446526089.0, + "step": 17269 + }, + { + "epoch": 1.896551724137931, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.607900381088257, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7236784100532532, + "num_tokens": 446546020.0, + "step": 17270 + }, + { + "epoch": 1.8966615418405448, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.087322950363159, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.723900318145752, + "num_tokens": 446573748.0, + "step": 17271 + }, + { + "epoch": 1.8967713595431583, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2501137256622314, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7284080982208252, + "num_tokens": 446600106.0, + "step": 17272 + }, + { + "epoch": 1.8968811772457719, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.622256278991699, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7332789897918701, + "num_tokens": 446621102.0, + "step": 17273 + }, + { + "epoch": 1.8969909949483856, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1750378608703613, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.691420316696167, + "num_tokens": 446647799.0, + "step": 17274 + }, + { + "epoch": 1.8971008126509994, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1690256595611572, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7169551849365234, + "num_tokens": 446673387.0, + "step": 17275 + }, + { + "epoch": 1.8972106303536131, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2676138877868652, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7190198302268982, + "num_tokens": 446698085.0, + "step": 17276 + }, + { + "epoch": 1.8973204480562267, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.229435443878174, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6898373365402222, + "num_tokens": 446724016.0, + "step": 17277 + }, + { + "epoch": 1.8974302657588402, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.212669610977173, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7295805215835571, + "num_tokens": 446749350.0, + "step": 17278 + }, + { + "epoch": 1.897540083461454, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0335044860839844, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7266044616699219, + "num_tokens": 446779779.0, + "step": 17279 + }, + { + "epoch": 1.8976499011640677, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.188657522201538, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7287641763687134, + "num_tokens": 446805939.0, + "step": 17280 + }, + { + "epoch": 1.8977597188666813, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1156535148620605, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.711860179901123, + "num_tokens": 446837087.0, + "step": 17281 + }, + { + "epoch": 1.897869536569295, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.232689380645752, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7250620126724243, + "num_tokens": 446862282.0, + "step": 17282 + }, + { + "epoch": 1.8979793542719086, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0026986598968506, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7317650318145752, + "num_tokens": 446892472.0, + "step": 17283 + }, + { + "epoch": 1.8980891719745223, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2369396686553955, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7343900203704834, + "num_tokens": 446917216.0, + "step": 17284 + }, + { + "epoch": 1.898198989677136, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.142263174057007, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7235665321350098, + "num_tokens": 446942707.0, + "step": 17285 + }, + { + "epoch": 1.8983088073797496, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1370017528533936, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7171657085418701, + "num_tokens": 446970127.0, + "step": 17286 + }, + { + "epoch": 1.8984186250823631, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.6383697986602783, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7155016660690308, + "num_tokens": 446990487.0, + "step": 17287 + }, + { + "epoch": 1.898528442784977, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2090373039245605, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7310081720352173, + "num_tokens": 447016971.0, + "step": 17288 + }, + { + "epoch": 1.8986382604875907, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.994630217552185, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.687798023223877, + "num_tokens": 447049599.0, + "step": 17289 + }, + { + "epoch": 1.8987480781902044, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1261396408081055, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7195233702659607, + "num_tokens": 447077909.0, + "step": 17290 + }, + { + "epoch": 1.898857895892818, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2462100982666016, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7463232278823853, + "num_tokens": 447102460.0, + "step": 17291 + }, + { + "epoch": 1.8989677135954315, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.260302782058716, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7247171401977539, + "num_tokens": 447127190.0, + "step": 17292 + }, + { + "epoch": 1.8990775312980452, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2464985847473145, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7366583347320557, + "num_tokens": 447152851.0, + "step": 17293 + }, + { + "epoch": 1.899187349000659, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.109663963317871, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7124426364898682, + "num_tokens": 447181968.0, + "step": 17294 + }, + { + "epoch": 1.8992971667032725, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3749139308929443, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7290548086166382, + "num_tokens": 447204277.0, + "step": 17295 + }, + { + "epoch": 1.899406984405886, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3993136882781982, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7334381341934204, + "num_tokens": 447226752.0, + "step": 17296 + }, + { + "epoch": 1.8995168021084998, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3806684017181396, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7485373020172119, + "num_tokens": 447248056.0, + "step": 17297 + }, + { + "epoch": 1.8996266198111136, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.293818950653076, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7114569544792175, + "num_tokens": 447273500.0, + "step": 17298 + }, + { + "epoch": 1.8997364375137273, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.163649320602417, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.717101514339447, + "num_tokens": 447302646.0, + "step": 17299 + }, + { + "epoch": 1.8998462552163409, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2261784076690674, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7205395698547363, + "num_tokens": 447328778.0, + "step": 17300 + }, + { + "epoch": 1.8999560729189544, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.3103950023651123, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7253824472427368, + "num_tokens": 447354432.0, + "step": 17301 + }, + { + "epoch": 1.9000658906215682, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1423072814941406, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7285100221633911, + "num_tokens": 447381128.0, + "step": 17302 + }, + { + "epoch": 1.900175708324182, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4039599895477295, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7178386449813843, + "num_tokens": 447406027.0, + "step": 17303 + }, + { + "epoch": 1.9002855260267957, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2117691040039062, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7504891157150269, + "num_tokens": 447430713.0, + "step": 17304 + }, + { + "epoch": 1.9003953437294092, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.355343818664551, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.728881299495697, + "num_tokens": 447455069.0, + "step": 17305 + }, + { + "epoch": 1.9005051614320227, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1198794841766357, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7123883366584778, + "num_tokens": 447483421.0, + "step": 17306 + }, + { + "epoch": 1.9006149791346365, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0355396270751953, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7093645334243774, + "num_tokens": 447511348.0, + "step": 17307 + }, + { + "epoch": 1.9007247968372503, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4386510848999023, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7338515520095825, + "num_tokens": 447533441.0, + "step": 17308 + }, + { + "epoch": 1.9008346145398638, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1790544986724854, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.722586989402771, + "num_tokens": 447558387.0, + "step": 17309 + }, + { + "epoch": 1.9009444322424773, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9314228296279907, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7040976881980896, + "num_tokens": 447592242.0, + "step": 17310 + }, + { + "epoch": 1.901054249945091, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.375089645385742, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7057231068611145, + "num_tokens": 447615537.0, + "step": 17311 + }, + { + "epoch": 1.9011640676477048, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.061462879180908, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7259977459907532, + "num_tokens": 447645715.0, + "step": 17312 + }, + { + "epoch": 1.9012738853503186, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.105400562286377, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7306624054908752, + "num_tokens": 447674361.0, + "step": 17313 + }, + { + "epoch": 1.9013837030529321, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.153881072998047, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7115121483802795, + "num_tokens": 447700851.0, + "step": 17314 + }, + { + "epoch": 1.9014935207555457, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0692341327667236, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7186635732650757, + "num_tokens": 447729638.0, + "step": 17315 + }, + { + "epoch": 1.9016033384581594, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.9942094087600708, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7042539715766907, + "num_tokens": 447760890.0, + "step": 17316 + }, + { + "epoch": 1.9017131561607732, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.069641590118408, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7451856136322021, + "num_tokens": 447785658.0, + "step": 17317 + }, + { + "epoch": 1.9018229738633867, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.276055335998535, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7092705965042114, + "num_tokens": 447811159.0, + "step": 17318 + }, + { + "epoch": 1.9019327915660005, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.12062931060791, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7305129170417786, + "num_tokens": 447838044.0, + "step": 17319 + }, + { + "epoch": 1.902042609268614, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.281691074371338, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.731757402420044, + "num_tokens": 447865087.0, + "step": 17320 + }, + { + "epoch": 1.9021524269712278, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.215900421142578, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7121626138687134, + "num_tokens": 447890684.0, + "step": 17321 + }, + { + "epoch": 1.9022622446738415, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.5312376022338867, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7245477437973022, + "num_tokens": 447915223.0, + "step": 17322 + }, + { + "epoch": 1.902372062376455, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1393637657165527, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7072998881340027, + "num_tokens": 447943562.0, + "step": 17323 + }, + { + "epoch": 1.9024818800790686, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.4111483097076416, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7274554967880249, + "num_tokens": 447967291.0, + "step": 17324 + }, + { + "epoch": 1.9025916977816824, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.2028889656066895, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7095804214477539, + "num_tokens": 447993169.0, + "step": 17325 + }, + { + "epoch": 1.9027015154842961, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0882270336151123, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7081620693206787, + "num_tokens": 448021760.0, + "step": 17326 + }, + { + "epoch": 1.9028113331869099, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 1.999646544456482, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6978039741516113, + "num_tokens": 448054045.0, + "step": 17327 + }, + { + "epoch": 1.9029211508895234, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.022449016571045, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7205913662910461, + "num_tokens": 448083280.0, + "step": 17328 + }, + { + "epoch": 1.903030968592137, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1951231956481934, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7197222709655762, + "num_tokens": 448110152.0, + "step": 17329 + }, + { + "epoch": 1.9031407862947507, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.0344367027282715, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.6984829306602478, + "num_tokens": 448141465.0, + "step": 17330 + }, + { + "epoch": 1.9032506039973645, + "ewc_loss": 0.00014781951904296875, + "grad_norm": 2.1945571899414062, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.737318754196167, + "num_tokens": 448165574.0, + "step": 17331 + }, + { + "epoch": 1.903360421699978, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0601816177368164, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7013466954231262, + "num_tokens": 448193919.0, + "step": 17332 + }, + { + "epoch": 1.9034702394025917, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1449389457702637, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.72868812084198, + "num_tokens": 448219763.0, + "step": 17333 + }, + { + "epoch": 1.9035800571052053, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0367448329925537, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6939963698387146, + "num_tokens": 448251785.0, + "step": 17334 + }, + { + "epoch": 1.903689874807819, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2056243419647217, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7217360734939575, + "num_tokens": 448278044.0, + "step": 17335 + }, + { + "epoch": 1.9037996925104328, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2121050357818604, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7298908829689026, + "num_tokens": 448303866.0, + "step": 17336 + }, + { + "epoch": 1.9039095102130463, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.006760835647583, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7088067531585693, + "num_tokens": 448334191.0, + "step": 17337 + }, + { + "epoch": 1.9040193279156599, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.270720958709717, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.713592529296875, + "num_tokens": 448361400.0, + "step": 17338 + }, + { + "epoch": 1.9041291456182736, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.626420021057129, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7130740284919739, + "num_tokens": 448381714.0, + "step": 17339 + }, + { + "epoch": 1.9042389633208874, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.230290651321411, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7147360444068909, + "num_tokens": 448407440.0, + "step": 17340 + }, + { + "epoch": 1.9043487810235011, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4927542209625244, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7239407300949097, + "num_tokens": 448429954.0, + "step": 17341 + }, + { + "epoch": 1.9044585987261147, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.261298894882202, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7371087074279785, + "num_tokens": 448455485.0, + "step": 17342 + }, + { + "epoch": 1.9045684164287282, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3658382892608643, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7431432008743286, + "num_tokens": 448480074.0, + "step": 17343 + }, + { + "epoch": 1.904678234131342, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1194889545440674, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7372305393218994, + "num_tokens": 448507473.0, + "step": 17344 + }, + { + "epoch": 1.9047880518339557, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4592208862304688, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7190878391265869, + "num_tokens": 448529085.0, + "step": 17345 + }, + { + "epoch": 1.9048978695365693, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3961198329925537, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7188199758529663, + "num_tokens": 448553224.0, + "step": 17346 + }, + { + "epoch": 1.9050076872391828, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.366590976715088, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7038203477859497, + "num_tokens": 448577920.0, + "step": 17347 + }, + { + "epoch": 1.9051175049417965, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1584980487823486, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7269595861434937, + "num_tokens": 448604254.0, + "step": 17348 + }, + { + "epoch": 1.9052273226444103, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4074912071228027, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7214154005050659, + "num_tokens": 448626666.0, + "step": 17349 + }, + { + "epoch": 1.905337140347024, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.8843492269515991, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.709673285484314, + "num_tokens": 448661643.0, + "step": 17350 + }, + { + "epoch": 1.9054469580496376, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2302942276000977, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7145971059799194, + "num_tokens": 448688672.0, + "step": 17351 + }, + { + "epoch": 1.9055567757522511, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.8564350605010986, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7247766852378845, + "num_tokens": 448707192.0, + "step": 17352 + }, + { + "epoch": 1.9056665934548649, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.322110176086426, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7027184963226318, + "num_tokens": 448733459.0, + "step": 17353 + }, + { + "epoch": 1.9057764111574786, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1562602519989014, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7323458194732666, + "num_tokens": 448758398.0, + "step": 17354 + }, + { + "epoch": 1.9058862288600924, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4660871028900146, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7193267345428467, + "num_tokens": 448779754.0, + "step": 17355 + }, + { + "epoch": 1.905996046562706, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3998560905456543, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7100006341934204, + "num_tokens": 448804077.0, + "step": 17356 + }, + { + "epoch": 1.9061058642653195, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.6008684635162354, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7265727519989014, + "num_tokens": 448825247.0, + "step": 17357 + }, + { + "epoch": 1.9062156819679332, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1753885746002197, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7258797287940979, + "num_tokens": 448850877.0, + "step": 17358 + }, + { + "epoch": 1.906325499670547, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.36044979095459, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7093870043754578, + "num_tokens": 448873759.0, + "step": 17359 + }, + { + "epoch": 1.9064353173731605, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.103459596633911, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7021011710166931, + "num_tokens": 448902244.0, + "step": 17360 + }, + { + "epoch": 1.906545135075774, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9009286165237427, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7209334373474121, + "num_tokens": 448937166.0, + "step": 17361 + }, + { + "epoch": 1.9066549527783878, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.318419933319092, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7202273607254028, + "num_tokens": 448961982.0, + "step": 17362 + }, + { + "epoch": 1.9067647704810016, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1141703128814697, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7084190845489502, + "num_tokens": 448993537.0, + "step": 17363 + }, + { + "epoch": 1.9068745881836153, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.214151382446289, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7236239314079285, + "num_tokens": 449020416.0, + "step": 17364 + }, + { + "epoch": 1.9069844058862289, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1989588737487793, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7377758622169495, + "num_tokens": 449046315.0, + "step": 17365 + }, + { + "epoch": 1.9070942235888424, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3014352321624756, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7058621644973755, + "num_tokens": 449072172.0, + "step": 17366 + }, + { + "epoch": 1.9072040412914562, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4463143348693848, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7283191084861755, + "num_tokens": 449094238.0, + "step": 17367 + }, + { + "epoch": 1.90731385899407, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3128743171691895, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7193322777748108, + "num_tokens": 449118465.0, + "step": 17368 + }, + { + "epoch": 1.9074236766966837, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.426797389984131, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7078655362129211, + "num_tokens": 449140523.0, + "step": 17369 + }, + { + "epoch": 1.9075334943992972, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.332543134689331, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.709757924079895, + "num_tokens": 449167458.0, + "step": 17370 + }, + { + "epoch": 1.9076433121019107, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.550943613052368, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7309776544570923, + "num_tokens": 449187811.0, + "step": 17371 + }, + { + "epoch": 1.9077531298045245, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.214256525039673, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.738696813583374, + "num_tokens": 449212999.0, + "step": 17372 + }, + { + "epoch": 1.9078629475071383, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.261852741241455, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7003004550933838, + "num_tokens": 449238024.0, + "step": 17373 + }, + { + "epoch": 1.9079727652097518, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0628252029418945, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7183787226676941, + "num_tokens": 449266862.0, + "step": 17374 + }, + { + "epoch": 1.9080825829123653, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0601675510406494, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7336355447769165, + "num_tokens": 449296136.0, + "step": 17375 + }, + { + "epoch": 1.908192400614979, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.226267099380493, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7057554721832275, + "num_tokens": 449322698.0, + "step": 17376 + }, + { + "epoch": 1.9083022183175928, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1504275798797607, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.6978635787963867, + "num_tokens": 449351229.0, + "step": 17377 + }, + { + "epoch": 1.9084120360202066, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3426015377044678, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7322795391082764, + "num_tokens": 449374457.0, + "step": 17378 + }, + { + "epoch": 1.9085218537228201, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1634442806243896, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7166991829872131, + "num_tokens": 449401738.0, + "step": 17379 + }, + { + "epoch": 1.9086316714254337, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.143526077270508, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7144932746887207, + "num_tokens": 449428375.0, + "step": 17380 + }, + { + "epoch": 1.9087414891280474, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0642378330230713, + "learning_rate": 1e-06, + "loss": 1.0679, + "mean_token_accuracy": 0.6843315362930298, + "num_tokens": 449460367.0, + "step": 17381 + }, + { + "epoch": 1.9088513068306612, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.418436288833618, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7356396913528442, + "num_tokens": 449481939.0, + "step": 17382 + }, + { + "epoch": 1.9089611245332747, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2105820178985596, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7148463726043701, + "num_tokens": 449507779.0, + "step": 17383 + }, + { + "epoch": 1.9090709422358885, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.24161434173584, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7253704071044922, + "num_tokens": 449536049.0, + "step": 17384 + }, + { + "epoch": 1.909180759938502, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.581360340118408, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7362061738967896, + "num_tokens": 449556604.0, + "step": 17385 + }, + { + "epoch": 1.9092905776411158, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0826871395111084, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7232133746147156, + "num_tokens": 449583690.0, + "step": 17386 + }, + { + "epoch": 1.9094003953437295, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1924901008605957, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7337794899940491, + "num_tokens": 449609306.0, + "step": 17387 + }, + { + "epoch": 1.909510213046343, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1125810146331787, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7241774797439575, + "num_tokens": 449637102.0, + "step": 17388 + }, + { + "epoch": 1.9096200307489566, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2869935035705566, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7347614765167236, + "num_tokens": 449659902.0, + "step": 17389 + }, + { + "epoch": 1.9097298484515703, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2305006980895996, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7248766422271729, + "num_tokens": 449685570.0, + "step": 17390 + }, + { + "epoch": 1.909839666154184, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2226836681365967, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7288159728050232, + "num_tokens": 449709286.0, + "step": 17391 + }, + { + "epoch": 1.9099494838567979, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.194038152694702, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7146682143211365, + "num_tokens": 449735615.0, + "step": 17392 + }, + { + "epoch": 1.9100593015594114, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.109938383102417, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7121521234512329, + "num_tokens": 449762331.0, + "step": 17393 + }, + { + "epoch": 1.910169119262025, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9831336736679077, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7117432355880737, + "num_tokens": 449795113.0, + "step": 17394 + }, + { + "epoch": 1.9102789369646387, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2696890830993652, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7294260263442993, + "num_tokens": 449821302.0, + "step": 17395 + }, + { + "epoch": 1.9103887546672524, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2094430923461914, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7128512859344482, + "num_tokens": 449847719.0, + "step": 17396 + }, + { + "epoch": 1.910498572369866, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.19209885597229, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.700665295124054, + "num_tokens": 449874132.0, + "step": 17397 + }, + { + "epoch": 1.9106083900724797, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.658090353012085, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7310483455657959, + "num_tokens": 449892049.0, + "step": 17398 + }, + { + "epoch": 1.9107182077750933, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1296653747558594, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7166998386383057, + "num_tokens": 449918827.0, + "step": 17399 + }, + { + "epoch": 1.910828025477707, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.542158365249634, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.712860107421875, + "num_tokens": 449940306.0, + "step": 17400 + }, + { + "epoch": 1.9109378431803208, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0735487937927246, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7049373388290405, + "num_tokens": 449968701.0, + "step": 17401 + }, + { + "epoch": 1.9110476608829343, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.330709457397461, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.6989766955375671, + "num_tokens": 449994837.0, + "step": 17402 + }, + { + "epoch": 1.9111574785855479, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.230116128921509, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7060153484344482, + "num_tokens": 450021966.0, + "step": 17403 + }, + { + "epoch": 1.9112672962881616, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1226038932800293, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7052935361862183, + "num_tokens": 450050687.0, + "step": 17404 + }, + { + "epoch": 1.9113771139907754, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2580976486206055, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.721117377281189, + "num_tokens": 450076238.0, + "step": 17405 + }, + { + "epoch": 1.9114869316933891, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.501293420791626, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.721362829208374, + "num_tokens": 450096935.0, + "step": 17406 + }, + { + "epoch": 1.9115967493960027, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3646135330200195, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.719365119934082, + "num_tokens": 450120989.0, + "step": 17407 + }, + { + "epoch": 1.9117065670986162, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2559218406677246, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7251842021942139, + "num_tokens": 450148570.0, + "step": 17408 + }, + { + "epoch": 1.91181638480123, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.8451112508773804, + "learning_rate": 1e-06, + "loss": 0.7634, + "mean_token_accuracy": 0.7585256099700928, + "num_tokens": 450180517.0, + "step": 17409 + }, + { + "epoch": 1.9119262025038437, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3999404907226562, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7341801524162292, + "num_tokens": 450204352.0, + "step": 17410 + }, + { + "epoch": 1.9120360202064572, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5361523628234863, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7198046445846558, + "num_tokens": 450223829.0, + "step": 17411 + }, + { + "epoch": 1.9121458379090708, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3197598457336426, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7135305404663086, + "num_tokens": 450249659.0, + "step": 17412 + }, + { + "epoch": 1.9122556556116845, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3848345279693604, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7406346797943115, + "num_tokens": 450271616.0, + "step": 17413 + }, + { + "epoch": 1.9123654733142983, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3489081859588623, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7401632070541382, + "num_tokens": 450293969.0, + "step": 17414 + }, + { + "epoch": 1.912475291016912, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4245615005493164, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7289460301399231, + "num_tokens": 450317278.0, + "step": 17415 + }, + { + "epoch": 1.9125851087195256, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3638033866882324, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.709801435470581, + "num_tokens": 450343184.0, + "step": 17416 + }, + { + "epoch": 1.9126949264221391, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.6856637001037598, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7489378452301025, + "num_tokens": 450360937.0, + "step": 17417 + }, + { + "epoch": 1.9128047441247529, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2385292053222656, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.719355583190918, + "num_tokens": 450384911.0, + "step": 17418 + }, + { + "epoch": 1.9129145618273666, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.823423385620117, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7480767369270325, + "num_tokens": 450403271.0, + "step": 17419 + }, + { + "epoch": 1.9130243795299804, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4311864376068115, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7419049143791199, + "num_tokens": 450425048.0, + "step": 17420 + }, + { + "epoch": 1.913134197232594, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1848723888397217, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7572032809257507, + "num_tokens": 450449798.0, + "step": 17421 + }, + { + "epoch": 1.9132440149352075, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.507155179977417, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7257033586502075, + "num_tokens": 450471716.0, + "step": 17422 + }, + { + "epoch": 1.9133538326378212, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.352693557739258, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.6925909519195557, + "num_tokens": 450499535.0, + "step": 17423 + }, + { + "epoch": 1.913463650340435, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2768442630767822, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7219787240028381, + "num_tokens": 450523739.0, + "step": 17424 + }, + { + "epoch": 1.9135734680430485, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.138583183288574, + "learning_rate": 1e-06, + "loss": 1.0612, + "mean_token_accuracy": 0.6792571544647217, + "num_tokens": 450556393.0, + "step": 17425 + }, + { + "epoch": 1.913683285745662, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.429250955581665, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.728870689868927, + "num_tokens": 450577971.0, + "step": 17426 + }, + { + "epoch": 1.9137931034482758, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0171027183532715, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7094625234603882, + "num_tokens": 450609518.0, + "step": 17427 + }, + { + "epoch": 1.9139029211508896, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5820391178131104, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7379465699195862, + "num_tokens": 450629031.0, + "step": 17428 + }, + { + "epoch": 1.9140127388535033, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1179120540618896, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7246242165565491, + "num_tokens": 450656033.0, + "step": 17429 + }, + { + "epoch": 1.9141225565561168, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2549421787261963, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7124804854393005, + "num_tokens": 450681630.0, + "step": 17430 + }, + { + "epoch": 1.9142323742587304, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5044209957122803, + "learning_rate": 1e-06, + "loss": 0.7905, + "mean_token_accuracy": 0.749582827091217, + "num_tokens": 450701588.0, + "step": 17431 + }, + { + "epoch": 1.9143421919613441, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.369668960571289, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7302019596099854, + "num_tokens": 450724736.0, + "step": 17432 + }, + { + "epoch": 1.914452009663958, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.340731620788574, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7253084778785706, + "num_tokens": 450748266.0, + "step": 17433 + }, + { + "epoch": 1.9145618273665717, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.084643840789795, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.723934531211853, + "num_tokens": 450777544.0, + "step": 17434 + }, + { + "epoch": 1.9146716450691852, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.281618118286133, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.724913477897644, + "num_tokens": 450802881.0, + "step": 17435 + }, + { + "epoch": 1.9147814627717987, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.158864974975586, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7070600986480713, + "num_tokens": 450832711.0, + "step": 17436 + }, + { + "epoch": 1.9148912804744125, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0728847980499268, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7102845907211304, + "num_tokens": 450862156.0, + "step": 17437 + }, + { + "epoch": 1.9150010981770262, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.177553415298462, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7014411091804504, + "num_tokens": 450887857.0, + "step": 17438 + }, + { + "epoch": 1.9151109158796398, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.11185359954834, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7169652581214905, + "num_tokens": 450916891.0, + "step": 17439 + }, + { + "epoch": 1.9152207335822533, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1970348358154297, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7018598914146423, + "num_tokens": 450943925.0, + "step": 17440 + }, + { + "epoch": 1.915330551284867, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4383037090301514, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7481086254119873, + "num_tokens": 450964413.0, + "step": 17441 + }, + { + "epoch": 1.9154403689874808, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.273084878921509, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7207770347595215, + "num_tokens": 450989912.0, + "step": 17442 + }, + { + "epoch": 1.9155501866900946, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3629913330078125, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7146362662315369, + "num_tokens": 451014695.0, + "step": 17443 + }, + { + "epoch": 1.9156600043927081, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.497605323791504, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.707415759563446, + "num_tokens": 451035290.0, + "step": 17444 + }, + { + "epoch": 1.9157698220953217, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5441231727600098, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7283493876457214, + "num_tokens": 451057017.0, + "step": 17445 + }, + { + "epoch": 1.9158796397979354, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2943274974823, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7295438051223755, + "num_tokens": 451081191.0, + "step": 17446 + }, + { + "epoch": 1.9159894575005492, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.174466133117676, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7276479005813599, + "num_tokens": 451107356.0, + "step": 17447 + }, + { + "epoch": 1.9160992752031627, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9342021942138672, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7070070505142212, + "num_tokens": 451142762.0, + "step": 17448 + }, + { + "epoch": 1.9162090929057765, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.168405532836914, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7253127098083496, + "num_tokens": 451170603.0, + "step": 17449 + }, + { + "epoch": 1.91631891060839, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.139563798904419, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7057245969772339, + "num_tokens": 451200087.0, + "step": 17450 + }, + { + "epoch": 1.9164287283110037, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3459596633911133, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7085138559341431, + "num_tokens": 451223252.0, + "step": 17451 + }, + { + "epoch": 1.9165385460136175, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3960626125335693, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7127920389175415, + "num_tokens": 451245659.0, + "step": 17452 + }, + { + "epoch": 1.916648363716231, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.370500326156616, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7360073924064636, + "num_tokens": 451267037.0, + "step": 17453 + }, + { + "epoch": 1.9167581814188446, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.572309732437134, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7257649302482605, + "num_tokens": 451288055.0, + "step": 17454 + }, + { + "epoch": 1.9168679991214583, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5876412391662598, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7071249485015869, + "num_tokens": 451309635.0, + "step": 17455 + }, + { + "epoch": 1.916977816824072, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0355653762817383, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7201809883117676, + "num_tokens": 451341564.0, + "step": 17456 + }, + { + "epoch": 1.9170876345266858, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1053123474121094, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6944444179534912, + "num_tokens": 451369371.0, + "step": 17457 + }, + { + "epoch": 1.9171974522292994, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2564609050750732, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7279609441757202, + "num_tokens": 451393862.0, + "step": 17458 + }, + { + "epoch": 1.917307269931913, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2145941257476807, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7165528535842896, + "num_tokens": 451421284.0, + "step": 17459 + }, + { + "epoch": 1.9174170876345267, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4094226360321045, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7097669243812561, + "num_tokens": 451445744.0, + "step": 17460 + }, + { + "epoch": 1.9175269053371404, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.26249623298645, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7258959412574768, + "num_tokens": 451469195.0, + "step": 17461 + }, + { + "epoch": 1.917636723039754, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3745388984680176, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7201604843139648, + "num_tokens": 451492045.0, + "step": 17462 + }, + { + "epoch": 1.9177465407423677, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4825704097747803, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.719180166721344, + "num_tokens": 451514767.0, + "step": 17463 + }, + { + "epoch": 1.9178563584449813, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.43090558052063, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.723845899105072, + "num_tokens": 451541390.0, + "step": 17464 + }, + { + "epoch": 1.917966176147595, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2691617012023926, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7250745296478271, + "num_tokens": 451565160.0, + "step": 17465 + }, + { + "epoch": 1.9180759938502088, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4545063972473145, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7077151536941528, + "num_tokens": 451587544.0, + "step": 17466 + }, + { + "epoch": 1.9181858115528223, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3252155780792236, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7456310987472534, + "num_tokens": 451612067.0, + "step": 17467 + }, + { + "epoch": 1.9182956292554358, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1555309295654297, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7354086637496948, + "num_tokens": 451638093.0, + "step": 17468 + }, + { + "epoch": 1.9184054469580496, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.144704818725586, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7223719954490662, + "num_tokens": 451665022.0, + "step": 17469 + }, + { + "epoch": 1.9185152646606634, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2493488788604736, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7080468535423279, + "num_tokens": 451692246.0, + "step": 17470 + }, + { + "epoch": 1.9186250823632771, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.337775707244873, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7460517883300781, + "num_tokens": 451715934.0, + "step": 17471 + }, + { + "epoch": 1.9187349000658906, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9555411338806152, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6930785775184631, + "num_tokens": 451751060.0, + "step": 17472 + }, + { + "epoch": 1.9188447177685042, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9772121906280518, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7105790376663208, + "num_tokens": 451782865.0, + "step": 17473 + }, + { + "epoch": 1.918954535471118, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0781610012054443, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.726154088973999, + "num_tokens": 451812176.0, + "step": 17474 + }, + { + "epoch": 1.9190643531737317, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.243635892868042, + "learning_rate": 1e-06, + "loss": 0.8133, + "mean_token_accuracy": 0.7408850193023682, + "num_tokens": 451835676.0, + "step": 17475 + }, + { + "epoch": 1.9191741708763452, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0647270679473877, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7151692509651184, + "num_tokens": 451864951.0, + "step": 17476 + }, + { + "epoch": 1.9192839885789588, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2848000526428223, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7182720899581909, + "num_tokens": 451888418.0, + "step": 17477 + }, + { + "epoch": 1.9193938062815725, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0861687660217285, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7026963829994202, + "num_tokens": 451917245.0, + "step": 17478 + }, + { + "epoch": 1.9195036239841863, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2967488765716553, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7497739791870117, + "num_tokens": 451940714.0, + "step": 17479 + }, + { + "epoch": 1.9196134416868, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.114201307296753, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7104216814041138, + "num_tokens": 451971952.0, + "step": 17480 + }, + { + "epoch": 1.9197232593894136, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.136108160018921, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7028138637542725, + "num_tokens": 452002024.0, + "step": 17481 + }, + { + "epoch": 1.919833077092027, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.201110601425171, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.720740556716919, + "num_tokens": 452029407.0, + "step": 17482 + }, + { + "epoch": 1.9199428947946409, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0873236656188965, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7138735055923462, + "num_tokens": 452056850.0, + "step": 17483 + }, + { + "epoch": 1.9200527124972546, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1321160793304443, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7155374884605408, + "num_tokens": 452084598.0, + "step": 17484 + }, + { + "epoch": 1.9201625301998684, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1207282543182373, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7292249202728271, + "num_tokens": 452112922.0, + "step": 17485 + }, + { + "epoch": 1.920272347902482, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.028531312942505, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7084421515464783, + "num_tokens": 452142665.0, + "step": 17486 + }, + { + "epoch": 1.9203821656050954, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.326716899871826, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7220010757446289, + "num_tokens": 452165606.0, + "step": 17487 + }, + { + "epoch": 1.9204919833077092, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.548007011413574, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7473312616348267, + "num_tokens": 452184198.0, + "step": 17488 + }, + { + "epoch": 1.920601801010323, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4720351696014404, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7199323177337646, + "num_tokens": 452206770.0, + "step": 17489 + }, + { + "epoch": 1.9207116187129365, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2350916862487793, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7151809334754944, + "num_tokens": 452233802.0, + "step": 17490 + }, + { + "epoch": 1.92082143641555, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4041240215301514, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7241488099098206, + "num_tokens": 452257788.0, + "step": 17491 + }, + { + "epoch": 1.9209312541181638, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0623395442962646, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7139718532562256, + "num_tokens": 452285869.0, + "step": 17492 + }, + { + "epoch": 1.9210410718207775, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9477367401123047, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.726252555847168, + "num_tokens": 452317624.0, + "step": 17493 + }, + { + "epoch": 1.9211508895233913, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.301558256149292, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7245701551437378, + "num_tokens": 452341393.0, + "step": 17494 + }, + { + "epoch": 1.9212607072260048, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.360227584838867, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7202632427215576, + "num_tokens": 452366505.0, + "step": 17495 + }, + { + "epoch": 1.9213705249286184, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.494739055633545, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7435152530670166, + "num_tokens": 452387653.0, + "step": 17496 + }, + { + "epoch": 1.9214803426312321, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3586716651916504, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.738442063331604, + "num_tokens": 452410836.0, + "step": 17497 + }, + { + "epoch": 1.9215901603338459, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0048983097076416, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7127212882041931, + "num_tokens": 452441639.0, + "step": 17498 + }, + { + "epoch": 1.9216999780364594, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.50555157661438, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7173875570297241, + "num_tokens": 452462803.0, + "step": 17499 + }, + { + "epoch": 1.9218097957390732, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3167006969451904, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7358393669128418, + "num_tokens": 452485567.0, + "step": 17500 + }, + { + "epoch": 1.9219196134416867, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2466671466827393, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7049810886383057, + "num_tokens": 452512177.0, + "step": 17501 + }, + { + "epoch": 1.9220294311443005, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.065331220626831, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.725875198841095, + "num_tokens": 452540671.0, + "step": 17502 + }, + { + "epoch": 1.9221392488469142, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0000112056732178, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7196888327598572, + "num_tokens": 452572954.0, + "step": 17503 + }, + { + "epoch": 1.9222490665495278, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.465056896209717, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7165272235870361, + "num_tokens": 452595664.0, + "step": 17504 + }, + { + "epoch": 1.9223588842521413, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.507099151611328, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.729012668132782, + "num_tokens": 452616085.0, + "step": 17505 + }, + { + "epoch": 1.922468701954755, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4580538272857666, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7029380798339844, + "num_tokens": 452641245.0, + "step": 17506 + }, + { + "epoch": 1.9225785196573688, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.278136968612671, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7116146087646484, + "num_tokens": 452666302.0, + "step": 17507 + }, + { + "epoch": 1.9226883373599826, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3160839080810547, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7270526885986328, + "num_tokens": 452689351.0, + "step": 17508 + }, + { + "epoch": 1.922798155062596, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.598050355911255, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7291587591171265, + "num_tokens": 452709981.0, + "step": 17509 + }, + { + "epoch": 1.9229079727652096, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0273725986480713, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7166920304298401, + "num_tokens": 452739967.0, + "step": 17510 + }, + { + "epoch": 1.9230177904678234, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.25469708442688, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7239814400672913, + "num_tokens": 452766530.0, + "step": 17511 + }, + { + "epoch": 1.9231276081704372, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.141157388687134, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.697167158126831, + "num_tokens": 452795231.0, + "step": 17512 + }, + { + "epoch": 1.9232374258730507, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1253297328948975, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7161028385162354, + "num_tokens": 452823578.0, + "step": 17513 + }, + { + "epoch": 1.9233472435756644, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0240509510040283, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7094724178314209, + "num_tokens": 452855213.0, + "step": 17514 + }, + { + "epoch": 1.923457061278278, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.226823329925537, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6912131309509277, + "num_tokens": 452885761.0, + "step": 17515 + }, + { + "epoch": 1.9235668789808917, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2968289852142334, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7344244718551636, + "num_tokens": 452909656.0, + "step": 17516 + }, + { + "epoch": 1.9236766966835055, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9478486776351929, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7217479944229126, + "num_tokens": 452945783.0, + "step": 17517 + }, + { + "epoch": 1.923786514386119, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3705008029937744, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7259337306022644, + "num_tokens": 452967176.0, + "step": 17518 + }, + { + "epoch": 1.9238963320887326, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3795173168182373, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7016459703445435, + "num_tokens": 452991130.0, + "step": 17519 + }, + { + "epoch": 1.9240061497913463, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.147502899169922, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7308539748191833, + "num_tokens": 453017318.0, + "step": 17520 + }, + { + "epoch": 1.92411596749396, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3548083305358887, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7051389813423157, + "num_tokens": 453042097.0, + "step": 17521 + }, + { + "epoch": 1.9242257851965738, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.163564920425415, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.710608720779419, + "num_tokens": 453070639.0, + "step": 17522 + }, + { + "epoch": 1.9243356028991874, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3571763038635254, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7432805299758911, + "num_tokens": 453092593.0, + "step": 17523 + }, + { + "epoch": 1.924445420601801, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9867976903915405, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7433832883834839, + "num_tokens": 453121695.0, + "step": 17524 + }, + { + "epoch": 1.9245552383044147, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0857813358306885, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7064571380615234, + "num_tokens": 453150979.0, + "step": 17525 + }, + { + "epoch": 1.9246650560070284, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.903804063796997, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7047773003578186, + "num_tokens": 453184250.0, + "step": 17526 + }, + { + "epoch": 1.924774873709642, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0246975421905518, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.6889151334762573, + "num_tokens": 453215976.0, + "step": 17527 + }, + { + "epoch": 1.9248846914122557, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3243801593780518, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7488977909088135, + "num_tokens": 453239722.0, + "step": 17528 + }, + { + "epoch": 1.9249945091148692, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.139951705932617, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7035434246063232, + "num_tokens": 453267887.0, + "step": 17529 + }, + { + "epoch": 1.925104326817483, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.076659917831421, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7084430456161499, + "num_tokens": 453297401.0, + "step": 17530 + }, + { + "epoch": 1.9252141445200968, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.947097897529602, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7215303182601929, + "num_tokens": 453328276.0, + "step": 17531 + }, + { + "epoch": 1.9253239622227103, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0221354961395264, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.7032521963119507, + "num_tokens": 453361354.0, + "step": 17532 + }, + { + "epoch": 1.9254337799253238, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.364332914352417, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7232148051261902, + "num_tokens": 453385604.0, + "step": 17533 + }, + { + "epoch": 1.9255435976279376, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1832146644592285, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7287991642951965, + "num_tokens": 453411241.0, + "step": 17534 + }, + { + "epoch": 1.9256534153305513, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3725459575653076, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7511346340179443, + "num_tokens": 453432239.0, + "step": 17535 + }, + { + "epoch": 1.925763233033165, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.063015937805176, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7179723978042603, + "num_tokens": 453461818.0, + "step": 17536 + }, + { + "epoch": 1.9258730507357786, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.6309351921081543, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7298663854598999, + "num_tokens": 453482038.0, + "step": 17537 + }, + { + "epoch": 1.9259828684383922, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.159245491027832, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.728686511516571, + "num_tokens": 453508784.0, + "step": 17538 + }, + { + "epoch": 1.926092686141006, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0430400371551514, + "learning_rate": 1e-06, + "loss": 1.0286, + "mean_token_accuracy": 0.6956212520599365, + "num_tokens": 453538687.0, + "step": 17539 + }, + { + "epoch": 1.9262025038436197, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2260243892669678, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.6990039944648743, + "num_tokens": 453564911.0, + "step": 17540 + }, + { + "epoch": 1.9263123215462332, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9555485248565674, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6949383020401001, + "num_tokens": 453599779.0, + "step": 17541 + }, + { + "epoch": 1.9264221392488468, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2430758476257324, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7105966806411743, + "num_tokens": 453624549.0, + "step": 17542 + }, + { + "epoch": 1.9265319569514605, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3271484375, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7206366062164307, + "num_tokens": 453648337.0, + "step": 17543 + }, + { + "epoch": 1.9266417746540743, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.177157402038574, + "learning_rate": 1e-06, + "loss": 0.7704, + "mean_token_accuracy": 0.752378523349762, + "num_tokens": 453675954.0, + "step": 17544 + }, + { + "epoch": 1.926751592356688, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.37062406539917, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.6980314254760742, + "num_tokens": 453702509.0, + "step": 17545 + }, + { + "epoch": 1.9268614100593016, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0838170051574707, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7370460033416748, + "num_tokens": 453730629.0, + "step": 17546 + }, + { + "epoch": 1.926971227761915, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4561102390289307, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7295929193496704, + "num_tokens": 453752699.0, + "step": 17547 + }, + { + "epoch": 1.9270810454645289, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.309436798095703, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.694466233253479, + "num_tokens": 453779843.0, + "step": 17548 + }, + { + "epoch": 1.9271908631671426, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1702966690063477, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7251087427139282, + "num_tokens": 453805125.0, + "step": 17549 + }, + { + "epoch": 1.9273006808697564, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0852508544921875, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7062459588050842, + "num_tokens": 453834983.0, + "step": 17550 + }, + { + "epoch": 1.92741049857237, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.201258420944214, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.732038140296936, + "num_tokens": 453861241.0, + "step": 17551 + }, + { + "epoch": 1.9275203162749834, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.296171188354492, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7146017551422119, + "num_tokens": 453887612.0, + "step": 17552 + }, + { + "epoch": 1.9276301339775972, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4897396564483643, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7448984384536743, + "num_tokens": 453909501.0, + "step": 17553 + }, + { + "epoch": 1.927739951680211, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.6008150577545166, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7336764335632324, + "num_tokens": 453930561.0, + "step": 17554 + }, + { + "epoch": 1.9278497693828245, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.085935592651367, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6923688054084778, + "num_tokens": 453960082.0, + "step": 17555 + }, + { + "epoch": 1.927959587085438, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.8374892473220825, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7165951132774353, + "num_tokens": 453994622.0, + "step": 17556 + }, + { + "epoch": 1.9280694047880518, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2104763984680176, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7373056411743164, + "num_tokens": 454019594.0, + "step": 17557 + }, + { + "epoch": 1.9281792224906655, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3928375244140625, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7499937415122986, + "num_tokens": 454042410.0, + "step": 17558 + }, + { + "epoch": 1.9282890401932793, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9392892122268677, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7288752794265747, + "num_tokens": 454074885.0, + "step": 17559 + }, + { + "epoch": 1.9283988578958928, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3202438354492188, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7039899826049805, + "num_tokens": 454099407.0, + "step": 17560 + }, + { + "epoch": 1.9285086755985064, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.192631959915161, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7211271524429321, + "num_tokens": 454125813.0, + "step": 17561 + }, + { + "epoch": 1.9286184933011201, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.104872703552246, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.746532678604126, + "num_tokens": 454152954.0, + "step": 17562 + }, + { + "epoch": 1.9287283110037339, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.22501277923584, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7276272177696228, + "num_tokens": 454179351.0, + "step": 17563 + }, + { + "epoch": 1.9288381287063474, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.169435977935791, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7146531939506531, + "num_tokens": 454205510.0, + "step": 17564 + }, + { + "epoch": 1.9289479464089612, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3797924518585205, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7337088584899902, + "num_tokens": 454228764.0, + "step": 17565 + }, + { + "epoch": 1.9290577641115747, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.551011323928833, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.735115647315979, + "num_tokens": 454248196.0, + "step": 17566 + }, + { + "epoch": 1.9291675818141885, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3713223934173584, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7345196604728699, + "num_tokens": 454269712.0, + "step": 17567 + }, + { + "epoch": 1.9292773995168022, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.159780263900757, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7003779411315918, + "num_tokens": 454297607.0, + "step": 17568 + }, + { + "epoch": 1.9293872172194158, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4225540161132812, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7219763994216919, + "num_tokens": 454321168.0, + "step": 17569 + }, + { + "epoch": 1.9294970349220293, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.989184021949768, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7187202572822571, + "num_tokens": 454352864.0, + "step": 17570 + }, + { + "epoch": 1.929606852624643, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2387216091156006, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7332193851470947, + "num_tokens": 454377003.0, + "step": 17571 + }, + { + "epoch": 1.9297166703272568, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.313734292984009, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7092485427856445, + "num_tokens": 454403443.0, + "step": 17572 + }, + { + "epoch": 1.9298264880298706, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1749861240386963, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7248722910881042, + "num_tokens": 454428688.0, + "step": 17573 + }, + { + "epoch": 1.929936305732484, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.435420274734497, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7485909461975098, + "num_tokens": 454448632.0, + "step": 17574 + }, + { + "epoch": 1.9300461234350976, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0590484142303467, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7283252477645874, + "num_tokens": 454478079.0, + "step": 17575 + }, + { + "epoch": 1.9301559411377114, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.250577688217163, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7291941046714783, + "num_tokens": 454503560.0, + "step": 17576 + }, + { + "epoch": 1.9302657588403251, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1784603595733643, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7061092853546143, + "num_tokens": 454532357.0, + "step": 17577 + }, + { + "epoch": 1.9303755765429387, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0706393718719482, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7119321823120117, + "num_tokens": 454561608.0, + "step": 17578 + }, + { + "epoch": 1.9304853942455524, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.172722816467285, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7135382294654846, + "num_tokens": 454590535.0, + "step": 17579 + }, + { + "epoch": 1.930595211948166, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3591346740722656, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7390779852867126, + "num_tokens": 454613021.0, + "step": 17580 + }, + { + "epoch": 1.9307050296507797, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3002848625183105, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7024713158607483, + "num_tokens": 454638183.0, + "step": 17581 + }, + { + "epoch": 1.9308148473533935, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1968963146209717, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7246736288070679, + "num_tokens": 454664254.0, + "step": 17582 + }, + { + "epoch": 1.930924665056007, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.876814365386963, + "learning_rate": 1e-06, + "loss": 1.0121, + "mean_token_accuracy": 0.6978042125701904, + "num_tokens": 454702713.0, + "step": 17583 + }, + { + "epoch": 1.9310344827586206, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1332573890686035, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7146216630935669, + "num_tokens": 454729770.0, + "step": 17584 + }, + { + "epoch": 1.9311443004612343, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1580073833465576, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7226521968841553, + "num_tokens": 454754170.0, + "step": 17585 + }, + { + "epoch": 1.931254118163848, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.291048526763916, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7127573490142822, + "num_tokens": 454779132.0, + "step": 17586 + }, + { + "epoch": 1.9313639358664618, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2784390449523926, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7079707384109497, + "num_tokens": 454803807.0, + "step": 17587 + }, + { + "epoch": 1.9314737535690754, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.238758087158203, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7030053734779358, + "num_tokens": 454831878.0, + "step": 17588 + }, + { + "epoch": 1.931583571271689, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1927883625030518, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7126017808914185, + "num_tokens": 454859770.0, + "step": 17589 + }, + { + "epoch": 1.9316933889743026, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.6510099172592163, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7068969011306763, + "num_tokens": 454905089.0, + "step": 17590 + }, + { + "epoch": 1.9318032066769164, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0934786796569824, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7178082466125488, + "num_tokens": 454934153.0, + "step": 17591 + }, + { + "epoch": 1.93191302437953, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.026376247406006, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6880865097045898, + "num_tokens": 454965708.0, + "step": 17592 + }, + { + "epoch": 1.9320228420821435, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1309194564819336, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7014061212539673, + "num_tokens": 454994316.0, + "step": 17593 + }, + { + "epoch": 1.9321326597847572, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.41101336479187, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7338274717330933, + "num_tokens": 455016511.0, + "step": 17594 + }, + { + "epoch": 1.932242477487371, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.451885223388672, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7516178488731384, + "num_tokens": 455036822.0, + "step": 17595 + }, + { + "epoch": 1.9323522951899847, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1391499042510986, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7216953039169312, + "num_tokens": 455065624.0, + "step": 17596 + }, + { + "epoch": 1.9324621128925983, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.271480083465576, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7174105048179626, + "num_tokens": 455092259.0, + "step": 17597 + }, + { + "epoch": 1.9325719305952118, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.5216243267059326, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7199302911758423, + "num_tokens": 455114492.0, + "step": 17598 + }, + { + "epoch": 1.9326817482978256, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4192399978637695, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7376214265823364, + "num_tokens": 455135634.0, + "step": 17599 + }, + { + "epoch": 1.9327915660004393, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0748989582061768, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7181440591812134, + "num_tokens": 455167182.0, + "step": 17600 + }, + { + "epoch": 1.932901383703053, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3053231239318848, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7183266282081604, + "num_tokens": 455191465.0, + "step": 17601 + }, + { + "epoch": 1.9330112014056666, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9514176845550537, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7215768098831177, + "num_tokens": 455223966.0, + "step": 17602 + }, + { + "epoch": 1.9331210191082802, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1083123683929443, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7185823917388916, + "num_tokens": 455250006.0, + "step": 17603 + }, + { + "epoch": 1.933230836810894, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.354452610015869, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7130347490310669, + "num_tokens": 455273183.0, + "step": 17604 + }, + { + "epoch": 1.9333406545135077, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.220019578933716, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7046710848808289, + "num_tokens": 455302408.0, + "step": 17605 + }, + { + "epoch": 1.9334504722161212, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2039623260498047, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7151672840118408, + "num_tokens": 455330213.0, + "step": 17606 + }, + { + "epoch": 1.9335602899187347, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4397573471069336, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7002657651901245, + "num_tokens": 455354541.0, + "step": 17607 + }, + { + "epoch": 1.9336701076213485, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1616032123565674, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7083953619003296, + "num_tokens": 455382122.0, + "step": 17608 + }, + { + "epoch": 1.9337799253239623, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.211794137954712, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7120369672775269, + "num_tokens": 455408109.0, + "step": 17609 + }, + { + "epoch": 1.933889743026576, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4265472888946533, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7316314578056335, + "num_tokens": 455429011.0, + "step": 17610 + }, + { + "epoch": 1.9339995607291895, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.3130054473876953, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7232246994972229, + "num_tokens": 455453465.0, + "step": 17611 + }, + { + "epoch": 1.934109378431803, + "ewc_loss": 0.000148773193359375, + "grad_norm": 1.9885977506637573, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7136601209640503, + "num_tokens": 455484298.0, + "step": 17612 + }, + { + "epoch": 1.9342191961344168, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.215684413909912, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7067672610282898, + "num_tokens": 455510085.0, + "step": 17613 + }, + { + "epoch": 1.9343290138370306, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2883481979370117, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7039128541946411, + "num_tokens": 455534673.0, + "step": 17614 + }, + { + "epoch": 1.9344388315396444, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2759902477264404, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7156382203102112, + "num_tokens": 455559873.0, + "step": 17615 + }, + { + "epoch": 1.934548649242258, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0806307792663574, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7150859236717224, + "num_tokens": 455588007.0, + "step": 17616 + }, + { + "epoch": 1.9346584669448714, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2417781352996826, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7398054599761963, + "num_tokens": 455612967.0, + "step": 17617 + }, + { + "epoch": 1.9347682846474852, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.184098720550537, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7139598727226257, + "num_tokens": 455641419.0, + "step": 17618 + }, + { + "epoch": 1.934878102350099, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.4240314960479736, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7257227897644043, + "num_tokens": 455663530.0, + "step": 17619 + }, + { + "epoch": 1.9349879200527125, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1274099349975586, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7208966612815857, + "num_tokens": 455690929.0, + "step": 17620 + }, + { + "epoch": 1.935097737755326, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.240041494369507, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7217435836791992, + "num_tokens": 455716597.0, + "step": 17621 + }, + { + "epoch": 1.9352075554579398, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1497151851654053, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7065342664718628, + "num_tokens": 455745358.0, + "step": 17622 + }, + { + "epoch": 1.9353173731605535, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0536892414093018, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7029106020927429, + "num_tokens": 455773826.0, + "step": 17623 + }, + { + "epoch": 1.9354271908631673, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0732545852661133, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7015502452850342, + "num_tokens": 455802065.0, + "step": 17624 + }, + { + "epoch": 1.9355370085657808, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0364339351654053, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7170908451080322, + "num_tokens": 455831596.0, + "step": 17625 + }, + { + "epoch": 1.9356468262683943, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.055638074874878, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.6979315876960754, + "num_tokens": 455860969.0, + "step": 17626 + }, + { + "epoch": 1.935756643971008, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.388457775115967, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7202098965644836, + "num_tokens": 455884957.0, + "step": 17627 + }, + { + "epoch": 1.9358664616736219, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1526613235473633, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.6968833208084106, + "num_tokens": 455912229.0, + "step": 17628 + }, + { + "epoch": 1.9359762793762354, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1858747005462646, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6998743414878845, + "num_tokens": 455939141.0, + "step": 17629 + }, + { + "epoch": 1.9360860970788492, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.013180732727051, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7055066227912903, + "num_tokens": 455970042.0, + "step": 17630 + }, + { + "epoch": 1.9361959147814627, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.429149627685547, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7376713752746582, + "num_tokens": 455991574.0, + "step": 17631 + }, + { + "epoch": 1.9363057324840764, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.33297061920166, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7266645431518555, + "num_tokens": 456014313.0, + "step": 17632 + }, + { + "epoch": 1.9364155501866902, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.2642552852630615, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7351824045181274, + "num_tokens": 456038844.0, + "step": 17633 + }, + { + "epoch": 1.9365253678893037, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.1169545650482178, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7091479301452637, + "num_tokens": 456065470.0, + "step": 17634 + }, + { + "epoch": 1.9366351855919173, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.0869314670562744, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7254565358161926, + "num_tokens": 456094501.0, + "step": 17635 + }, + { + "epoch": 1.936745003294531, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.214043617248535, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7182590961456299, + "num_tokens": 456121203.0, + "step": 17636 + }, + { + "epoch": 1.9368548209971448, + "ewc_loss": 0.000148773193359375, + "grad_norm": 2.182464838027954, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6995754241943359, + "num_tokens": 456147979.0, + "step": 17637 + }, + { + "epoch": 1.9369646386997585, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3276119232177734, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7473648190498352, + "num_tokens": 456169219.0, + "step": 17638 + }, + { + "epoch": 1.937074456402372, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.109954357147217, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7087429761886597, + "num_tokens": 456197959.0, + "step": 17639 + }, + { + "epoch": 1.9371842741049856, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1691198348999023, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7171816229820251, + "num_tokens": 456225101.0, + "step": 17640 + }, + { + "epoch": 1.9372940918075994, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2799630165100098, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.7032735347747803, + "num_tokens": 456251271.0, + "step": 17641 + }, + { + "epoch": 1.9374039095102131, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2051565647125244, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7020907402038574, + "num_tokens": 456279573.0, + "step": 17642 + }, + { + "epoch": 1.9375137272128267, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.29038405418396, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7200836539268494, + "num_tokens": 456304106.0, + "step": 17643 + }, + { + "epoch": 1.9376235449154404, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.479767322540283, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.71382737159729, + "num_tokens": 456326000.0, + "step": 17644 + }, + { + "epoch": 1.937733362618054, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.07834529876709, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7028206586837769, + "num_tokens": 456357308.0, + "step": 17645 + }, + { + "epoch": 1.9378431803206677, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3367364406585693, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7151143550872803, + "num_tokens": 456379764.0, + "step": 17646 + }, + { + "epoch": 1.9379529980232815, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.222080945968628, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7315397262573242, + "num_tokens": 456404692.0, + "step": 17647 + }, + { + "epoch": 1.938062815725895, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2683072090148926, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7043880224227905, + "num_tokens": 456429190.0, + "step": 17648 + }, + { + "epoch": 1.9381726334285085, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.017529010772705, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7185844779014587, + "num_tokens": 456459855.0, + "step": 17649 + }, + { + "epoch": 1.9382824511311223, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9864779710769653, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7156225442886353, + "num_tokens": 456489508.0, + "step": 17650 + }, + { + "epoch": 1.938392268833736, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2520058155059814, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7049679756164551, + "num_tokens": 456514346.0, + "step": 17651 + }, + { + "epoch": 1.9385020865363498, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9384278059005737, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.693945050239563, + "num_tokens": 456547450.0, + "step": 17652 + }, + { + "epoch": 1.9386119042389633, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.6672239303588867, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7457585334777832, + "num_tokens": 456565621.0, + "step": 17653 + }, + { + "epoch": 1.9387217219415769, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.537052869796753, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7100883722305298, + "num_tokens": 456585911.0, + "step": 17654 + }, + { + "epoch": 1.9388315396441906, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.143629550933838, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.7018319368362427, + "num_tokens": 456614068.0, + "step": 17655 + }, + { + "epoch": 1.9389413573468044, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9112141132354736, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7136290669441223, + "num_tokens": 456646528.0, + "step": 17656 + }, + { + "epoch": 1.939051175049418, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.247407913208008, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.7025972604751587, + "num_tokens": 456672194.0, + "step": 17657 + }, + { + "epoch": 1.9391609927520315, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4870123863220215, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7185162901878357, + "num_tokens": 456692698.0, + "step": 17658 + }, + { + "epoch": 1.9392708104546452, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.325261354446411, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7155724763870239, + "num_tokens": 456715664.0, + "step": 17659 + }, + { + "epoch": 1.939380628157259, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1282997131347656, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7225245237350464, + "num_tokens": 456741854.0, + "step": 17660 + }, + { + "epoch": 1.9394904458598727, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1018130779266357, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7097649574279785, + "num_tokens": 456768364.0, + "step": 17661 + }, + { + "epoch": 1.9396002635624863, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0683040618896484, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6945596933364868, + "num_tokens": 456799532.0, + "step": 17662 + }, + { + "epoch": 1.9397100812650998, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.303834915161133, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7409441471099854, + "num_tokens": 456823868.0, + "step": 17663 + }, + { + "epoch": 1.9398198989677136, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.717158794403076, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7502163052558899, + "num_tokens": 456843968.0, + "step": 17664 + }, + { + "epoch": 1.9399297166703273, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.701514720916748, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7332810759544373, + "num_tokens": 456862883.0, + "step": 17665 + }, + { + "epoch": 1.940039534372941, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.334409475326538, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7167089581489563, + "num_tokens": 456886445.0, + "step": 17666 + }, + { + "epoch": 1.9401493520755546, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4673168659210205, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7214020490646362, + "num_tokens": 456908085.0, + "step": 17667 + }, + { + "epoch": 1.9402591697781681, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1424078941345215, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7225489616394043, + "num_tokens": 456937381.0, + "step": 17668 + }, + { + "epoch": 1.940368987480782, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.279393434524536, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.73075270652771, + "num_tokens": 456961903.0, + "step": 17669 + }, + { + "epoch": 1.9404788051833957, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.316321849822998, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6958153247833252, + "num_tokens": 456986959.0, + "step": 17670 + }, + { + "epoch": 1.9405886228860092, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.524747133255005, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7136808633804321, + "num_tokens": 457006516.0, + "step": 17671 + }, + { + "epoch": 1.9406984405886227, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.89816415309906, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.731229841709137, + "num_tokens": 457038982.0, + "step": 17672 + }, + { + "epoch": 1.9408082582912365, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.946903944015503, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7287147045135498, + "num_tokens": 457070470.0, + "step": 17673 + }, + { + "epoch": 1.9409180759938502, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.590841293334961, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7396734952926636, + "num_tokens": 457090603.0, + "step": 17674 + }, + { + "epoch": 1.941027893696464, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.316833257675171, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.720219612121582, + "num_tokens": 457116720.0, + "step": 17675 + }, + { + "epoch": 1.9411377113990775, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4509947299957275, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.738606870174408, + "num_tokens": 457138482.0, + "step": 17676 + }, + { + "epoch": 1.941247529101691, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2329154014587402, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7299029231071472, + "num_tokens": 457162724.0, + "step": 17677 + }, + { + "epoch": 1.9413573468043048, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.256645679473877, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7312548160552979, + "num_tokens": 457187631.0, + "step": 17678 + }, + { + "epoch": 1.9414671645069186, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0995936393737793, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.716649055480957, + "num_tokens": 457217336.0, + "step": 17679 + }, + { + "epoch": 1.9415769822095323, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.342984437942505, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7380086183547974, + "num_tokens": 457239708.0, + "step": 17680 + }, + { + "epoch": 1.9416867999121459, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1322453022003174, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.694077730178833, + "num_tokens": 457273370.0, + "step": 17681 + }, + { + "epoch": 1.9417966176147594, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.022322654724121, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7201372385025024, + "num_tokens": 457303652.0, + "step": 17682 + }, + { + "epoch": 1.9419064353173732, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.297492742538452, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7181994915008545, + "num_tokens": 457329030.0, + "step": 17683 + }, + { + "epoch": 1.942016253019987, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.795348644256592, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7355016469955444, + "num_tokens": 457346283.0, + "step": 17684 + }, + { + "epoch": 1.9421260707226005, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.7459945678710938, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7298798561096191, + "num_tokens": 457366083.0, + "step": 17685 + }, + { + "epoch": 1.942235888425214, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0965824127197266, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.701460599899292, + "num_tokens": 457397897.0, + "step": 17686 + }, + { + "epoch": 1.9423457061278278, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2066187858581543, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.71678227186203, + "num_tokens": 457422345.0, + "step": 17687 + }, + { + "epoch": 1.9424555238304415, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.137878656387329, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7267447113990784, + "num_tokens": 457448229.0, + "step": 17688 + }, + { + "epoch": 1.9425653415330553, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0268261432647705, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7127315998077393, + "num_tokens": 457477440.0, + "step": 17689 + }, + { + "epoch": 1.9426751592356688, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0960569381713867, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7058184742927551, + "num_tokens": 457505166.0, + "step": 17690 + }, + { + "epoch": 1.9427849769382823, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0542218685150146, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.7002027034759521, + "num_tokens": 457536226.0, + "step": 17691 + }, + { + "epoch": 1.942894794640896, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0397536754608154, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7189990282058716, + "num_tokens": 457565023.0, + "step": 17692 + }, + { + "epoch": 1.9430046123435099, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1512703895568848, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7250526547431946, + "num_tokens": 457592425.0, + "step": 17693 + }, + { + "epoch": 1.9431144300461234, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2384488582611084, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7082652449607849, + "num_tokens": 457621475.0, + "step": 17694 + }, + { + "epoch": 1.9432242477487371, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.079010248184204, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7059208154678345, + "num_tokens": 457652399.0, + "step": 17695 + }, + { + "epoch": 1.9433340654513507, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.237508535385132, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6829211711883545, + "num_tokens": 457680403.0, + "step": 17696 + }, + { + "epoch": 1.9434438831539644, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2381982803344727, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7134893536567688, + "num_tokens": 457706330.0, + "step": 17697 + }, + { + "epoch": 1.9435537008565782, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.364966630935669, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.6960093975067139, + "num_tokens": 457728851.0, + "step": 17698 + }, + { + "epoch": 1.9436635185591917, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.573740243911743, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7258919477462769, + "num_tokens": 457749633.0, + "step": 17699 + }, + { + "epoch": 1.9437733362618053, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.279353380203247, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.715807318687439, + "num_tokens": 457774304.0, + "step": 17700 + }, + { + "epoch": 1.943883153964419, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.123847007751465, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7422856092453003, + "num_tokens": 457801445.0, + "step": 17701 + }, + { + "epoch": 1.9439929716670328, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2488672733306885, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.729364275932312, + "num_tokens": 457825155.0, + "step": 17702 + }, + { + "epoch": 1.9441027893696465, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1790499687194824, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7178499698638916, + "num_tokens": 457852221.0, + "step": 17703 + }, + { + "epoch": 1.94421260707226, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.175082206726074, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7126941084861755, + "num_tokens": 457879946.0, + "step": 17704 + }, + { + "epoch": 1.9443224247748736, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.084272861480713, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7171920537948608, + "num_tokens": 457910546.0, + "step": 17705 + }, + { + "epoch": 1.9444322424774874, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.403259515762329, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7316970825195312, + "num_tokens": 457933478.0, + "step": 17706 + }, + { + "epoch": 1.9445420601801011, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1195881366729736, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7146180868148804, + "num_tokens": 457964229.0, + "step": 17707 + }, + { + "epoch": 1.9446518778827147, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.304368019104004, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7073575258255005, + "num_tokens": 457988781.0, + "step": 17708 + }, + { + "epoch": 1.9447616955853284, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.165736198425293, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7055963277816772, + "num_tokens": 458016238.0, + "step": 17709 + }, + { + "epoch": 1.944871513287942, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.077756881713867, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6950052976608276, + "num_tokens": 458047596.0, + "step": 17710 + }, + { + "epoch": 1.9449813309905557, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0624663829803467, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6969804763793945, + "num_tokens": 458077736.0, + "step": 17711 + }, + { + "epoch": 1.9450911486931695, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.430040121078491, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7150555849075317, + "num_tokens": 458102156.0, + "step": 17712 + }, + { + "epoch": 1.945200966395783, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.187897205352783, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7239254713058472, + "num_tokens": 458132768.0, + "step": 17713 + }, + { + "epoch": 1.9453107840983965, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1875240802764893, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7069385051727295, + "num_tokens": 458159677.0, + "step": 17714 + }, + { + "epoch": 1.9454206018010103, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.284041404724121, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7171723246574402, + "num_tokens": 458184722.0, + "step": 17715 + }, + { + "epoch": 1.945530419503624, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9137382507324219, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7189840078353882, + "num_tokens": 458217156.0, + "step": 17716 + }, + { + "epoch": 1.9456402372062378, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4528775215148926, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7194757461547852, + "num_tokens": 458238832.0, + "step": 17717 + }, + { + "epoch": 1.9457500549088513, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3535358905792236, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7088615894317627, + "num_tokens": 458262370.0, + "step": 17718 + }, + { + "epoch": 1.9458598726114649, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.643397331237793, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.7404468059539795, + "num_tokens": 458281145.0, + "step": 17719 + }, + { + "epoch": 1.9459696903140786, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1734209060668945, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7209641933441162, + "num_tokens": 458308876.0, + "step": 17720 + }, + { + "epoch": 1.9460795080166924, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0126118659973145, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7014113068580627, + "num_tokens": 458338453.0, + "step": 17721 + }, + { + "epoch": 1.946189325719306, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.128690242767334, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7069207429885864, + "num_tokens": 458364645.0, + "step": 17722 + }, + { + "epoch": 1.9462991434219195, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3412632942199707, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.707362174987793, + "num_tokens": 458389238.0, + "step": 17723 + }, + { + "epoch": 1.9464089611245332, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2887089252471924, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7175079584121704, + "num_tokens": 458412718.0, + "step": 17724 + }, + { + "epoch": 1.946518778827147, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1790928840637207, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7360432147979736, + "num_tokens": 458440883.0, + "step": 17725 + }, + { + "epoch": 1.9466285965297607, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9998910427093506, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7106776237487793, + "num_tokens": 458471234.0, + "step": 17726 + }, + { + "epoch": 1.9467384142323743, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.58793044090271, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7115738391876221, + "num_tokens": 458491313.0, + "step": 17727 + }, + { + "epoch": 1.9468482319349878, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.984907627105713, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.728357195854187, + "num_tokens": 458521269.0, + "step": 17728 + }, + { + "epoch": 1.9469580496376016, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.249633550643921, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7060453295707703, + "num_tokens": 458549404.0, + "step": 17729 + }, + { + "epoch": 1.9470678673402153, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2732369899749756, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7039810419082642, + "num_tokens": 458576406.0, + "step": 17730 + }, + { + "epoch": 1.947177685042829, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.337155342102051, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7226738929748535, + "num_tokens": 458601780.0, + "step": 17731 + }, + { + "epoch": 1.9472875027454426, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2029871940612793, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7365863919258118, + "num_tokens": 458626965.0, + "step": 17732 + }, + { + "epoch": 1.9473973204480561, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2614643573760986, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7341965436935425, + "num_tokens": 458652567.0, + "step": 17733 + }, + { + "epoch": 1.94750713815067, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.38877272605896, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7099158763885498, + "num_tokens": 458678063.0, + "step": 17734 + }, + { + "epoch": 1.9476169558532836, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2157773971557617, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7286522388458252, + "num_tokens": 458703002.0, + "step": 17735 + }, + { + "epoch": 1.9477267735558972, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3049495220184326, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7273552417755127, + "num_tokens": 458726273.0, + "step": 17736 + }, + { + "epoch": 1.9478365912585107, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3558096885681152, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7204363346099854, + "num_tokens": 458748947.0, + "step": 17737 + }, + { + "epoch": 1.9479464089611245, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3427376747131348, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7188405990600586, + "num_tokens": 458773159.0, + "step": 17738 + }, + { + "epoch": 1.9480562266637382, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.093071699142456, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7106726169586182, + "num_tokens": 458802385.0, + "step": 17739 + }, + { + "epoch": 1.948166044366352, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2881200313568115, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7207578420639038, + "num_tokens": 458828374.0, + "step": 17740 + }, + { + "epoch": 1.9482758620689655, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2926647663116455, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7058596611022949, + "num_tokens": 458852194.0, + "step": 17741 + }, + { + "epoch": 1.948385679771579, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.136920213699341, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7364943623542786, + "num_tokens": 458879585.0, + "step": 17742 + }, + { + "epoch": 1.9484954974741928, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0213823318481445, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7208608984947205, + "num_tokens": 458910244.0, + "step": 17743 + }, + { + "epoch": 1.9486053151768066, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1223785877227783, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7176525592803955, + "num_tokens": 458937185.0, + "step": 17744 + }, + { + "epoch": 1.94871513287942, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.101372241973877, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7373161911964417, + "num_tokens": 458963727.0, + "step": 17745 + }, + { + "epoch": 1.9488249505820339, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9579602479934692, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.703748345375061, + "num_tokens": 458996251.0, + "step": 17746 + }, + { + "epoch": 1.9489347682846474, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0069761276245117, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.714523196220398, + "num_tokens": 459027137.0, + "step": 17747 + }, + { + "epoch": 1.9490445859872612, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.308211088180542, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7339819669723511, + "num_tokens": 459050086.0, + "step": 17748 + }, + { + "epoch": 1.949154403689875, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2880561351776123, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.734542191028595, + "num_tokens": 459073889.0, + "step": 17749 + }, + { + "epoch": 1.9492642213924884, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.204744577407837, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7121304869651794, + "num_tokens": 459100458.0, + "step": 17750 + }, + { + "epoch": 1.949374039095102, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.556089401245117, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7349422574043274, + "num_tokens": 459119890.0, + "step": 17751 + }, + { + "epoch": 1.9494838567977157, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2254652976989746, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7067040801048279, + "num_tokens": 459147450.0, + "step": 17752 + }, + { + "epoch": 1.9495936745003295, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1662285327911377, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7503025531768799, + "num_tokens": 459172636.0, + "step": 17753 + }, + { + "epoch": 1.9497034922029433, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.225865364074707, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7071624398231506, + "num_tokens": 459200885.0, + "step": 17754 + }, + { + "epoch": 1.9498133099055568, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0478687286376953, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.728460431098938, + "num_tokens": 459229408.0, + "step": 17755 + }, + { + "epoch": 1.9499231276081703, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.418926477432251, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7393133640289307, + "num_tokens": 459254265.0, + "step": 17756 + }, + { + "epoch": 1.950032945310784, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2142183780670166, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7096635103225708, + "num_tokens": 459280332.0, + "step": 17757 + }, + { + "epoch": 1.9501427630133978, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.220562696456909, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7255450487136841, + "num_tokens": 459305595.0, + "step": 17758 + }, + { + "epoch": 1.9502525807160114, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.5905473232269287, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7444339394569397, + "num_tokens": 459325299.0, + "step": 17759 + }, + { + "epoch": 1.9503623984186251, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2629356384277344, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7170921564102173, + "num_tokens": 459352836.0, + "step": 17760 + }, + { + "epoch": 1.9504722161212387, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.26867413520813, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7050729393959045, + "num_tokens": 459379227.0, + "step": 17761 + }, + { + "epoch": 1.9505820338238524, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0537898540496826, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.6967493891716003, + "num_tokens": 459411692.0, + "step": 17762 + }, + { + "epoch": 1.9506918515264662, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.8891972303390503, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.725275993347168, + "num_tokens": 459445915.0, + "step": 17763 + }, + { + "epoch": 1.9508016692290797, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.38372802734375, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7035608291625977, + "num_tokens": 459471396.0, + "step": 17764 + }, + { + "epoch": 1.9509114869316933, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.5046539306640625, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.721081018447876, + "num_tokens": 459493047.0, + "step": 17765 + }, + { + "epoch": 1.951021304634307, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.592578649520874, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7289418578147888, + "num_tokens": 459515684.0, + "step": 17766 + }, + { + "epoch": 1.9511311223369208, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3057663440704346, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7264386415481567, + "num_tokens": 459540503.0, + "step": 17767 + }, + { + "epoch": 1.9512409400395345, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.090811014175415, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7141115069389343, + "num_tokens": 459569352.0, + "step": 17768 + }, + { + "epoch": 1.951350757742148, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.503006935119629, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7462226748466492, + "num_tokens": 459591727.0, + "step": 17769 + }, + { + "epoch": 1.9514605754447616, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.507866859436035, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7339133620262146, + "num_tokens": 459613008.0, + "step": 17770 + }, + { + "epoch": 1.9515703931473753, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.113858222961426, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7221064567565918, + "num_tokens": 459640519.0, + "step": 17771 + }, + { + "epoch": 1.951680210849989, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2103710174560547, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7156171798706055, + "num_tokens": 459668946.0, + "step": 17772 + }, + { + "epoch": 1.9517900285526026, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.8931328058242798, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7230148315429688, + "num_tokens": 459701833.0, + "step": 17773 + }, + { + "epoch": 1.9518998462552162, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1444449424743652, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7292807102203369, + "num_tokens": 459729092.0, + "step": 17774 + }, + { + "epoch": 1.95200966395783, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3038246631622314, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7065306901931763, + "num_tokens": 459755210.0, + "step": 17775 + }, + { + "epoch": 1.9521194816604437, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.048759698867798, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6965909600257874, + "num_tokens": 459787978.0, + "step": 17776 + }, + { + "epoch": 1.9522292993630574, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1564977169036865, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7314039468765259, + "num_tokens": 459814198.0, + "step": 17777 + }, + { + "epoch": 1.952339117065671, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0162839889526367, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7481158971786499, + "num_tokens": 459843631.0, + "step": 17778 + }, + { + "epoch": 1.9524489347682845, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3980917930603027, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7305930852890015, + "num_tokens": 459865903.0, + "step": 17779 + }, + { + "epoch": 1.9525587524708983, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.439352512359619, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7019262313842773, + "num_tokens": 459890911.0, + "step": 17780 + }, + { + "epoch": 1.952668570173512, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2071540355682373, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7316724061965942, + "num_tokens": 459916634.0, + "step": 17781 + }, + { + "epoch": 1.9527783878761258, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.073826789855957, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7061598300933838, + "num_tokens": 459945596.0, + "step": 17782 + }, + { + "epoch": 1.9528882055787393, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.201004981994629, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6870112419128418, + "num_tokens": 459971509.0, + "step": 17783 + }, + { + "epoch": 1.9529980232813529, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2668333053588867, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7137976288795471, + "num_tokens": 459995866.0, + "step": 17784 + }, + { + "epoch": 1.9531078409839666, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4319710731506348, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7186816930770874, + "num_tokens": 460019422.0, + "step": 17785 + }, + { + "epoch": 1.9532176586865804, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.091524839401245, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7108268141746521, + "num_tokens": 460051120.0, + "step": 17786 + }, + { + "epoch": 1.953327476389194, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.236015796661377, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.728717029094696, + "num_tokens": 460075998.0, + "step": 17787 + }, + { + "epoch": 1.9534372940918074, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.318145751953125, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7400690317153931, + "num_tokens": 460102115.0, + "step": 17788 + }, + { + "epoch": 1.9535471117944212, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3909549713134766, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.718824565410614, + "num_tokens": 460124918.0, + "step": 17789 + }, + { + "epoch": 1.953656929497035, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1537084579467773, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7137146592140198, + "num_tokens": 460151773.0, + "step": 17790 + }, + { + "epoch": 1.9537667471996487, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0789530277252197, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7130541205406189, + "num_tokens": 460180962.0, + "step": 17791 + }, + { + "epoch": 1.9538765649022622, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.361171245574951, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6946194171905518, + "num_tokens": 460204462.0, + "step": 17792 + }, + { + "epoch": 1.9539863826048758, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.195655107498169, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7050666809082031, + "num_tokens": 460231154.0, + "step": 17793 + }, + { + "epoch": 1.9540962003074895, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.548074960708618, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7359710931777954, + "num_tokens": 460251325.0, + "step": 17794 + }, + { + "epoch": 1.9542060180101033, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.370450496673584, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7124873399734497, + "num_tokens": 460275225.0, + "step": 17795 + }, + { + "epoch": 1.954315835712717, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2389676570892334, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7329145073890686, + "num_tokens": 460300141.0, + "step": 17796 + }, + { + "epoch": 1.9544256534153306, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4546003341674805, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.730820894241333, + "num_tokens": 460321850.0, + "step": 17797 + }, + { + "epoch": 1.9545354711179441, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.9082047939300537, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7375478744506836, + "num_tokens": 460353093.0, + "step": 17798 + }, + { + "epoch": 1.9546452888205579, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.231119155883789, + "learning_rate": 1e-06, + "loss": 1.0614, + "mean_token_accuracy": 0.6861120462417603, + "num_tokens": 460380598.0, + "step": 17799 + }, + { + "epoch": 1.9547551065231716, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1853013038635254, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7122046947479248, + "num_tokens": 460405826.0, + "step": 17800 + }, + { + "epoch": 1.9548649242257852, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.295302629470825, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7237244844436646, + "num_tokens": 460429248.0, + "step": 17801 + }, + { + "epoch": 1.9549747419283987, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1955366134643555, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7407961487770081, + "num_tokens": 460453280.0, + "step": 17802 + }, + { + "epoch": 1.9550845596310125, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.370917797088623, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7242923974990845, + "num_tokens": 460476265.0, + "step": 17803 + }, + { + "epoch": 1.9551943773336262, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 1.970326542854309, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7154204249382019, + "num_tokens": 460506821.0, + "step": 17804 + }, + { + "epoch": 1.95530419503624, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.0621933937072754, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7047708630561829, + "num_tokens": 460535198.0, + "step": 17805 + }, + { + "epoch": 1.9554140127388535, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.3478739261627197, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7068142890930176, + "num_tokens": 460559697.0, + "step": 17806 + }, + { + "epoch": 1.955523830441467, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1468043327331543, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7063258290290833, + "num_tokens": 460587873.0, + "step": 17807 + }, + { + "epoch": 1.9556336481440808, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.2144737243652344, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6976019144058228, + "num_tokens": 460613492.0, + "step": 17808 + }, + { + "epoch": 1.9557434658466946, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.4934158325195312, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.736828088760376, + "num_tokens": 460633636.0, + "step": 17809 + }, + { + "epoch": 1.955853283549308, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.132805585861206, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.733107328414917, + "num_tokens": 460660312.0, + "step": 17810 + }, + { + "epoch": 1.9559631012519219, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.685455322265625, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.721530020236969, + "num_tokens": 460678619.0, + "step": 17811 + }, + { + "epoch": 1.9560729189545354, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.5020456314086914, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7337747812271118, + "num_tokens": 460699511.0, + "step": 17812 + }, + { + "epoch": 1.9561827366571491, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1657915115356445, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7221633791923523, + "num_tokens": 460726563.0, + "step": 17813 + }, + { + "epoch": 1.956292554359763, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.011674165725708, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.73527991771698, + "num_tokens": 460755421.0, + "step": 17814 + }, + { + "epoch": 1.9564023720623764, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.152651786804199, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7200902700424194, + "num_tokens": 460781504.0, + "step": 17815 + }, + { + "epoch": 1.95651218976499, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.1829288005828857, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7246955633163452, + "num_tokens": 460808805.0, + "step": 17816 + }, + { + "epoch": 1.9566220074676037, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.096371650695801, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7113665342330933, + "num_tokens": 460836943.0, + "step": 17817 + }, + { + "epoch": 1.9567318251702175, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.21506929397583, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.707813560962677, + "num_tokens": 460863149.0, + "step": 17818 + }, + { + "epoch": 1.9568416428728312, + "ewc_loss": 0.00014972686767578125, + "grad_norm": 2.123795747756958, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7320082783699036, + "num_tokens": 460891102.0, + "step": 17819 + }, + { + "epoch": 1.9569514605754448, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.147012710571289, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7075359225273132, + "num_tokens": 460917830.0, + "step": 17820 + }, + { + "epoch": 1.9570612782780583, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.160850763320923, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7111284732818604, + "num_tokens": 460943706.0, + "step": 17821 + }, + { + "epoch": 1.957171095980672, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.0946152210235596, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7354826927185059, + "num_tokens": 460970926.0, + "step": 17822 + }, + { + "epoch": 1.9572809136832858, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.167275905609131, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7295855283737183, + "num_tokens": 460995967.0, + "step": 17823 + }, + { + "epoch": 1.9573907313858994, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.02327561378479, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7133383750915527, + "num_tokens": 461026643.0, + "step": 17824 + }, + { + "epoch": 1.9575005490885131, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.26528000831604, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7092214226722717, + "num_tokens": 461052393.0, + "step": 17825 + }, + { + "epoch": 1.9576103667911267, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.003577470779419, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7120466232299805, + "num_tokens": 461083688.0, + "step": 17826 + }, + { + "epoch": 1.9577201844937404, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.348468780517578, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.71571946144104, + "num_tokens": 461106965.0, + "step": 17827 + }, + { + "epoch": 1.9578300021963542, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.182607412338257, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6905757188796997, + "num_tokens": 461134676.0, + "step": 17828 + }, + { + "epoch": 1.9579398198989677, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1892476081848145, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6991732120513916, + "num_tokens": 461163256.0, + "step": 17829 + }, + { + "epoch": 1.9580496376015812, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.291271924972534, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7354490160942078, + "num_tokens": 461186258.0, + "step": 17830 + }, + { + "epoch": 1.958159455304195, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.0202929973602295, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.70585036277771, + "num_tokens": 461218074.0, + "step": 17831 + }, + { + "epoch": 1.9582692730068088, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1628024578094482, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7279850840568542, + "num_tokens": 461243234.0, + "step": 17832 + }, + { + "epoch": 1.9583790907094225, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.0550522804260254, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7120870351791382, + "num_tokens": 461272954.0, + "step": 17833 + }, + { + "epoch": 1.958488908412036, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.3309109210968018, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7402828931808472, + "num_tokens": 461295956.0, + "step": 17834 + }, + { + "epoch": 1.9585987261146496, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1478538513183594, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.718238353729248, + "num_tokens": 461323378.0, + "step": 17835 + }, + { + "epoch": 1.9587085438172633, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1514816284179688, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7121429443359375, + "num_tokens": 461349861.0, + "step": 17836 + }, + { + "epoch": 1.958818361519877, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.5651402473449707, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.731268048286438, + "num_tokens": 461370899.0, + "step": 17837 + }, + { + "epoch": 1.9589281792224906, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.285027503967285, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7343341112136841, + "num_tokens": 461397994.0, + "step": 17838 + }, + { + "epoch": 1.9590379969251042, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.4390814304351807, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7221375107765198, + "num_tokens": 461421247.0, + "step": 17839 + }, + { + "epoch": 1.959147814627718, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.3854451179504395, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7155182361602783, + "num_tokens": 461443922.0, + "step": 17840 + }, + { + "epoch": 1.9592576323303317, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.441032648086548, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7233278155326843, + "num_tokens": 461466502.0, + "step": 17841 + }, + { + "epoch": 1.9593674500329454, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 1.9596598148345947, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7008659839630127, + "num_tokens": 461502352.0, + "step": 17842 + }, + { + "epoch": 1.959477267735559, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.245574951171875, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7266046404838562, + "num_tokens": 461526845.0, + "step": 17843 + }, + { + "epoch": 1.9595870854381725, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1322836875915527, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7372872233390808, + "num_tokens": 461552144.0, + "step": 17844 + }, + { + "epoch": 1.9596969031407863, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.035672426223755, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7233850955963135, + "num_tokens": 461581473.0, + "step": 17845 + }, + { + "epoch": 1.9598067208434, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.178816318511963, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7304157614707947, + "num_tokens": 461607343.0, + "step": 17846 + }, + { + "epoch": 1.9599165385460138, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.129343032836914, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7155023217201233, + "num_tokens": 461634676.0, + "step": 17847 + }, + { + "epoch": 1.9600263562486273, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 1.9777531623840332, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7072793245315552, + "num_tokens": 461667066.0, + "step": 17848 + }, + { + "epoch": 1.9601361739512408, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.5440213680267334, + "learning_rate": 1e-06, + "loss": 0.7249, + "mean_token_accuracy": 0.7704769372940063, + "num_tokens": 461687263.0, + "step": 17849 + }, + { + "epoch": 1.9602459916538546, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1406760215759277, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7389780282974243, + "num_tokens": 461712768.0, + "step": 17850 + }, + { + "epoch": 1.9603558093564684, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.2459192276000977, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6988333463668823, + "num_tokens": 461738675.0, + "step": 17851 + }, + { + "epoch": 1.960465627059082, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.2526612281799316, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7174240946769714, + "num_tokens": 461764686.0, + "step": 17852 + }, + { + "epoch": 1.9605754447616954, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1590027809143066, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7203478217124939, + "num_tokens": 461792000.0, + "step": 17853 + }, + { + "epoch": 1.9606852624643092, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.421583414077759, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7176833152770996, + "num_tokens": 461815714.0, + "step": 17854 + }, + { + "epoch": 1.960795080166923, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.249774932861328, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7156277894973755, + "num_tokens": 461840276.0, + "step": 17855 + }, + { + "epoch": 1.9609048978695367, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.0133025646209717, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6943541169166565, + "num_tokens": 461871940.0, + "step": 17856 + }, + { + "epoch": 1.9610147155721502, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.1341376304626465, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7363216280937195, + "num_tokens": 461900224.0, + "step": 17857 + }, + { + "epoch": 1.9611245332747638, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.242044687271118, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7195459008216858, + "num_tokens": 461925930.0, + "step": 17858 + }, + { + "epoch": 1.9612343509773775, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.305194616317749, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7223944664001465, + "num_tokens": 461951480.0, + "step": 17859 + }, + { + "epoch": 1.9613441686799913, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2494277954101562, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7183843851089478, + "num_tokens": 461975796.0, + "step": 17860 + }, + { + "epoch": 1.961453986382605, + "ewc_loss": 0.0001506805419921875, + "grad_norm": 2.555929660797119, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.717928409576416, + "num_tokens": 461996344.0, + "step": 17861 + }, + { + "epoch": 1.9615638040852186, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.159358263015747, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7325230836868286, + "num_tokens": 462023553.0, + "step": 17862 + }, + { + "epoch": 1.9616736217878321, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.093125820159912, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7048267126083374, + "num_tokens": 462049915.0, + "step": 17863 + }, + { + "epoch": 1.9617834394904459, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.110738754272461, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7040234208106995, + "num_tokens": 462079966.0, + "step": 17864 + }, + { + "epoch": 1.9618932571930596, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.406200408935547, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7151139378547668, + "num_tokens": 462103348.0, + "step": 17865 + }, + { + "epoch": 1.9620030748956732, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2183916568756104, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7194513082504272, + "num_tokens": 462131788.0, + "step": 17866 + }, + { + "epoch": 1.9621128925982867, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.485997438430786, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7189037799835205, + "num_tokens": 462154304.0, + "step": 17867 + }, + { + "epoch": 1.9622227103009005, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1776015758514404, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7330825924873352, + "num_tokens": 462180987.0, + "step": 17868 + }, + { + "epoch": 1.9623325280035142, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.083099842071533, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.717296838760376, + "num_tokens": 462210465.0, + "step": 17869 + }, + { + "epoch": 1.962442345706128, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.206322193145752, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7153054475784302, + "num_tokens": 462236827.0, + "step": 17870 + }, + { + "epoch": 1.9625521634087415, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.106492280960083, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7231839299201965, + "num_tokens": 462264662.0, + "step": 17871 + }, + { + "epoch": 1.962661981111355, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3764491081237793, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7259981632232666, + "num_tokens": 462288633.0, + "step": 17872 + }, + { + "epoch": 1.9627717988139688, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1612801551818848, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7096070051193237, + "num_tokens": 462315207.0, + "step": 17873 + }, + { + "epoch": 1.9628816165165826, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.219660520553589, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7065303921699524, + "num_tokens": 462343361.0, + "step": 17874 + }, + { + "epoch": 1.962991434219196, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4360311031341553, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7148042917251587, + "num_tokens": 462365597.0, + "step": 17875 + }, + { + "epoch": 1.9631012519218098, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.254955530166626, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7032105922698975, + "num_tokens": 462391846.0, + "step": 17876 + }, + { + "epoch": 1.9632110696244234, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.11517596244812, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7173893451690674, + "num_tokens": 462418944.0, + "step": 17877 + }, + { + "epoch": 1.9633208873270371, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.253293991088867, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7138948440551758, + "num_tokens": 462446632.0, + "step": 17878 + }, + { + "epoch": 1.963430705029651, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.33502459526062, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7317763566970825, + "num_tokens": 462471488.0, + "step": 17879 + }, + { + "epoch": 1.9635405227322644, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.12400221824646, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.692601203918457, + "num_tokens": 462499777.0, + "step": 17880 + }, + { + "epoch": 1.963650340434878, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4335787296295166, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7170490026473999, + "num_tokens": 462521390.0, + "step": 17881 + }, + { + "epoch": 1.9637601581374917, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.225464105606079, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7248920202255249, + "num_tokens": 462548402.0, + "step": 17882 + }, + { + "epoch": 1.9638699758401055, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.06522274017334, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7220009565353394, + "num_tokens": 462579022.0, + "step": 17883 + }, + { + "epoch": 1.9639797935427192, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.502021551132202, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7381299734115601, + "num_tokens": 462600350.0, + "step": 17884 + }, + { + "epoch": 1.9640896112453328, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1074862480163574, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7152373194694519, + "num_tokens": 462628836.0, + "step": 17885 + }, + { + "epoch": 1.9641994289479463, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.591266393661499, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7286142110824585, + "num_tokens": 462648404.0, + "step": 17886 + }, + { + "epoch": 1.96430924665056, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.034888744354248, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.732012152671814, + "num_tokens": 462675662.0, + "step": 17887 + }, + { + "epoch": 1.9644190643531738, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1454787254333496, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7072461843490601, + "num_tokens": 462702373.0, + "step": 17888 + }, + { + "epoch": 1.9645288820557874, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4370648860931396, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7256067991256714, + "num_tokens": 462724181.0, + "step": 17889 + }, + { + "epoch": 1.964638699758401, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.164334535598755, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.711350679397583, + "num_tokens": 462751807.0, + "step": 17890 + }, + { + "epoch": 1.9647485174610146, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.027507781982422, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7179652452468872, + "num_tokens": 462783331.0, + "step": 17891 + }, + { + "epoch": 1.9648583351636284, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.255120277404785, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7329632043838501, + "num_tokens": 462812266.0, + "step": 17892 + }, + { + "epoch": 1.9649681528662422, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.202357053756714, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7233150005340576, + "num_tokens": 462841811.0, + "step": 17893 + }, + { + "epoch": 1.9650779705688557, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0372939109802246, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7207491397857666, + "num_tokens": 462870788.0, + "step": 17894 + }, + { + "epoch": 1.9651877882714692, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1673059463500977, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6966840028762817, + "num_tokens": 462899120.0, + "step": 17895 + }, + { + "epoch": 1.965297605974083, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9626106023788452, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7309006452560425, + "num_tokens": 462928270.0, + "step": 17896 + }, + { + "epoch": 1.9654074236766967, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9360196590423584, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.7013527154922485, + "num_tokens": 462960055.0, + "step": 17897 + }, + { + "epoch": 1.9655172413793105, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1435599327087402, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7219031453132629, + "num_tokens": 462989831.0, + "step": 17898 + }, + { + "epoch": 1.965627059081924, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.5315351486206055, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7017079591751099, + "num_tokens": 463011219.0, + "step": 17899 + }, + { + "epoch": 1.9657368767845376, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.85878324508667, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.694467306137085, + "num_tokens": 463047470.0, + "step": 17900 + }, + { + "epoch": 1.9658466944871513, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.56850004196167, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7239309549331665, + "num_tokens": 463068746.0, + "step": 17901 + }, + { + "epoch": 1.965956512189765, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.265049695968628, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7271594405174255, + "num_tokens": 463095132.0, + "step": 17902 + }, + { + "epoch": 1.9660663298923786, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9532670974731445, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7331763505935669, + "num_tokens": 463124970.0, + "step": 17903 + }, + { + "epoch": 1.9661761475949922, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1360671520233154, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7152310609817505, + "num_tokens": 463154144.0, + "step": 17904 + }, + { + "epoch": 1.966285965297606, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4044899940490723, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7154263854026794, + "num_tokens": 463179158.0, + "step": 17905 + }, + { + "epoch": 1.9663957830002197, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0363221168518066, + "learning_rate": 1e-06, + "loss": 0.7858, + "mean_token_accuracy": 0.7504950165748596, + "num_tokens": 463206314.0, + "step": 17906 + }, + { + "epoch": 1.9665056007028334, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.59193754196167, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7203973531723022, + "num_tokens": 463228266.0, + "step": 17907 + }, + { + "epoch": 1.966615418405447, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.186570644378662, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7060393691062927, + "num_tokens": 463255744.0, + "step": 17908 + }, + { + "epoch": 1.9667252361080605, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.037532329559326, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7117974162101746, + "num_tokens": 463287159.0, + "step": 17909 + }, + { + "epoch": 1.9668350538106742, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.128349781036377, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7457352876663208, + "num_tokens": 463313671.0, + "step": 17910 + }, + { + "epoch": 1.966944871513288, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.295682191848755, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7157395482063293, + "num_tokens": 463338766.0, + "step": 17911 + }, + { + "epoch": 1.9670546892159018, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1230475902557373, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.726633608341217, + "num_tokens": 463365966.0, + "step": 17912 + }, + { + "epoch": 1.9671645069185153, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2550997734069824, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7167138457298279, + "num_tokens": 463389649.0, + "step": 17913 + }, + { + "epoch": 1.9672743246211288, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.101746082305908, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7148061394691467, + "num_tokens": 463418898.0, + "step": 17914 + }, + { + "epoch": 1.9673841423237426, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4086766242980957, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7127165794372559, + "num_tokens": 463442296.0, + "step": 17915 + }, + { + "epoch": 1.9674939600263563, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0787079334259033, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7152719497680664, + "num_tokens": 463470572.0, + "step": 17916 + }, + { + "epoch": 1.9676037777289699, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0370664596557617, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7124641537666321, + "num_tokens": 463501542.0, + "step": 17917 + }, + { + "epoch": 1.9677135954315834, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2693159580230713, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7014011144638062, + "num_tokens": 463528562.0, + "step": 17918 + }, + { + "epoch": 1.9678234131341972, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1056432723999023, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7220131158828735, + "num_tokens": 463557046.0, + "step": 17919 + }, + { + "epoch": 1.967933230836811, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2779314517974854, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7261285781860352, + "num_tokens": 463581800.0, + "step": 17920 + }, + { + "epoch": 1.9680430485394247, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.208312511444092, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7222833037376404, + "num_tokens": 463607472.0, + "step": 17921 + }, + { + "epoch": 1.9681528662420382, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0981075763702393, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.699782133102417, + "num_tokens": 463635685.0, + "step": 17922 + }, + { + "epoch": 1.9682626839446518, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0459654331207275, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6956156492233276, + "num_tokens": 463667783.0, + "step": 17923 + }, + { + "epoch": 1.9683725016472655, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.047375202178955, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7181946635246277, + "num_tokens": 463699725.0, + "step": 17924 + }, + { + "epoch": 1.9684823193498793, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3962883949279785, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.734932005405426, + "num_tokens": 463721407.0, + "step": 17925 + }, + { + "epoch": 1.9685921370524928, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2283923625946045, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.708986759185791, + "num_tokens": 463749097.0, + "step": 17926 + }, + { + "epoch": 1.9687019547551066, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1109113693237305, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7238545417785645, + "num_tokens": 463777892.0, + "step": 17927 + }, + { + "epoch": 1.96881177245772, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2561607360839844, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7229562997817993, + "num_tokens": 463802494.0, + "step": 17928 + }, + { + "epoch": 1.9689215901603339, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1896767616271973, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7116501331329346, + "num_tokens": 463829345.0, + "step": 17929 + }, + { + "epoch": 1.9690314078629476, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.355160713195801, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.713736891746521, + "num_tokens": 463852750.0, + "step": 17930 + }, + { + "epoch": 1.9691412255655611, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.290677309036255, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7374783158302307, + "num_tokens": 463876341.0, + "step": 17931 + }, + { + "epoch": 1.9692510432681747, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2067322731018066, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.695456862449646, + "num_tokens": 463903977.0, + "step": 17932 + }, + { + "epoch": 1.9693608609707884, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9788471460342407, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7045462131500244, + "num_tokens": 463935305.0, + "step": 17933 + }, + { + "epoch": 1.9694706786734022, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.105323314666748, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7242681980133057, + "num_tokens": 463964995.0, + "step": 17934 + }, + { + "epoch": 1.969580496376016, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1287240982055664, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7164438962936401, + "num_tokens": 463995158.0, + "step": 17935 + }, + { + "epoch": 1.9696903140786295, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1682920455932617, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.717950701713562, + "num_tokens": 464022617.0, + "step": 17936 + }, + { + "epoch": 1.969800131781243, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.454489231109619, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7279566526412964, + "num_tokens": 464044568.0, + "step": 17937 + }, + { + "epoch": 1.9699099494838568, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3749608993530273, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7307999134063721, + "num_tokens": 464066711.0, + "step": 17938 + }, + { + "epoch": 1.9700197671864705, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0173072814941406, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7141366004943848, + "num_tokens": 464096915.0, + "step": 17939 + }, + { + "epoch": 1.970129584889084, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.164015769958496, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7161109447479248, + "num_tokens": 464123541.0, + "step": 17940 + }, + { + "epoch": 1.9702394025916978, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.151231527328491, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7097620368003845, + "num_tokens": 464151300.0, + "step": 17941 + }, + { + "epoch": 1.9703492202943114, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1550240516662598, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7117621898651123, + "num_tokens": 464178356.0, + "step": 17942 + }, + { + "epoch": 1.9704590379969251, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2523767948150635, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6931849718093872, + "num_tokens": 464206079.0, + "step": 17943 + }, + { + "epoch": 1.9705688556995389, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9494571685791016, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6891726851463318, + "num_tokens": 464238899.0, + "step": 17944 + }, + { + "epoch": 1.9706786734021524, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.074370861053467, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.70834881067276, + "num_tokens": 464269439.0, + "step": 17945 + }, + { + "epoch": 1.970788491104766, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.35463809967041, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7261318564414978, + "num_tokens": 464292768.0, + "step": 17946 + }, + { + "epoch": 1.9708983088073797, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.106858015060425, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7173621654510498, + "num_tokens": 464322796.0, + "step": 17947 + }, + { + "epoch": 1.9710081265099935, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0002079010009766, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.715185284614563, + "num_tokens": 464353679.0, + "step": 17948 + }, + { + "epoch": 1.9711179442126072, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.710787773132324, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7282009124755859, + "num_tokens": 464372732.0, + "step": 17949 + }, + { + "epoch": 1.9712277619152208, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9206149578094482, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7312049269676208, + "num_tokens": 464405373.0, + "step": 17950 + }, + { + "epoch": 1.9713375796178343, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.376638650894165, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7104878425598145, + "num_tokens": 464430443.0, + "step": 17951 + }, + { + "epoch": 1.971447397320448, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.139622926712036, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7022226452827454, + "num_tokens": 464457515.0, + "step": 17952 + }, + { + "epoch": 1.9715572150230618, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.689742088317871, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7154675722122192, + "num_tokens": 464477028.0, + "step": 17953 + }, + { + "epoch": 1.9716670327256753, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.178001880645752, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7163766622543335, + "num_tokens": 464502140.0, + "step": 17954 + }, + { + "epoch": 1.9717768504282889, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0610697269439697, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7254559993743896, + "num_tokens": 464530992.0, + "step": 17955 + }, + { + "epoch": 1.9718866681309026, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0236167907714844, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6917657852172852, + "num_tokens": 464565271.0, + "step": 17956 + }, + { + "epoch": 1.9719964858335164, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1899068355560303, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7452466487884521, + "num_tokens": 464592186.0, + "step": 17957 + }, + { + "epoch": 1.9721063035361301, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2626490592956543, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7290797233581543, + "num_tokens": 464616573.0, + "step": 17958 + }, + { + "epoch": 1.9722161212387437, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.354905843734741, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7012386322021484, + "num_tokens": 464641020.0, + "step": 17959 + }, + { + "epoch": 1.9723259389413572, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.05064058303833, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7372039556503296, + "num_tokens": 464671457.0, + "step": 17960 + }, + { + "epoch": 1.972435756643971, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9948476552963257, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7321358919143677, + "num_tokens": 464701479.0, + "step": 17961 + }, + { + "epoch": 1.9725455743465847, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1044225692749023, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.708169162273407, + "num_tokens": 464730943.0, + "step": 17962 + }, + { + "epoch": 1.9726553920491985, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4864115715026855, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7115215063095093, + "num_tokens": 464753566.0, + "step": 17963 + }, + { + "epoch": 1.972765209751812, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2687432765960693, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6975013017654419, + "num_tokens": 464781410.0, + "step": 17964 + }, + { + "epoch": 1.9728750274544256, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.328401803970337, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7259646654129028, + "num_tokens": 464806484.0, + "step": 17965 + }, + { + "epoch": 1.9729848451570393, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.188605308532715, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7121307849884033, + "num_tokens": 464833348.0, + "step": 17966 + }, + { + "epoch": 1.973094662859653, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4152705669403076, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.711579442024231, + "num_tokens": 464856986.0, + "step": 17967 + }, + { + "epoch": 1.9732044805622666, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9866042137145996, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7435919642448425, + "num_tokens": 464887986.0, + "step": 17968 + }, + { + "epoch": 1.9733142982648801, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.5637593269348145, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7345666885375977, + "num_tokens": 464907957.0, + "step": 17969 + }, + { + "epoch": 1.973424115967494, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0992608070373535, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.737053632736206, + "num_tokens": 464933222.0, + "step": 17970 + }, + { + "epoch": 1.9735339336701077, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1992571353912354, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7165043354034424, + "num_tokens": 464959829.0, + "step": 17971 + }, + { + "epoch": 1.9736437513727214, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3195412158966064, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7242147326469421, + "num_tokens": 464983039.0, + "step": 17972 + }, + { + "epoch": 1.973753569075335, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0169732570648193, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.700478196144104, + "num_tokens": 465012388.0, + "step": 17973 + }, + { + "epoch": 1.9738633867779485, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0295615196228027, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7044566869735718, + "num_tokens": 465042863.0, + "step": 17974 + }, + { + "epoch": 1.9739732044805622, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.248288154602051, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7272803783416748, + "num_tokens": 465066140.0, + "step": 17975 + }, + { + "epoch": 1.974083022183176, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2220067977905273, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7298179864883423, + "num_tokens": 465091795.0, + "step": 17976 + }, + { + "epoch": 1.9741928398857898, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4361460208892822, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7316966652870178, + "num_tokens": 465115843.0, + "step": 17977 + }, + { + "epoch": 1.9743026575884033, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9953142404556274, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7181932926177979, + "num_tokens": 465143631.0, + "step": 17978 + }, + { + "epoch": 1.9744124752910168, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.214677333831787, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.6983053684234619, + "num_tokens": 465170006.0, + "step": 17979 + }, + { + "epoch": 1.9745222929936306, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.135145664215088, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.719956636428833, + "num_tokens": 465196759.0, + "step": 17980 + }, + { + "epoch": 1.9746321106962443, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0861294269561768, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7456651329994202, + "num_tokens": 465223186.0, + "step": 17981 + }, + { + "epoch": 1.9747419283988579, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.430856466293335, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7161704301834106, + "num_tokens": 465244970.0, + "step": 17982 + }, + { + "epoch": 1.9748517461014714, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3124892711639404, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.730018138885498, + "num_tokens": 465267666.0, + "step": 17983 + }, + { + "epoch": 1.9749615638040852, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.939426064491272, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6934312582015991, + "num_tokens": 465300502.0, + "step": 17984 + }, + { + "epoch": 1.975071381506699, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.119821548461914, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6903420686721802, + "num_tokens": 465330538.0, + "step": 17985 + }, + { + "epoch": 1.9751811992093127, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1542117595672607, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7213876843452454, + "num_tokens": 465356579.0, + "step": 17986 + }, + { + "epoch": 1.9752910169119262, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.557940721511841, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7343225479125977, + "num_tokens": 465375455.0, + "step": 17987 + }, + { + "epoch": 1.9754008346145397, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.076441526412964, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6884604692459106, + "num_tokens": 465406089.0, + "step": 17988 + }, + { + "epoch": 1.9755106523171535, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4070563316345215, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7405344247817993, + "num_tokens": 465428073.0, + "step": 17989 + }, + { + "epoch": 1.9756204700197673, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1305510997772217, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7147363424301147, + "num_tokens": 465454586.0, + "step": 17990 + }, + { + "epoch": 1.9757302877223808, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.22631573677063, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.718180775642395, + "num_tokens": 465479020.0, + "step": 17991 + }, + { + "epoch": 1.9758401054249946, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.476273536682129, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7295955419540405, + "num_tokens": 465499620.0, + "step": 17992 + }, + { + "epoch": 1.975949923127608, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3344974517822266, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7286274433135986, + "num_tokens": 465523489.0, + "step": 17993 + }, + { + "epoch": 1.9760597408302218, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2910611629486084, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.727170467376709, + "num_tokens": 465550394.0, + "step": 17994 + }, + { + "epoch": 1.9761695585328356, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4033873081207275, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7353076338768005, + "num_tokens": 465574074.0, + "step": 17995 + }, + { + "epoch": 1.9762793762354491, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.8014439344406128, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7369496822357178, + "num_tokens": 465610552.0, + "step": 17996 + }, + { + "epoch": 1.9763891939380627, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2747673988342285, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.707060694694519, + "num_tokens": 465638687.0, + "step": 17997 + }, + { + "epoch": 1.9764990116406764, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.680823802947998, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.720002293586731, + "num_tokens": 465659469.0, + "step": 17998 + }, + { + "epoch": 1.9766088293432902, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.001554489135742, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.715674877166748, + "num_tokens": 465690690.0, + "step": 17999 + }, + { + "epoch": 1.976718647045904, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.433908700942993, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7325592041015625, + "num_tokens": 465713521.0, + "step": 18000 + }, + { + "epoch": 1.9768284647485175, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0920259952545166, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7086973786354065, + "num_tokens": 465741454.0, + "step": 18001 + }, + { + "epoch": 1.976938282451131, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.221158027648926, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7210613489151001, + "num_tokens": 465767905.0, + "step": 18002 + }, + { + "epoch": 1.9770481001537448, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1802966594696045, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6941243410110474, + "num_tokens": 465795347.0, + "step": 18003 + }, + { + "epoch": 1.9771579178563585, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.8096442222595215, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7280038595199585, + "num_tokens": 465812978.0, + "step": 18004 + }, + { + "epoch": 1.977267735558972, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.252068519592285, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7104206085205078, + "num_tokens": 465838833.0, + "step": 18005 + }, + { + "epoch": 1.9773775532615858, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1695430278778076, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7397828102111816, + "num_tokens": 465865017.0, + "step": 18006 + }, + { + "epoch": 1.9774873709641994, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.304980516433716, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7102393507957458, + "num_tokens": 465887636.0, + "step": 18007 + }, + { + "epoch": 1.977597188666813, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.023106098175049, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.698259711265564, + "num_tokens": 465917893.0, + "step": 18008 + }, + { + "epoch": 1.9777070063694269, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.07065749168396, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7258349657058716, + "num_tokens": 465946956.0, + "step": 18009 + }, + { + "epoch": 1.9778168240720404, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3318750858306885, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7257322669029236, + "num_tokens": 465971921.0, + "step": 18010 + }, + { + "epoch": 1.977926641774654, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2077431678771973, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7175185680389404, + "num_tokens": 465998156.0, + "step": 18011 + }, + { + "epoch": 1.9780364594772677, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9813306331634521, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7144314050674438, + "num_tokens": 466029346.0, + "step": 18012 + }, + { + "epoch": 1.9781462771798815, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9919373989105225, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7306689023971558, + "num_tokens": 466059698.0, + "step": 18013 + }, + { + "epoch": 1.9782560948824952, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.37796688079834, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.7018253803253174, + "num_tokens": 466084530.0, + "step": 18014 + }, + { + "epoch": 1.9783659125851087, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9173625707626343, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6927905082702637, + "num_tokens": 466118536.0, + "step": 18015 + }, + { + "epoch": 1.9784757302877223, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2929489612579346, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.731389045715332, + "num_tokens": 466142907.0, + "step": 18016 + }, + { + "epoch": 1.978585547990336, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9534265995025635, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7236998677253723, + "num_tokens": 466175850.0, + "step": 18017 + }, + { + "epoch": 1.9786953656929498, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3899106979370117, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6976420879364014, + "num_tokens": 466201036.0, + "step": 18018 + }, + { + "epoch": 1.9788051833955633, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.393310546875, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7309795618057251, + "num_tokens": 466221700.0, + "step": 18019 + }, + { + "epoch": 1.9789150010981769, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.6412465572357178, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7218209505081177, + "num_tokens": 466241325.0, + "step": 18020 + }, + { + "epoch": 1.9790248188007906, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9993330240249634, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7135323286056519, + "num_tokens": 466272510.0, + "step": 18021 + }, + { + "epoch": 1.9791346365034044, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2011144161224365, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7161766886711121, + "num_tokens": 466297912.0, + "step": 18022 + }, + { + "epoch": 1.9792444542060181, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2032275199890137, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7151010036468506, + "num_tokens": 466324169.0, + "step": 18023 + }, + { + "epoch": 1.9793542719086317, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1108484268188477, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7119706273078918, + "num_tokens": 466351938.0, + "step": 18024 + }, + { + "epoch": 1.9794640896112452, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1428825855255127, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7115921974182129, + "num_tokens": 466381266.0, + "step": 18025 + }, + { + "epoch": 1.979573907313859, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0950539112091064, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7308048009872437, + "num_tokens": 466408555.0, + "step": 18026 + }, + { + "epoch": 1.9796837250164727, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4950103759765625, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7206225991249084, + "num_tokens": 466429737.0, + "step": 18027 + }, + { + "epoch": 1.9797935427190865, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2263619899749756, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7215915322303772, + "num_tokens": 466454995.0, + "step": 18028 + }, + { + "epoch": 1.9799033604217, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0386180877685547, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7067404389381409, + "num_tokens": 466485326.0, + "step": 18029 + }, + { + "epoch": 1.9800131781243135, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9719839096069336, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.6996139287948608, + "num_tokens": 466517105.0, + "step": 18030 + }, + { + "epoch": 1.9801229958269273, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3083741664886475, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7315056324005127, + "num_tokens": 466542822.0, + "step": 18031 + }, + { + "epoch": 1.980232813529541, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4200239181518555, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7246698141098022, + "num_tokens": 466566406.0, + "step": 18032 + }, + { + "epoch": 1.9803426312321546, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.154076337814331, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.721503496170044, + "num_tokens": 466594203.0, + "step": 18033 + }, + { + "epoch": 1.9804524489347681, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.039712905883789, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.714321494102478, + "num_tokens": 466624685.0, + "step": 18034 + }, + { + "epoch": 1.9805622666373819, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0701229572296143, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.720097005367279, + "num_tokens": 466652801.0, + "step": 18035 + }, + { + "epoch": 1.9806720843399956, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.241011619567871, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7192341089248657, + "num_tokens": 466678910.0, + "step": 18036 + }, + { + "epoch": 1.9807819020426094, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.162715196609497, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7244741320610046, + "num_tokens": 466707742.0, + "step": 18037 + }, + { + "epoch": 1.980891719745223, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.646472692489624, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7099457383155823, + "num_tokens": 466730741.0, + "step": 18038 + }, + { + "epoch": 1.9810015374478365, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.130108118057251, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7119851112365723, + "num_tokens": 466757928.0, + "step": 18039 + }, + { + "epoch": 1.9811113551504502, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3423449993133545, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7151986956596375, + "num_tokens": 466781909.0, + "step": 18040 + }, + { + "epoch": 1.981221172853064, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3668956756591797, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7095497846603394, + "num_tokens": 466807173.0, + "step": 18041 + }, + { + "epoch": 1.9813309905556777, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.327261447906494, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.729661226272583, + "num_tokens": 466829869.0, + "step": 18042 + }, + { + "epoch": 1.9814408082582913, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.071962594985962, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7267602682113647, + "num_tokens": 466857696.0, + "step": 18043 + }, + { + "epoch": 1.9815506259609048, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.3982934951782227, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7340610027313232, + "num_tokens": 466879323.0, + "step": 18044 + }, + { + "epoch": 1.9816604436635186, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2648632526397705, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7184962630271912, + "num_tokens": 466904558.0, + "step": 18045 + }, + { + "epoch": 1.9817702613661323, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.5126547813415527, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7225774526596069, + "num_tokens": 466924408.0, + "step": 18046 + }, + { + "epoch": 1.9818800790687459, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.324272394180298, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7470198273658752, + "num_tokens": 466945415.0, + "step": 18047 + }, + { + "epoch": 1.9819898967713594, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1410810947418213, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7239301800727844, + "num_tokens": 466971936.0, + "step": 18048 + }, + { + "epoch": 1.9820997144739732, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1801772117614746, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7104079723358154, + "num_tokens": 466999690.0, + "step": 18049 + }, + { + "epoch": 1.982209532176587, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.5480382442474365, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7429287433624268, + "num_tokens": 467019527.0, + "step": 18050 + }, + { + "epoch": 1.9823193498792007, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.057786464691162, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7294520139694214, + "num_tokens": 467047899.0, + "step": 18051 + }, + { + "epoch": 1.9824291675818142, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.503873586654663, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7327715754508972, + "num_tokens": 467069050.0, + "step": 18052 + }, + { + "epoch": 1.9825389852844277, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2898950576782227, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7423890829086304, + "num_tokens": 467091624.0, + "step": 18053 + }, + { + "epoch": 1.9826488029870415, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.2195029258728027, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.6980053782463074, + "num_tokens": 467119004.0, + "step": 18054 + }, + { + "epoch": 1.9827586206896552, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.087347984313965, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.684861421585083, + "num_tokens": 467150163.0, + "step": 18055 + }, + { + "epoch": 1.9828684383922688, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.024505853652954, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7422475814819336, + "num_tokens": 467180353.0, + "step": 18056 + }, + { + "epoch": 1.9829782560948825, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.5669708251953125, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7561780214309692, + "num_tokens": 467200651.0, + "step": 18057 + }, + { + "epoch": 1.983088073797496, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.214136838912964, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7381426095962524, + "num_tokens": 467225799.0, + "step": 18058 + }, + { + "epoch": 1.9831978915001098, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.848554849624634, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7281458377838135, + "num_tokens": 467243075.0, + "step": 18059 + }, + { + "epoch": 1.9833077092027236, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.0899879932403564, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6977415084838867, + "num_tokens": 467272737.0, + "step": 18060 + }, + { + "epoch": 1.9834175269053371, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.950093150138855, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7327595353126526, + "num_tokens": 467301245.0, + "step": 18061 + }, + { + "epoch": 1.9835273446079507, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 1.9875426292419434, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7042897939682007, + "num_tokens": 467331176.0, + "step": 18062 + }, + { + "epoch": 1.9836371623105644, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.4638428688049316, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7473771572113037, + "num_tokens": 467350422.0, + "step": 18063 + }, + { + "epoch": 1.9837469800131782, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1321089267730713, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7309987545013428, + "num_tokens": 467377058.0, + "step": 18064 + }, + { + "epoch": 1.983856797715792, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.156777858734131, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7029025554656982, + "num_tokens": 467405378.0, + "step": 18065 + }, + { + "epoch": 1.9839666154184055, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.196885585784912, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7178033590316772, + "num_tokens": 467433938.0, + "step": 18066 + }, + { + "epoch": 1.984076433121019, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.457577705383301, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.6991889476776123, + "num_tokens": 467456400.0, + "step": 18067 + }, + { + "epoch": 1.9841862508236328, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1088945865631104, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7130352258682251, + "num_tokens": 467485415.0, + "step": 18068 + }, + { + "epoch": 1.9842960685262465, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.057136058807373, + "learning_rate": 1e-06, + "loss": 1.0333, + "mean_token_accuracy": 0.6899559497833252, + "num_tokens": 467516613.0, + "step": 18069 + }, + { + "epoch": 1.98440588622886, + "ewc_loss": 0.00015163421630859375, + "grad_norm": 2.1726081371307373, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7220075130462646, + "num_tokens": 467541554.0, + "step": 18070 + }, + { + "epoch": 1.9845157039314738, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.5130016803741455, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7362017035484314, + "num_tokens": 467562551.0, + "step": 18071 + }, + { + "epoch": 1.9846255216340873, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.2735955715179443, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7139378786087036, + "num_tokens": 467587875.0, + "step": 18072 + }, + { + "epoch": 1.984735339336701, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4234771728515625, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7384934425354004, + "num_tokens": 467609473.0, + "step": 18073 + }, + { + "epoch": 1.9848451570393149, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.0878329277038574, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7315228581428528, + "num_tokens": 467638318.0, + "step": 18074 + }, + { + "epoch": 1.9849549747419284, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.0748798847198486, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6935241222381592, + "num_tokens": 467669113.0, + "step": 18075 + }, + { + "epoch": 1.985064792444542, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.310788869857788, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.701810896396637, + "num_tokens": 467695282.0, + "step": 18076 + }, + { + "epoch": 1.9851746101471557, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.407775402069092, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7218013405799866, + "num_tokens": 467718225.0, + "step": 18077 + }, + { + "epoch": 1.9852844278497694, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.5006744861602783, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.6974473595619202, + "num_tokens": 467741915.0, + "step": 18078 + }, + { + "epoch": 1.9853942455523832, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4395880699157715, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7309499979019165, + "num_tokens": 467762792.0, + "step": 18079 + }, + { + "epoch": 1.9855040632549967, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4620378017425537, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7185351848602295, + "num_tokens": 467784467.0, + "step": 18080 + }, + { + "epoch": 1.9856138809576103, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.2489681243896484, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7311460971832275, + "num_tokens": 467809037.0, + "step": 18081 + }, + { + "epoch": 1.985723698660224, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.377805709838867, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7353482246398926, + "num_tokens": 467833099.0, + "step": 18082 + }, + { + "epoch": 1.9858335163628378, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.217256784439087, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7135151624679565, + "num_tokens": 467861147.0, + "step": 18083 + }, + { + "epoch": 1.9859433340654513, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.2796733379364014, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7103517055511475, + "num_tokens": 467886383.0, + "step": 18084 + }, + { + "epoch": 1.9860531517680649, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.708348035812378, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7469159364700317, + "num_tokens": 467904883.0, + "step": 18085 + }, + { + "epoch": 1.9861629694706786, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4623944759368896, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7207999229431152, + "num_tokens": 467927722.0, + "step": 18086 + }, + { + "epoch": 1.9862727871732924, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.2397806644439697, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7397778034210205, + "num_tokens": 467954533.0, + "step": 18087 + }, + { + "epoch": 1.9863826048759061, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.09464168548584, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7493075728416443, + "num_tokens": 467980580.0, + "step": 18088 + }, + { + "epoch": 1.9864924225785197, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.540396213531494, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7412831783294678, + "num_tokens": 468001046.0, + "step": 18089 + }, + { + "epoch": 1.9866022402811332, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4874002933502197, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7351561188697815, + "num_tokens": 468022546.0, + "step": 18090 + }, + { + "epoch": 1.986712057983747, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4623489379882812, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7186920642852783, + "num_tokens": 468045996.0, + "step": 18091 + }, + { + "epoch": 1.9868218756863607, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.1252269744873047, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7374687790870667, + "num_tokens": 468073492.0, + "step": 18092 + }, + { + "epoch": 1.9869316933889745, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4737045764923096, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7061187028884888, + "num_tokens": 468097229.0, + "step": 18093 + }, + { + "epoch": 1.987041511091588, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.1185832023620605, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.6978265643119812, + "num_tokens": 468124449.0, + "step": 18094 + }, + { + "epoch": 1.9871513287942015, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.1231980323791504, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7032322883605957, + "num_tokens": 468153002.0, + "step": 18095 + }, + { + "epoch": 1.9872611464968153, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.376579761505127, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7279250025749207, + "num_tokens": 468175728.0, + "step": 18096 + }, + { + "epoch": 1.987370964199429, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.0815954208374023, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6888065338134766, + "num_tokens": 468208067.0, + "step": 18097 + }, + { + "epoch": 1.9874807819020426, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.540398597717285, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7344108819961548, + "num_tokens": 468229495.0, + "step": 18098 + }, + { + "epoch": 1.9875905996046561, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.21121883392334, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7308465838432312, + "num_tokens": 468256501.0, + "step": 18099 + }, + { + "epoch": 1.9877004173072699, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.3327949047088623, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7398257851600647, + "num_tokens": 468280381.0, + "step": 18100 + }, + { + "epoch": 1.9878102350098836, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.007996082305908, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7111862897872925, + "num_tokens": 468311940.0, + "step": 18101 + }, + { + "epoch": 1.9879200527124974, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4712183475494385, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7212809920310974, + "num_tokens": 468335383.0, + "step": 18102 + }, + { + "epoch": 1.988029870415111, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.37896466255188, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.70721036195755, + "num_tokens": 468358251.0, + "step": 18103 + }, + { + "epoch": 1.9881396881177245, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.3336572647094727, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7491481304168701, + "num_tokens": 468381560.0, + "step": 18104 + }, + { + "epoch": 1.9882495058203382, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.452904224395752, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.706749439239502, + "num_tokens": 468405235.0, + "step": 18105 + }, + { + "epoch": 1.988359323522952, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.1475183963775635, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.727086067199707, + "num_tokens": 468433779.0, + "step": 18106 + }, + { + "epoch": 1.9884691412255655, + "ewc_loss": 0.000152587890625, + "grad_norm": 1.935729742050171, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7316420674324036, + "num_tokens": 468465278.0, + "step": 18107 + }, + { + "epoch": 1.9885789589281793, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4319467544555664, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7365716695785522, + "num_tokens": 468489088.0, + "step": 18108 + }, + { + "epoch": 1.9886887766307928, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.0929300785064697, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7113720178604126, + "num_tokens": 468521256.0, + "step": 18109 + }, + { + "epoch": 1.9887985943334066, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4790759086608887, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7367209792137146, + "num_tokens": 468541124.0, + "step": 18110 + }, + { + "epoch": 1.9889084120360203, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.3569226264953613, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7244746088981628, + "num_tokens": 468565961.0, + "step": 18111 + }, + { + "epoch": 1.9890182297386338, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.0062479972839355, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.709953248500824, + "num_tokens": 468595579.0, + "step": 18112 + }, + { + "epoch": 1.9891280474412474, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.2315738201141357, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7239004373550415, + "num_tokens": 468621118.0, + "step": 18113 + }, + { + "epoch": 1.9892378651438611, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.4939677715301514, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7311395406723022, + "num_tokens": 468641545.0, + "step": 18114 + }, + { + "epoch": 1.989347682846475, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.140052556991577, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7195448875427246, + "num_tokens": 468668558.0, + "step": 18115 + }, + { + "epoch": 1.9894575005490887, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.1263227462768555, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.6998261213302612, + "num_tokens": 468697598.0, + "step": 18116 + }, + { + "epoch": 1.9895673182517022, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.091801166534424, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7087885737419128, + "num_tokens": 468726401.0, + "step": 18117 + }, + { + "epoch": 1.9896771359543157, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.480860710144043, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7087497115135193, + "num_tokens": 468748943.0, + "step": 18118 + }, + { + "epoch": 1.9897869536569295, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.017033815383911, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7158291339874268, + "num_tokens": 468777685.0, + "step": 18119 + }, + { + "epoch": 1.9898967713595432, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.409797430038452, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7237173318862915, + "num_tokens": 468800301.0, + "step": 18120 + }, + { + "epoch": 1.9900065890621568, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.115391492843628, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7236666083335876, + "num_tokens": 468826461.0, + "step": 18121 + }, + { + "epoch": 1.9901164067647705, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.066045045852661, + "learning_rate": 1e-06, + "loss": 1.0526, + "mean_token_accuracy": 0.6963715553283691, + "num_tokens": 468858553.0, + "step": 18122 + }, + { + "epoch": 1.990226224467384, + "ewc_loss": 0.000152587890625, + "grad_norm": 2.003929376602173, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6997032165527344, + "num_tokens": 468888908.0, + "step": 18123 + }, + { + "epoch": 1.9903360421699978, + "ewc_loss": 0.000152587890625, + "grad_norm": 1.9339861869812012, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6907650232315063, + "num_tokens": 468922491.0, + "step": 18124 + }, + { + "epoch": 1.9904458598726116, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9904239177703857, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7098324298858643, + "num_tokens": 468954283.0, + "step": 18125 + }, + { + "epoch": 1.9905556775752251, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3983726501464844, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7125917077064514, + "num_tokens": 468977253.0, + "step": 18126 + }, + { + "epoch": 1.9906654952778386, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.451931953430176, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7305866479873657, + "num_tokens": 468997728.0, + "step": 18127 + }, + { + "epoch": 1.9907753129804524, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2919957637786865, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7219493389129639, + "num_tokens": 469022017.0, + "step": 18128 + }, + { + "epoch": 1.9908851306830662, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.286240577697754, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7229909896850586, + "num_tokens": 469045797.0, + "step": 18129 + }, + { + "epoch": 1.99099494838568, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.104278802871704, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6956034898757935, + "num_tokens": 469077174.0, + "step": 18130 + }, + { + "epoch": 1.9911047660882935, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.105379104614258, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6942716836929321, + "num_tokens": 469104610.0, + "step": 18131 + }, + { + "epoch": 1.991214583790907, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9461183547973633, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7131091356277466, + "num_tokens": 469136746.0, + "step": 18132 + }, + { + "epoch": 1.9913244014935207, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0260837078094482, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.708444356918335, + "num_tokens": 469167149.0, + "step": 18133 + }, + { + "epoch": 1.9914342191961345, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.075965642929077, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7136033177375793, + "num_tokens": 469194948.0, + "step": 18134 + }, + { + "epoch": 1.991544036898748, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1481804847717285, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7155680656433105, + "num_tokens": 469220938.0, + "step": 18135 + }, + { + "epoch": 1.9916538546013618, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.861282229423523, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7241982221603394, + "num_tokens": 469257117.0, + "step": 18136 + }, + { + "epoch": 1.9917636723039753, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1731390953063965, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7179157137870789, + "num_tokens": 469284383.0, + "step": 18137 + }, + { + "epoch": 1.991873490006589, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3245975971221924, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7259819507598877, + "num_tokens": 469307692.0, + "step": 18138 + }, + { + "epoch": 1.9919833077092028, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.370014190673828, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7174892425537109, + "num_tokens": 469330147.0, + "step": 18139 + }, + { + "epoch": 1.9920931254118164, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2024476528167725, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7221359014511108, + "num_tokens": 469355986.0, + "step": 18140 + }, + { + "epoch": 1.99220294311443, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.173048257827759, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7228673696517944, + "num_tokens": 469384324.0, + "step": 18141 + }, + { + "epoch": 1.9923127608170437, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9242290258407593, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7251238226890564, + "num_tokens": 469414795.0, + "step": 18142 + }, + { + "epoch": 1.9924225785196574, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4005212783813477, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7109897136688232, + "num_tokens": 469437569.0, + "step": 18143 + }, + { + "epoch": 1.9925323962222712, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.349236249923706, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7432736158370972, + "num_tokens": 469461788.0, + "step": 18144 + }, + { + "epoch": 1.9926422139248847, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.130617141723633, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7118721008300781, + "num_tokens": 469489058.0, + "step": 18145 + }, + { + "epoch": 1.9927520316274983, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.072103977203369, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6915703415870667, + "num_tokens": 469518606.0, + "step": 18146 + }, + { + "epoch": 1.992861849330112, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.27081298828125, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.72266685962677, + "num_tokens": 469542224.0, + "step": 18147 + }, + { + "epoch": 1.9929716670327258, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.075016498565674, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7159337401390076, + "num_tokens": 469571020.0, + "step": 18148 + }, + { + "epoch": 1.9930814847353393, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.093271255493164, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7256163358688354, + "num_tokens": 469598688.0, + "step": 18149 + }, + { + "epoch": 1.9931913024379528, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.983614206314087, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.725567102432251, + "num_tokens": 469628287.0, + "step": 18150 + }, + { + "epoch": 1.9933011201405666, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9569008350372314, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7069788575172424, + "num_tokens": 469661848.0, + "step": 18151 + }, + { + "epoch": 1.9934109378431804, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0540449619293213, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7271169424057007, + "num_tokens": 469691358.0, + "step": 18152 + }, + { + "epoch": 1.993520755545794, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2338433265686035, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7208092212677002, + "num_tokens": 469719859.0, + "step": 18153 + }, + { + "epoch": 1.9936305732484076, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2716362476348877, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.705963134765625, + "num_tokens": 469744208.0, + "step": 18154 + }, + { + "epoch": 1.9937403909510212, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.053677797317505, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.704555332660675, + "num_tokens": 469775148.0, + "step": 18155 + }, + { + "epoch": 1.993850208653635, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.579831123352051, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7553868293762207, + "num_tokens": 469796709.0, + "step": 18156 + }, + { + "epoch": 1.9939600263562487, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.5203044414520264, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7166200280189514, + "num_tokens": 469819795.0, + "step": 18157 + }, + { + "epoch": 1.9940698440588625, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.236433982849121, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7406466603279114, + "num_tokens": 469844011.0, + "step": 18158 + }, + { + "epoch": 1.994179661761476, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.337275743484497, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.720526933670044, + "num_tokens": 469867553.0, + "step": 18159 + }, + { + "epoch": 1.9942894794640895, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1652865409851074, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7249459028244019, + "num_tokens": 469896905.0, + "step": 18160 + }, + { + "epoch": 1.9943992971667033, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.416614294052124, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7032510638237, + "num_tokens": 469920200.0, + "step": 18161 + }, + { + "epoch": 1.994509114869317, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4967966079711914, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7187358736991882, + "num_tokens": 469944573.0, + "step": 18162 + }, + { + "epoch": 1.9946189325719306, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3228445053100586, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7217340469360352, + "num_tokens": 469967220.0, + "step": 18163 + }, + { + "epoch": 1.994728750274544, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.5381617546081543, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7299467325210571, + "num_tokens": 469987942.0, + "step": 18164 + }, + { + "epoch": 1.9948385679771579, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1290018558502197, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.719687819480896, + "num_tokens": 470017012.0, + "step": 18165 + }, + { + "epoch": 1.9949483856797716, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.212148666381836, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7259466648101807, + "num_tokens": 470042843.0, + "step": 18166 + }, + { + "epoch": 1.9950582033823854, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.150348663330078, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7429347038269043, + "num_tokens": 470069002.0, + "step": 18167 + }, + { + "epoch": 1.995168021084999, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4544425010681152, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7331194281578064, + "num_tokens": 470089815.0, + "step": 18168 + }, + { + "epoch": 1.9952778387876124, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.036635637283325, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7088027000427246, + "num_tokens": 470121345.0, + "step": 18169 + }, + { + "epoch": 1.9953876564902262, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.107105016708374, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7320244312286377, + "num_tokens": 470148531.0, + "step": 18170 + }, + { + "epoch": 1.99549747419284, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.338620185852051, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6986086368560791, + "num_tokens": 470174290.0, + "step": 18171 + }, + { + "epoch": 1.9956072918954535, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2052860260009766, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7152029871940613, + "num_tokens": 470201166.0, + "step": 18172 + }, + { + "epoch": 1.9957171095980673, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.475552797317505, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7226418256759644, + "num_tokens": 470222186.0, + "step": 18173 + }, + { + "epoch": 1.9958269273006808, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9712971448898315, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.734968900680542, + "num_tokens": 470253168.0, + "step": 18174 + }, + { + "epoch": 1.9959367450032945, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.260152578353882, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7158777713775635, + "num_tokens": 470281135.0, + "step": 18175 + }, + { + "epoch": 1.9960465627059083, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.22389817237854, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7147312164306641, + "num_tokens": 470307947.0, + "step": 18176 + }, + { + "epoch": 1.9961563804085218, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0906171798706055, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7510856986045837, + "num_tokens": 470335331.0, + "step": 18177 + }, + { + "epoch": 1.9962661981111354, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.041565179824829, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7107733488082886, + "num_tokens": 470365413.0, + "step": 18178 + }, + { + "epoch": 1.9963760158137491, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0061585903167725, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7031246423721313, + "num_tokens": 470396737.0, + "step": 18179 + }, + { + "epoch": 1.9964858335163629, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.309246778488159, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7113720178604126, + "num_tokens": 470422654.0, + "step": 18180 + }, + { + "epoch": 1.9965956512189766, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1287739276885986, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7280591726303101, + "num_tokens": 470449058.0, + "step": 18181 + }, + { + "epoch": 1.9967054689215902, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3817293643951416, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7403831481933594, + "num_tokens": 470471733.0, + "step": 18182 + }, + { + "epoch": 1.9968152866242037, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4472529888153076, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7179374694824219, + "num_tokens": 470494067.0, + "step": 18183 + }, + { + "epoch": 1.9969251043268175, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1507976055145264, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7176200747489929, + "num_tokens": 470521583.0, + "step": 18184 + }, + { + "epoch": 1.9970349220294312, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2982428073883057, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7157676815986633, + "num_tokens": 470545232.0, + "step": 18185 + }, + { + "epoch": 1.9971447397320448, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.064544439315796, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7245476245880127, + "num_tokens": 470573581.0, + "step": 18186 + }, + { + "epoch": 1.9972545574346585, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.507676362991333, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7271699905395508, + "num_tokens": 470594915.0, + "step": 18187 + }, + { + "epoch": 1.997364375137272, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2656965255737305, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7242270708084106, + "num_tokens": 470619696.0, + "step": 18188 + }, + { + "epoch": 1.9974741928398858, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.155189037322998, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6946436166763306, + "num_tokens": 470648134.0, + "step": 18189 + }, + { + "epoch": 1.9975840105424996, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2427172660827637, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7327835559844971, + "num_tokens": 470672694.0, + "step": 18190 + }, + { + "epoch": 1.997693828245113, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1590495109558105, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7557474374771118, + "num_tokens": 470698698.0, + "step": 18191 + }, + { + "epoch": 1.9978036459477266, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1086294651031494, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7296398878097534, + "num_tokens": 470725993.0, + "step": 18192 + }, + { + "epoch": 1.9979134636503404, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2086896896362305, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7145887613296509, + "num_tokens": 470751735.0, + "step": 18193 + }, + { + "epoch": 1.9980232813529542, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 3.032885789871216, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7387788891792297, + "num_tokens": 470768977.0, + "step": 18194 + }, + { + "epoch": 1.998133099055568, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0215978622436523, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7144108414649963, + "num_tokens": 470797383.0, + "step": 18195 + }, + { + "epoch": 1.9982429167581814, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4610705375671387, + "learning_rate": 1e-06, + "loss": 0.782, + "mean_token_accuracy": 0.7456194162368774, + "num_tokens": 470818262.0, + "step": 18196 + }, + { + "epoch": 1.998352734460795, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1719348430633545, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7088733315467834, + "num_tokens": 470845484.0, + "step": 18197 + }, + { + "epoch": 1.9984625521634087, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.353689193725586, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.6995820999145508, + "num_tokens": 470870283.0, + "step": 18198 + }, + { + "epoch": 1.9985723698660225, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4655375480651855, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7307403087615967, + "num_tokens": 470891092.0, + "step": 18199 + }, + { + "epoch": 1.998682187568636, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.204967975616455, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7272311449050903, + "num_tokens": 470916592.0, + "step": 18200 + }, + { + "epoch": 1.9987920052712496, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1811137199401855, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.739243745803833, + "num_tokens": 470943251.0, + "step": 18201 + }, + { + "epoch": 1.9989018229738633, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2229416370391846, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7349469065666199, + "num_tokens": 470970150.0, + "step": 18202 + }, + { + "epoch": 1.999011640676477, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.965766191482544, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6931681036949158, + "num_tokens": 471001184.0, + "step": 18203 + }, + { + "epoch": 1.9991214583790908, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3232922554016113, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7235739231109619, + "num_tokens": 471025144.0, + "step": 18204 + }, + { + "epoch": 1.9992312760817044, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.929687261581421, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.7058576345443726, + "num_tokens": 471059042.0, + "step": 18205 + }, + { + "epoch": 1.999341093784318, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.001511573791504, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6917412281036377, + "num_tokens": 471090984.0, + "step": 18206 + }, + { + "epoch": 1.9994509114869317, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4262123107910156, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7318830490112305, + "num_tokens": 471111508.0, + "step": 18207 + }, + { + "epoch": 1.9995607291895454, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2700271606445312, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7083932161331177, + "num_tokens": 471136394.0, + "step": 18208 + }, + { + "epoch": 1.9996705468921592, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.202165126800537, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7171175479888916, + "num_tokens": 471164069.0, + "step": 18209 + }, + { + "epoch": 1.9997803645947727, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.3184103965759277, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7195074558258057, + "num_tokens": 471189023.0, + "step": 18210 + }, + { + "epoch": 1.9998901822973862, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4848783016204834, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7347540259361267, + "num_tokens": 471209443.0, + "step": 18211 + }, + { + "epoch": 2.0, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.245919704437256, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7071663737297058, + "num_tokens": 471235014.0, + "step": 18212 + }, + { + "epoch": 2.0001098177026138, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2112131118774414, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7293740510940552, + "num_tokens": 471260654.0, + "step": 18213 + }, + { + "epoch": 2.0002196354052275, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0782275199890137, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7274888753890991, + "num_tokens": 471286421.0, + "step": 18214 + }, + { + "epoch": 2.000329453107841, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.218068838119507, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7279843091964722, + "num_tokens": 471310258.0, + "step": 18215 + }, + { + "epoch": 2.0004392708104546, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0900626182556152, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.722556471824646, + "num_tokens": 471338723.0, + "step": 18216 + }, + { + "epoch": 2.0005490885130683, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.039153575897217, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7299952507019043, + "num_tokens": 471369007.0, + "step": 18217 + }, + { + "epoch": 2.000658906215682, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4381680488586426, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7452568411827087, + "num_tokens": 471390448.0, + "step": 18218 + }, + { + "epoch": 2.0007687239182954, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.036973714828491, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7202451229095459, + "num_tokens": 471420002.0, + "step": 18219 + }, + { + "epoch": 2.000878541620909, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.493375539779663, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7245859503746033, + "num_tokens": 471442498.0, + "step": 18220 + }, + { + "epoch": 2.000988359323523, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.541389226913452, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7336611151695251, + "num_tokens": 471463307.0, + "step": 18221 + }, + { + "epoch": 2.0010981770261367, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2961463928222656, + "learning_rate": 1e-06, + "loss": 0.7519, + "mean_token_accuracy": 0.753495991230011, + "num_tokens": 471487919.0, + "step": 18222 + }, + { + "epoch": 2.0012079947287504, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.216268301010132, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7574820518493652, + "num_tokens": 471514828.0, + "step": 18223 + }, + { + "epoch": 2.0013178124313638, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.491347312927246, + "learning_rate": 1e-06, + "loss": 0.7609, + "mean_token_accuracy": 0.7559542655944824, + "num_tokens": 471537025.0, + "step": 18224 + }, + { + "epoch": 2.0014276301339775, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.50994610786438, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7255059480667114, + "num_tokens": 471557930.0, + "step": 18225 + }, + { + "epoch": 2.0015374478365913, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2665328979492188, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7210662364959717, + "num_tokens": 471582641.0, + "step": 18226 + }, + { + "epoch": 2.001647265539205, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4536831378936768, + "learning_rate": 1e-06, + "loss": 0.7362, + "mean_token_accuracy": 0.7562721371650696, + "num_tokens": 471604738.0, + "step": 18227 + }, + { + "epoch": 2.001757083241819, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.207629919052124, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7014724016189575, + "num_tokens": 471634425.0, + "step": 18228 + }, + { + "epoch": 2.001866900944432, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2312049865722656, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7300125956535339, + "num_tokens": 471660871.0, + "step": 18229 + }, + { + "epoch": 2.001976718647046, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.179382085800171, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7286853790283203, + "num_tokens": 471688335.0, + "step": 18230 + }, + { + "epoch": 2.0020865363496596, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1090667247772217, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.708652675151825, + "num_tokens": 471718337.0, + "step": 18231 + }, + { + "epoch": 2.0021963540522734, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 1.9607278108596802, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7155229449272156, + "num_tokens": 471751966.0, + "step": 18232 + }, + { + "epoch": 2.0023061717548867, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.5261428356170654, + "learning_rate": 1e-06, + "loss": 0.7716, + "mean_token_accuracy": 0.7506772875785828, + "num_tokens": 471773386.0, + "step": 18233 + }, + { + "epoch": 2.0024159894575004, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.494875907897949, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7343564033508301, + "num_tokens": 471795477.0, + "step": 18234 + }, + { + "epoch": 2.002525807160114, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.0741984844207764, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7227680087089539, + "num_tokens": 471825957.0, + "step": 18235 + }, + { + "epoch": 2.002635624862728, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.233762741088867, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7224195003509521, + "num_tokens": 471853767.0, + "step": 18236 + }, + { + "epoch": 2.0027454425653417, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2571969032287598, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7173373699188232, + "num_tokens": 471883082.0, + "step": 18237 + }, + { + "epoch": 2.002855260267955, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.230452537536621, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7365097999572754, + "num_tokens": 471910218.0, + "step": 18238 + }, + { + "epoch": 2.0029650779705688, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.2795016765594482, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7262426614761353, + "num_tokens": 471936993.0, + "step": 18239 + }, + { + "epoch": 2.0030748956731825, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.289705753326416, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7037754058837891, + "num_tokens": 471964226.0, + "step": 18240 + }, + { + "epoch": 2.0031847133757963, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4342174530029297, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7303156852722168, + "num_tokens": 471986803.0, + "step": 18241 + }, + { + "epoch": 2.00329453107841, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.474125623703003, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7302505970001221, + "num_tokens": 472010234.0, + "step": 18242 + }, + { + "epoch": 2.0034043487810234, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.5899295806884766, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7338184118270874, + "num_tokens": 472032126.0, + "step": 18243 + }, + { + "epoch": 2.003514166483637, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.363290548324585, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7240661382675171, + "num_tokens": 472057020.0, + "step": 18244 + }, + { + "epoch": 2.003623984186251, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.5314724445343018, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7199987173080444, + "num_tokens": 472079754.0, + "step": 18245 + }, + { + "epoch": 2.0037338018888646, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.1096811294555664, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.748833954334259, + "num_tokens": 472107357.0, + "step": 18246 + }, + { + "epoch": 2.003843619591478, + "ewc_loss": 0.00015354156494140625, + "grad_norm": 2.4201817512512207, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7394146919250488, + "num_tokens": 472132314.0, + "step": 18247 + }, + { + "epoch": 2.0039534372940917, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.3301432132720947, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7202030420303345, + "num_tokens": 472159056.0, + "step": 18248 + }, + { + "epoch": 2.0040632549967055, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.174534559249878, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7469235062599182, + "num_tokens": 472187610.0, + "step": 18249 + }, + { + "epoch": 2.004173072699319, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.4455344676971436, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7331018447875977, + "num_tokens": 472211027.0, + "step": 18250 + }, + { + "epoch": 2.004282890401933, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.1604855060577393, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.727139949798584, + "num_tokens": 472240380.0, + "step": 18251 + }, + { + "epoch": 2.0043927081045463, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.1160635948181152, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.750130832195282, + "num_tokens": 472270893.0, + "step": 18252 + }, + { + "epoch": 2.00450252580716, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.1866159439086914, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7472720146179199, + "num_tokens": 472296851.0, + "step": 18253 + }, + { + "epoch": 2.004612343509774, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2626607418060303, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7287604808807373, + "num_tokens": 472324270.0, + "step": 18254 + }, + { + "epoch": 2.0047221612123876, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2156434059143066, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7395567893981934, + "num_tokens": 472350052.0, + "step": 18255 + }, + { + "epoch": 2.0048319789150013, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.219496488571167, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7314751148223877, + "num_tokens": 472377760.0, + "step": 18256 + }, + { + "epoch": 2.0049417966176146, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 1.9361679553985596, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7211905717849731, + "num_tokens": 472413478.0, + "step": 18257 + }, + { + "epoch": 2.0050516143202284, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.286081314086914, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7029461860656738, + "num_tokens": 472441679.0, + "step": 18258 + }, + { + "epoch": 2.005161432022842, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2908670902252197, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7460783123970032, + "num_tokens": 472467174.0, + "step": 18259 + }, + { + "epoch": 2.005271249725456, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.332176685333252, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7405035495758057, + "num_tokens": 472491450.0, + "step": 18260 + }, + { + "epoch": 2.005381067428069, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.5421314239501953, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7488235235214233, + "num_tokens": 472513867.0, + "step": 18261 + }, + { + "epoch": 2.005490885130683, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.299384355545044, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7390930652618408, + "num_tokens": 472538909.0, + "step": 18262 + }, + { + "epoch": 2.0056007028332967, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 1.9151920080184937, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.722349226474762, + "num_tokens": 472573516.0, + "step": 18263 + }, + { + "epoch": 2.0057105205359105, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.189314126968384, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7147842049598694, + "num_tokens": 472604040.0, + "step": 18264 + }, + { + "epoch": 2.0058203382385242, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.3541791439056396, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7386559844017029, + "num_tokens": 472628842.0, + "step": 18265 + }, + { + "epoch": 2.0059301559411375, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.4547855854034424, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7084048986434937, + "num_tokens": 472653793.0, + "step": 18266 + }, + { + "epoch": 2.0060399736437513, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2608554363250732, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.705172598361969, + "num_tokens": 472682750.0, + "step": 18267 + }, + { + "epoch": 2.006149791346365, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2837023735046387, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.705187976360321, + "num_tokens": 472708280.0, + "step": 18268 + }, + { + "epoch": 2.006259609048979, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.335411548614502, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7198734283447266, + "num_tokens": 472732535.0, + "step": 18269 + }, + { + "epoch": 2.0063694267515926, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.151563882827759, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7101804614067078, + "num_tokens": 472761119.0, + "step": 18270 + }, + { + "epoch": 2.006479244454206, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.695408344268799, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7420538663864136, + "num_tokens": 472782492.0, + "step": 18271 + }, + { + "epoch": 2.0065890621568196, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.2477707862854004, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7235153913497925, + "num_tokens": 472809592.0, + "step": 18272 + }, + { + "epoch": 2.0066988798594334, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.142915964126587, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7253700494766235, + "num_tokens": 472838346.0, + "step": 18273 + }, + { + "epoch": 2.006808697562047, + "ewc_loss": 0.00015544891357421875, + "grad_norm": 2.1932690143585205, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7136659622192383, + "num_tokens": 472868408.0, + "step": 18274 + }, + { + "epoch": 2.0069185152646605, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4947361946105957, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7443421483039856, + "num_tokens": 472891999.0, + "step": 18275 + }, + { + "epoch": 2.0070283329672742, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.729430913925171, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7468964457511902, + "num_tokens": 472912478.0, + "step": 18276 + }, + { + "epoch": 2.007138150669888, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3938848972320557, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7204784154891968, + "num_tokens": 472937871.0, + "step": 18277 + }, + { + "epoch": 2.0072479683725017, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1577019691467285, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.712329626083374, + "num_tokens": 472968653.0, + "step": 18278 + }, + { + "epoch": 2.0073577860751155, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2801668643951416, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7327532172203064, + "num_tokens": 472993966.0, + "step": 18279 + }, + { + "epoch": 2.007467603777729, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.5146701335906982, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7259130477905273, + "num_tokens": 473019198.0, + "step": 18280 + }, + { + "epoch": 2.0075774214803426, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2085115909576416, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7223799228668213, + "num_tokens": 473047784.0, + "step": 18281 + }, + { + "epoch": 2.0076872391829563, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3599746227264404, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7202295064926147, + "num_tokens": 473073103.0, + "step": 18282 + }, + { + "epoch": 2.00779705688557, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.348388433456421, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7456385493278503, + "num_tokens": 473098209.0, + "step": 18283 + }, + { + "epoch": 2.0079068745881834, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3676698207855225, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.6979783773422241, + "num_tokens": 473123539.0, + "step": 18284 + }, + { + "epoch": 2.008016692290797, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 1.9769777059555054, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7115403413772583, + "num_tokens": 473158145.0, + "step": 18285 + }, + { + "epoch": 2.008126509993411, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1718571186065674, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7207050323486328, + "num_tokens": 473185679.0, + "step": 18286 + }, + { + "epoch": 2.0082363276960247, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.654397964477539, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7390576601028442, + "num_tokens": 473205197.0, + "step": 18287 + }, + { + "epoch": 2.0083461453986384, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.326537847518921, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7391363382339478, + "num_tokens": 473230948.0, + "step": 18288 + }, + { + "epoch": 2.0084559631012517, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.563173532485962, + "learning_rate": 1e-06, + "loss": 0.7531, + "mean_token_accuracy": 0.7594137191772461, + "num_tokens": 473251055.0, + "step": 18289 + }, + { + "epoch": 2.0085657808038655, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4542317390441895, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7348136305809021, + "num_tokens": 473274792.0, + "step": 18290 + }, + { + "epoch": 2.0086755985064793, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0179998874664307, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7524409294128418, + "num_tokens": 473304591.0, + "step": 18291 + }, + { + "epoch": 2.008785416209093, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.190727710723877, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7353590726852417, + "num_tokens": 473333610.0, + "step": 18292 + }, + { + "epoch": 2.0088952339117068, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.754213809967041, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7372366189956665, + "num_tokens": 473352052.0, + "step": 18293 + }, + { + "epoch": 2.00900505161432, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0755059719085693, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7123619914054871, + "num_tokens": 473384444.0, + "step": 18294 + }, + { + "epoch": 2.009114869316934, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0559206008911133, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7212331891059875, + "num_tokens": 473415544.0, + "step": 18295 + }, + { + "epoch": 2.0092246870195476, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2471323013305664, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.724836528301239, + "num_tokens": 473443069.0, + "step": 18296 + }, + { + "epoch": 2.0093345047221614, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.330413579940796, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7405602335929871, + "num_tokens": 473467650.0, + "step": 18297 + }, + { + "epoch": 2.0094443224247747, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.110602617263794, + "learning_rate": 1e-06, + "loss": 0.7649, + "mean_token_accuracy": 0.7405745983123779, + "num_tokens": 473495979.0, + "step": 18298 + }, + { + "epoch": 2.0095541401273884, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2133281230926514, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7428234815597534, + "num_tokens": 473521867.0, + "step": 18299 + }, + { + "epoch": 2.009663957830002, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1320502758026123, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.735715389251709, + "num_tokens": 473551775.0, + "step": 18300 + }, + { + "epoch": 2.009773775532616, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.8422694206237793, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7530109882354736, + "num_tokens": 473570580.0, + "step": 18301 + }, + { + "epoch": 2.0098835932352297, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.744899272918701, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.740497350692749, + "num_tokens": 473589849.0, + "step": 18302 + }, + { + "epoch": 2.009993410937843, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0432310104370117, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7272977828979492, + "num_tokens": 473621943.0, + "step": 18303 + }, + { + "epoch": 2.0101032286404568, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4722883701324463, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7386721968650818, + "num_tokens": 473644412.0, + "step": 18304 + }, + { + "epoch": 2.0102130463430705, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 1.8880988359451294, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7411798238754272, + "num_tokens": 473681233.0, + "step": 18305 + }, + { + "epoch": 2.0103228640456843, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.5346016883850098, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7424108982086182, + "num_tokens": 473703760.0, + "step": 18306 + }, + { + "epoch": 2.010432681748298, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.317847490310669, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7296099662780762, + "num_tokens": 473728373.0, + "step": 18307 + }, + { + "epoch": 2.0105424994509113, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.222177505493164, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7275672554969788, + "num_tokens": 473754675.0, + "step": 18308 + }, + { + "epoch": 2.010652317153525, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4497125148773193, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7372864484786987, + "num_tokens": 473776730.0, + "step": 18309 + }, + { + "epoch": 2.010762134856139, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2073476314544678, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7239819765090942, + "num_tokens": 473806682.0, + "step": 18310 + }, + { + "epoch": 2.0108719525587526, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1945598125457764, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7078406810760498, + "num_tokens": 473837602.0, + "step": 18311 + }, + { + "epoch": 2.010981770261366, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3368358612060547, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7360749244689941, + "num_tokens": 473861516.0, + "step": 18312 + }, + { + "epoch": 2.0110915879639797, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2616422176361084, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7483052015304565, + "num_tokens": 473887367.0, + "step": 18313 + }, + { + "epoch": 2.0112014056665934, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.249636650085449, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7250378131866455, + "num_tokens": 473914911.0, + "step": 18314 + }, + { + "epoch": 2.011311223369207, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.410654067993164, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7239429950714111, + "num_tokens": 473941329.0, + "step": 18315 + }, + { + "epoch": 2.011421041071821, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.7108540534973145, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7247631549835205, + "num_tokens": 473962147.0, + "step": 18316 + }, + { + "epoch": 2.0115308587744343, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3049468994140625, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7206652760505676, + "num_tokens": 473989895.0, + "step": 18317 + }, + { + "epoch": 2.011640676477048, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4566705226898193, + "learning_rate": 1e-06, + "loss": 0.7353, + "mean_token_accuracy": 0.7647997140884399, + "num_tokens": 474010572.0, + "step": 18318 + }, + { + "epoch": 2.011750494179662, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.6098246574401855, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.729719877243042, + "num_tokens": 474032697.0, + "step": 18319 + }, + { + "epoch": 2.0118603118822755, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.026991128921509, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7512508034706116, + "num_tokens": 474063383.0, + "step": 18320 + }, + { + "epoch": 2.0119701295848893, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2092196941375732, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7283717393875122, + "num_tokens": 474091568.0, + "step": 18321 + }, + { + "epoch": 2.0120799472875026, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3802428245544434, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.757867693901062, + "num_tokens": 474113882.0, + "step": 18322 + }, + { + "epoch": 2.0121897649901164, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.18688702583313, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7320458889007568, + "num_tokens": 474143063.0, + "step": 18323 + }, + { + "epoch": 2.01229958269273, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.232804298400879, + "learning_rate": 1e-06, + "loss": 0.7283, + "mean_token_accuracy": 0.7712494730949402, + "num_tokens": 474167338.0, + "step": 18324 + }, + { + "epoch": 2.012409400395344, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.531975507736206, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7450897097587585, + "num_tokens": 474191108.0, + "step": 18325 + }, + { + "epoch": 2.012519218097957, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.318932294845581, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7304036617279053, + "num_tokens": 474216320.0, + "step": 18326 + }, + { + "epoch": 2.012629035800571, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0580556392669678, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7288153171539307, + "num_tokens": 474248732.0, + "step": 18327 + }, + { + "epoch": 2.0127388535031847, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.447676181793213, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7550297379493713, + "num_tokens": 474270619.0, + "step": 18328 + }, + { + "epoch": 2.0128486712057985, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.4887781143188477, + "learning_rate": 1e-06, + "loss": 0.7257, + "mean_token_accuracy": 0.7657907009124756, + "num_tokens": 474293168.0, + "step": 18329 + }, + { + "epoch": 2.0129584889084122, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3661766052246094, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7312021851539612, + "num_tokens": 474318595.0, + "step": 18330 + }, + { + "epoch": 2.0130683066110255, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.23496675491333, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7360051870346069, + "num_tokens": 474346507.0, + "step": 18331 + }, + { + "epoch": 2.0131781243136393, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.268495798110962, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7150135636329651, + "num_tokens": 474374731.0, + "step": 18332 + }, + { + "epoch": 2.013287942016253, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.542962074279785, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7484043836593628, + "num_tokens": 474396677.0, + "step": 18333 + }, + { + "epoch": 2.013397759718867, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2023873329162598, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7209113836288452, + "num_tokens": 474424538.0, + "step": 18334 + }, + { + "epoch": 2.01350757742148, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.6842870712280273, + "learning_rate": 1e-06, + "loss": 0.7723, + "mean_token_accuracy": 0.7551124095916748, + "num_tokens": 474444738.0, + "step": 18335 + }, + { + "epoch": 2.013617395124094, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.6053128242492676, + "learning_rate": 1e-06, + "loss": 0.7407, + "mean_token_accuracy": 0.7597637176513672, + "num_tokens": 474464517.0, + "step": 18336 + }, + { + "epoch": 2.0137272128267076, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1785666942596436, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7238239049911499, + "num_tokens": 474495613.0, + "step": 18337 + }, + { + "epoch": 2.0138370305293214, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.3748395442962646, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7072758078575134, + "num_tokens": 474521657.0, + "step": 18338 + }, + { + "epoch": 2.013946848231935, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.2192275524139404, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.725908637046814, + "num_tokens": 474549732.0, + "step": 18339 + }, + { + "epoch": 2.0140566659345485, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0898597240448, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7285369634628296, + "num_tokens": 474578853.0, + "step": 18340 + }, + { + "epoch": 2.014166483637162, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.0901122093200684, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7240983843803406, + "num_tokens": 474609898.0, + "step": 18341 + }, + { + "epoch": 2.014276301339776, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.6778926849365234, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7514089345932007, + "num_tokens": 474630096.0, + "step": 18342 + }, + { + "epoch": 2.0143861190423897, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.360471725463867, + "learning_rate": 1e-06, + "loss": 0.775, + "mean_token_accuracy": 0.7528198957443237, + "num_tokens": 474655330.0, + "step": 18343 + }, + { + "epoch": 2.0144959367450035, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4437811374664307, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.725442111492157, + "num_tokens": 474679402.0, + "step": 18344 + }, + { + "epoch": 2.014605754447617, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1715307235717773, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7396938800811768, + "num_tokens": 474709822.0, + "step": 18345 + }, + { + "epoch": 2.0147155721502306, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3217310905456543, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7582529187202454, + "num_tokens": 474734308.0, + "step": 18346 + }, + { + "epoch": 2.0148253898528443, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.46860671043396, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7495343685150146, + "num_tokens": 474757181.0, + "step": 18347 + }, + { + "epoch": 2.014935207555458, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.535146474838257, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7282273173332214, + "num_tokens": 474779652.0, + "step": 18348 + }, + { + "epoch": 2.0150450252580714, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5957367420196533, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7285259366035461, + "num_tokens": 474802679.0, + "step": 18349 + }, + { + "epoch": 2.015154842960685, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5289700031280518, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7296111583709717, + "num_tokens": 474826802.0, + "step": 18350 + }, + { + "epoch": 2.015264660663299, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3456132411956787, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7294614911079407, + "num_tokens": 474852591.0, + "step": 18351 + }, + { + "epoch": 2.0153744783659127, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1505889892578125, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.741736888885498, + "num_tokens": 474881174.0, + "step": 18352 + }, + { + "epoch": 2.0154842960685264, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0454251766204834, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7361765503883362, + "num_tokens": 474912081.0, + "step": 18353 + }, + { + "epoch": 2.0155941137711397, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5153679847717285, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7156988382339478, + "num_tokens": 474935198.0, + "step": 18354 + }, + { + "epoch": 2.0157039314737535, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4913954734802246, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7207516431808472, + "num_tokens": 474958871.0, + "step": 18355 + }, + { + "epoch": 2.0158137491763672, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.389716863632202, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7315816879272461, + "num_tokens": 474982605.0, + "step": 18356 + }, + { + "epoch": 2.015923566878981, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5053844451904297, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7242975234985352, + "num_tokens": 475006223.0, + "step": 18357 + }, + { + "epoch": 2.0160333845815948, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.215703248977661, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.730275571346283, + "num_tokens": 475034253.0, + "step": 18358 + }, + { + "epoch": 2.016143202284208, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1833879947662354, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7224345207214355, + "num_tokens": 475064327.0, + "step": 18359 + }, + { + "epoch": 2.016253019986822, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.340832471847534, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7380118370056152, + "num_tokens": 475088498.0, + "step": 18360 + }, + { + "epoch": 2.0163628376894356, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.265061616897583, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.73331618309021, + "num_tokens": 475113631.0, + "step": 18361 + }, + { + "epoch": 2.0164726553920493, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.1270523071289062, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7222851514816284, + "num_tokens": 475144330.0, + "step": 18362 + }, + { + "epoch": 2.0165824730946627, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.364370584487915, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7250845432281494, + "num_tokens": 475173602.0, + "step": 18363 + }, + { + "epoch": 2.0166922907972764, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.460798740386963, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7369059324264526, + "num_tokens": 475197199.0, + "step": 18364 + }, + { + "epoch": 2.01680210849989, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 1.99859619140625, + "learning_rate": 1e-06, + "loss": 0.7734, + "mean_token_accuracy": 0.7534838914871216, + "num_tokens": 475228261.0, + "step": 18365 + }, + { + "epoch": 2.016911926202504, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.269274950027466, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.720103919506073, + "num_tokens": 475255246.0, + "step": 18366 + }, + { + "epoch": 2.0170217439051177, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.502854347229004, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7564271688461304, + "num_tokens": 475278080.0, + "step": 18367 + }, + { + "epoch": 2.017131561607731, + "ewc_loss": 0.00015735626220703125, + "grad_norm": 2.7899951934814453, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7471631169319153, + "num_tokens": 475297692.0, + "step": 18368 + }, + { + "epoch": 2.0172413793103448, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.344353199005127, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7059352397918701, + "num_tokens": 475323342.0, + "step": 18369 + }, + { + "epoch": 2.0173511970129585, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3966846466064453, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.735001802444458, + "num_tokens": 475346994.0, + "step": 18370 + }, + { + "epoch": 2.0174610147155723, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2970683574676514, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7312588095664978, + "num_tokens": 475372315.0, + "step": 18371 + }, + { + "epoch": 2.017570832418186, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.352811574935913, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7459430694580078, + "num_tokens": 475396438.0, + "step": 18372 + }, + { + "epoch": 2.0176806501207993, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.358427047729492, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7174023389816284, + "num_tokens": 475422077.0, + "step": 18373 + }, + { + "epoch": 2.017790467823413, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.302462100982666, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7386808395385742, + "num_tokens": 475447617.0, + "step": 18374 + }, + { + "epoch": 2.017900285526027, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.624054431915283, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7404302358627319, + "num_tokens": 475471261.0, + "step": 18375 + }, + { + "epoch": 2.0180101032286406, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.608386278152466, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7386704087257385, + "num_tokens": 475492817.0, + "step": 18376 + }, + { + "epoch": 2.018119920931254, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.326643705368042, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7249189615249634, + "num_tokens": 475519515.0, + "step": 18377 + }, + { + "epoch": 2.0182297386338677, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.30844783782959, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7320852279663086, + "num_tokens": 475546017.0, + "step": 18378 + }, + { + "epoch": 2.0183395563364814, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2094287872314453, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7108293771743774, + "num_tokens": 475576814.0, + "step": 18379 + }, + { + "epoch": 2.018449374039095, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2917397022247314, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7183159589767456, + "num_tokens": 475604395.0, + "step": 18380 + }, + { + "epoch": 2.018559191741709, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.219902276992798, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7144553065299988, + "num_tokens": 475632406.0, + "step": 18381 + }, + { + "epoch": 2.0186690094443223, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.066077470779419, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7221836447715759, + "num_tokens": 475663063.0, + "step": 18382 + }, + { + "epoch": 2.018778827146936, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.659858465194702, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7404419183731079, + "num_tokens": 475683086.0, + "step": 18383 + }, + { + "epoch": 2.0188886448495498, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7658464908599854, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7438628673553467, + "num_tokens": 475701641.0, + "step": 18384 + }, + { + "epoch": 2.0189984625521635, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4590113162994385, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7200758457183838, + "num_tokens": 475725651.0, + "step": 18385 + }, + { + "epoch": 2.0191082802547773, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1058835983276367, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7308086156845093, + "num_tokens": 475755963.0, + "step": 18386 + }, + { + "epoch": 2.0192180979573906, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3751602172851562, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7156383991241455, + "num_tokens": 475782500.0, + "step": 18387 + }, + { + "epoch": 2.0193279156600044, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3174262046813965, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7248860597610474, + "num_tokens": 475810085.0, + "step": 18388 + }, + { + "epoch": 2.019437733362618, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5967564582824707, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7373895645141602, + "num_tokens": 475831769.0, + "step": 18389 + }, + { + "epoch": 2.019547551065232, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.215771436691284, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7199243307113647, + "num_tokens": 475860288.0, + "step": 18390 + }, + { + "epoch": 2.019657368767845, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.451167583465576, + "learning_rate": 1e-06, + "loss": 0.766, + "mean_token_accuracy": 0.7577810287475586, + "num_tokens": 475881887.0, + "step": 18391 + }, + { + "epoch": 2.019767186470459, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.566363573074341, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7179865837097168, + "num_tokens": 475907087.0, + "step": 18392 + }, + { + "epoch": 2.0198770041730727, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.558706521987915, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7265751361846924, + "num_tokens": 475930532.0, + "step": 18393 + }, + { + "epoch": 2.0199868218756865, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.300179958343506, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7447667121887207, + "num_tokens": 475956755.0, + "step": 18394 + }, + { + "epoch": 2.0200966395783, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.109483003616333, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7322391271591187, + "num_tokens": 475988953.0, + "step": 18395 + }, + { + "epoch": 2.0202064572809135, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6685638427734375, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7453574538230896, + "num_tokens": 476010678.0, + "step": 18396 + }, + { + "epoch": 2.0203162749835273, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2695701122283936, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7406073808670044, + "num_tokens": 476038599.0, + "step": 18397 + }, + { + "epoch": 2.020426092686141, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1885523796081543, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7187302112579346, + "num_tokens": 476067170.0, + "step": 18398 + }, + { + "epoch": 2.020535910388755, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.342506170272827, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7369002103805542, + "num_tokens": 476093353.0, + "step": 18399 + }, + { + "epoch": 2.020645728091368, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.203984260559082, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.727554202079773, + "num_tokens": 476120892.0, + "step": 18400 + }, + { + "epoch": 2.020755545793982, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2498230934143066, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7273646593093872, + "num_tokens": 476148830.0, + "step": 18401 + }, + { + "epoch": 2.0208653634965956, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2716126441955566, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7070451974868774, + "num_tokens": 476176269.0, + "step": 18402 + }, + { + "epoch": 2.0209751811992094, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3101236820220947, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7060842514038086, + "num_tokens": 476203739.0, + "step": 18403 + }, + { + "epoch": 2.021084998901823, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.557701349258423, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7390782833099365, + "num_tokens": 476227125.0, + "step": 18404 + }, + { + "epoch": 2.0211948166044365, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.360400438308716, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7245908379554749, + "num_tokens": 476253997.0, + "step": 18405 + }, + { + "epoch": 2.02130463430705, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3741369247436523, + "learning_rate": 1e-06, + "loss": 0.6952, + "mean_token_accuracy": 0.7773710489273071, + "num_tokens": 476276686.0, + "step": 18406 + }, + { + "epoch": 2.021414452009664, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.029743194580078, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7221508622169495, + "num_tokens": 476307945.0, + "step": 18407 + }, + { + "epoch": 2.0215242697122777, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 1.9036040306091309, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7167115211486816, + "num_tokens": 476344153.0, + "step": 18408 + }, + { + "epoch": 2.0216340874148915, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2269129753112793, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7490880489349365, + "num_tokens": 476372534.0, + "step": 18409 + }, + { + "epoch": 2.021743905117505, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2729616165161133, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.721001148223877, + "num_tokens": 476400143.0, + "step": 18410 + }, + { + "epoch": 2.0218537228201185, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.592797040939331, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7531027793884277, + "num_tokens": 476421887.0, + "step": 18411 + }, + { + "epoch": 2.0219635405227323, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2739274501800537, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7311272025108337, + "num_tokens": 476448408.0, + "step": 18412 + }, + { + "epoch": 2.022073358225346, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.862818956375122, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7441333532333374, + "num_tokens": 476467945.0, + "step": 18413 + }, + { + "epoch": 2.0221831759279594, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 1.9963533878326416, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7147353887557983, + "num_tokens": 476504064.0, + "step": 18414 + }, + { + "epoch": 2.022292993630573, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.412311553955078, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7405733466148376, + "num_tokens": 476528563.0, + "step": 18415 + }, + { + "epoch": 2.022402811333187, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5935845375061035, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.731839656829834, + "num_tokens": 476551278.0, + "step": 18416 + }, + { + "epoch": 2.0225126290358006, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.689675807952881, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7533373832702637, + "num_tokens": 476571059.0, + "step": 18417 + }, + { + "epoch": 2.0226224467384144, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1891093254089355, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7352490425109863, + "num_tokens": 476601332.0, + "step": 18418 + }, + { + "epoch": 2.0227322644410277, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.765381336212158, + "learning_rate": 1e-06, + "loss": 0.7567, + "mean_token_accuracy": 0.7573620080947876, + "num_tokens": 476620721.0, + "step": 18419 + }, + { + "epoch": 2.0228420821436415, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.419271469116211, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7337455749511719, + "num_tokens": 476647984.0, + "step": 18420 + }, + { + "epoch": 2.0229518998462552, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2967424392700195, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7369798421859741, + "num_tokens": 476675219.0, + "step": 18421 + }, + { + "epoch": 2.023061717548869, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.101754903793335, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7159234285354614, + "num_tokens": 476707059.0, + "step": 18422 + }, + { + "epoch": 2.0231715352514827, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.182703733444214, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.745945930480957, + "num_tokens": 476734933.0, + "step": 18423 + }, + { + "epoch": 2.023281352954096, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.441336154937744, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7458435297012329, + "num_tokens": 476758364.0, + "step": 18424 + }, + { + "epoch": 2.02339117065671, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1809232234954834, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7158863544464111, + "num_tokens": 476788896.0, + "step": 18425 + }, + { + "epoch": 2.0235009883593236, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4314820766448975, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7095524668693542, + "num_tokens": 476814130.0, + "step": 18426 + }, + { + "epoch": 2.0236108060619373, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.287646770477295, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7422479391098022, + "num_tokens": 476839446.0, + "step": 18427 + }, + { + "epoch": 2.0237206237645506, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.162280321121216, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7133455276489258, + "num_tokens": 476871147.0, + "step": 18428 + }, + { + "epoch": 2.0238304414671644, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2651116847991943, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7238346338272095, + "num_tokens": 476897450.0, + "step": 18429 + }, + { + "epoch": 2.023940259169778, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5397374629974365, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7194921970367432, + "num_tokens": 476921078.0, + "step": 18430 + }, + { + "epoch": 2.024050076872392, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4097700119018555, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7307865619659424, + "num_tokens": 476944763.0, + "step": 18431 + }, + { + "epoch": 2.0241598945750057, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7370412349700928, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.744570791721344, + "num_tokens": 476964532.0, + "step": 18432 + }, + { + "epoch": 2.024269712277619, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6601364612579346, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7429050207138062, + "num_tokens": 476984769.0, + "step": 18433 + }, + { + "epoch": 2.0243795299802327, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3601953983306885, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7209542393684387, + "num_tokens": 477010814.0, + "step": 18434 + }, + { + "epoch": 2.0244893476828465, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2955453395843506, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7315483093261719, + "num_tokens": 477037902.0, + "step": 18435 + }, + { + "epoch": 2.0245991653854603, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3912768363952637, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7353979349136353, + "num_tokens": 477062536.0, + "step": 18436 + }, + { + "epoch": 2.024708983088074, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.397148370742798, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.738242506980896, + "num_tokens": 477088327.0, + "step": 18437 + }, + { + "epoch": 2.0248188007906873, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4331047534942627, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7561089396476746, + "num_tokens": 477111255.0, + "step": 18438 + }, + { + "epoch": 2.024928618493301, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.105726957321167, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7362848520278931, + "num_tokens": 477140742.0, + "step": 18439 + }, + { + "epoch": 2.025038436195915, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.337597131729126, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7257257103919983, + "num_tokens": 477167026.0, + "step": 18440 + }, + { + "epoch": 2.0251482538985286, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.562547445297241, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7292741537094116, + "num_tokens": 477189984.0, + "step": 18441 + }, + { + "epoch": 2.025258071601142, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.369541645050049, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6957676410675049, + "num_tokens": 477216833.0, + "step": 18442 + }, + { + "epoch": 2.0253678893037557, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6130456924438477, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7261166572570801, + "num_tokens": 477239306.0, + "step": 18443 + }, + { + "epoch": 2.0254777070063694, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6044223308563232, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7326899766921997, + "num_tokens": 477261024.0, + "step": 18444 + }, + { + "epoch": 2.025587524708983, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.296759605407715, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7370604276657104, + "num_tokens": 477287270.0, + "step": 18445 + }, + { + "epoch": 2.025697342411597, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2952239513397217, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.733819842338562, + "num_tokens": 477315531.0, + "step": 18446 + }, + { + "epoch": 2.0258071601142102, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.327024459838867, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7368770241737366, + "num_tokens": 477340295.0, + "step": 18447 + }, + { + "epoch": 2.025916977816824, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5232937335968018, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7434028387069702, + "num_tokens": 477362513.0, + "step": 18448 + }, + { + "epoch": 2.0260267955194378, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1915597915649414, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7061344981193542, + "num_tokens": 477390622.0, + "step": 18449 + }, + { + "epoch": 2.0261366132220515, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.33311128616333, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.734045147895813, + "num_tokens": 477415906.0, + "step": 18450 + }, + { + "epoch": 2.0262464309246653, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2773327827453613, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7322883605957031, + "num_tokens": 477442984.0, + "step": 18451 + }, + { + "epoch": 2.0263562486272786, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.246882200241089, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7423413991928101, + "num_tokens": 477467160.0, + "step": 18452 + }, + { + "epoch": 2.0264660663298923, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1377274990081787, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7383193969726562, + "num_tokens": 477494952.0, + "step": 18453 + }, + { + "epoch": 2.026575884032506, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.256200075149536, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7220851182937622, + "num_tokens": 477523578.0, + "step": 18454 + }, + { + "epoch": 2.02668570173512, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.354300022125244, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7391365766525269, + "num_tokens": 477547502.0, + "step": 18455 + }, + { + "epoch": 2.026795519437733, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7024548053741455, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7189083099365234, + "num_tokens": 477568762.0, + "step": 18456 + }, + { + "epoch": 2.026905337140347, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.556983232498169, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7432208061218262, + "num_tokens": 477591095.0, + "step": 18457 + }, + { + "epoch": 2.0270151548429607, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3820226192474365, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7350040674209595, + "num_tokens": 477615159.0, + "step": 18458 + }, + { + "epoch": 2.0271249725455744, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1451516151428223, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7012861371040344, + "num_tokens": 477646241.0, + "step": 18459 + }, + { + "epoch": 2.027234790248188, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0243024826049805, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7004779577255249, + "num_tokens": 477678701.0, + "step": 18460 + }, + { + "epoch": 2.0273446079508015, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1925556659698486, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7083696722984314, + "num_tokens": 477708276.0, + "step": 18461 + }, + { + "epoch": 2.0274544256534153, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.212549924850464, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7290542125701904, + "num_tokens": 477734701.0, + "step": 18462 + }, + { + "epoch": 2.027564243356029, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2495689392089844, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.725008487701416, + "num_tokens": 477761108.0, + "step": 18463 + }, + { + "epoch": 2.027674061058643, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.251525402069092, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7461583614349365, + "num_tokens": 477787164.0, + "step": 18464 + }, + { + "epoch": 2.027783878761256, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.50211238861084, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7279977798461914, + "num_tokens": 477811238.0, + "step": 18465 + }, + { + "epoch": 2.02789369646387, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.154986619949341, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7276121377944946, + "num_tokens": 477840378.0, + "step": 18466 + }, + { + "epoch": 2.0280035141664836, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.36023211479187, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.718089759349823, + "num_tokens": 477866312.0, + "step": 18467 + }, + { + "epoch": 2.0281133318690974, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4333364963531494, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7474194765090942, + "num_tokens": 477889462.0, + "step": 18468 + }, + { + "epoch": 2.028223149571711, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3744876384735107, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.730745255947113, + "num_tokens": 477912523.0, + "step": 18469 + }, + { + "epoch": 2.0283329672743244, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3645005226135254, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7448350191116333, + "num_tokens": 477938584.0, + "step": 18470 + }, + { + "epoch": 2.028442784976938, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.149935245513916, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7236775755882263, + "num_tokens": 477969511.0, + "step": 18471 + }, + { + "epoch": 2.028552602679552, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.373858690261841, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.712457001209259, + "num_tokens": 477996399.0, + "step": 18472 + }, + { + "epoch": 2.0286624203821657, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.527740478515625, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7387877702713013, + "num_tokens": 478019229.0, + "step": 18473 + }, + { + "epoch": 2.0287722380847795, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.563384771347046, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7457289099693298, + "num_tokens": 478040209.0, + "step": 18474 + }, + { + "epoch": 2.028882055787393, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3948450088500977, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7480875849723816, + "num_tokens": 478062881.0, + "step": 18475 + }, + { + "epoch": 2.0289918734900065, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.472705125808716, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7491756677627563, + "num_tokens": 478085248.0, + "step": 18476 + }, + { + "epoch": 2.0291016911926203, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.277068614959717, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7280999422073364, + "num_tokens": 478111876.0, + "step": 18477 + }, + { + "epoch": 2.029211508895234, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.217222213745117, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7323735952377319, + "num_tokens": 478140060.0, + "step": 18478 + }, + { + "epoch": 2.0293213265978474, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.021787643432617, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7111619710922241, + "num_tokens": 478173027.0, + "step": 18479 + }, + { + "epoch": 2.029431144300461, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4415736198425293, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7332726716995239, + "num_tokens": 478197596.0, + "step": 18480 + }, + { + "epoch": 2.029540962003075, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2248218059539795, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.74295973777771, + "num_tokens": 478227789.0, + "step": 18481 + }, + { + "epoch": 2.0296507797056886, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.122856616973877, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7213119864463806, + "num_tokens": 478259865.0, + "step": 18482 + }, + { + "epoch": 2.0297605974083024, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.70711350440979, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7407747507095337, + "num_tokens": 478280288.0, + "step": 18483 + }, + { + "epoch": 2.0298704151109157, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6376872062683105, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.731224000453949, + "num_tokens": 478301159.0, + "step": 18484 + }, + { + "epoch": 2.0299802328135295, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.424565553665161, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7367141842842102, + "num_tokens": 478326072.0, + "step": 18485 + }, + { + "epoch": 2.030090050516143, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0925962924957275, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.693634569644928, + "num_tokens": 478362978.0, + "step": 18486 + }, + { + "epoch": 2.030199868218757, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.101275682449341, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7328590750694275, + "num_tokens": 478395259.0, + "step": 18487 + }, + { + "epoch": 2.0303096859213707, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2314047813415527, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7218650579452515, + "num_tokens": 478424449.0, + "step": 18488 + }, + { + "epoch": 2.030419503623984, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1157236099243164, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.716217577457428, + "num_tokens": 478454349.0, + "step": 18489 + }, + { + "epoch": 2.030529321326598, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6560654640197754, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7218726873397827, + "num_tokens": 478476519.0, + "step": 18490 + }, + { + "epoch": 2.0306391390292116, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4402358531951904, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.727121889591217, + "num_tokens": 478501584.0, + "step": 18491 + }, + { + "epoch": 2.0307489567318253, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.381195306777954, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.726190447807312, + "num_tokens": 478527028.0, + "step": 18492 + }, + { + "epoch": 2.0308587744344386, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1599650382995605, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7285834550857544, + "num_tokens": 478556497.0, + "step": 18493 + }, + { + "epoch": 2.0309685921370524, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3586630821228027, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7367773056030273, + "num_tokens": 478581600.0, + "step": 18494 + }, + { + "epoch": 2.031078409839666, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3738999366760254, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.718461811542511, + "num_tokens": 478608224.0, + "step": 18495 + }, + { + "epoch": 2.03118822754228, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6006622314453125, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7430093288421631, + "num_tokens": 478630025.0, + "step": 18496 + }, + { + "epoch": 2.0312980452448937, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.278491258621216, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7279927134513855, + "num_tokens": 478655944.0, + "step": 18497 + }, + { + "epoch": 2.031407862947507, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.311750888824463, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7184678316116333, + "num_tokens": 478682722.0, + "step": 18498 + }, + { + "epoch": 2.0315176806501207, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.219022274017334, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7229613661766052, + "num_tokens": 478712650.0, + "step": 18499 + }, + { + "epoch": 2.0316274983527345, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 1.8166813850402832, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7162981629371643, + "num_tokens": 478750751.0, + "step": 18500 + }, + { + "epoch": 2.0317373160553482, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2052648067474365, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7034794688224792, + "num_tokens": 478782213.0, + "step": 18501 + }, + { + "epoch": 2.031847133757962, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.560626745223999, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7242358922958374, + "num_tokens": 478806595.0, + "step": 18502 + }, + { + "epoch": 2.0319569514605753, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 1.9829550981521606, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7078122496604919, + "num_tokens": 478842045.0, + "step": 18503 + }, + { + "epoch": 2.032066769163189, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.267063617706299, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7242639064788818, + "num_tokens": 478870674.0, + "step": 18504 + }, + { + "epoch": 2.032176586865803, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6958580017089844, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7352895140647888, + "num_tokens": 478891731.0, + "step": 18505 + }, + { + "epoch": 2.0322864045684166, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1468169689178467, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7248921990394592, + "num_tokens": 478920241.0, + "step": 18506 + }, + { + "epoch": 2.03239622227103, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0870485305786133, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7095286250114441, + "num_tokens": 478951766.0, + "step": 18507 + }, + { + "epoch": 2.0325060399736437, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0939276218414307, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7203316688537598, + "num_tokens": 478984028.0, + "step": 18508 + }, + { + "epoch": 2.0326158576762574, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.194411039352417, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7179945111274719, + "num_tokens": 479013463.0, + "step": 18509 + }, + { + "epoch": 2.032725675378871, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3715522289276123, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7089841961860657, + "num_tokens": 479038882.0, + "step": 18510 + }, + { + "epoch": 2.032835493081485, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4665799140930176, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.718124270439148, + "num_tokens": 479063048.0, + "step": 18511 + }, + { + "epoch": 2.0329453107840982, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.222170352935791, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7157069444656372, + "num_tokens": 479092793.0, + "step": 18512 + }, + { + "epoch": 2.033055128486712, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.536017417907715, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7390145063400269, + "num_tokens": 479114094.0, + "step": 18513 + }, + { + "epoch": 2.0331649461893258, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2964119911193848, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7439953684806824, + "num_tokens": 479139288.0, + "step": 18514 + }, + { + "epoch": 2.0332747638919395, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.461513042449951, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7357238531112671, + "num_tokens": 479161789.0, + "step": 18515 + }, + { + "epoch": 2.033384581594553, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.314265727996826, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.720453143119812, + "num_tokens": 479187593.0, + "step": 18516 + }, + { + "epoch": 2.0334943992971666, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2575325965881348, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7334497570991516, + "num_tokens": 479215416.0, + "step": 18517 + }, + { + "epoch": 2.0336042169997803, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.242541790008545, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7271739840507507, + "num_tokens": 479242278.0, + "step": 18518 + }, + { + "epoch": 2.033714034702394, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1981618404388428, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7184385061264038, + "num_tokens": 479270308.0, + "step": 18519 + }, + { + "epoch": 2.033823852405008, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2794711589813232, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7419654130935669, + "num_tokens": 479297337.0, + "step": 18520 + }, + { + "epoch": 2.033933670107621, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.176159620285034, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7189940214157104, + "num_tokens": 479326319.0, + "step": 18521 + }, + { + "epoch": 2.034043487810235, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.281928300857544, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.734067976474762, + "num_tokens": 479353862.0, + "step": 18522 + }, + { + "epoch": 2.0341533055128487, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.247859001159668, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7302173376083374, + "num_tokens": 479380234.0, + "step": 18523 + }, + { + "epoch": 2.0342631232154624, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.616013526916504, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.740428626537323, + "num_tokens": 479400222.0, + "step": 18524 + }, + { + "epoch": 2.034372940918076, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.318380355834961, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7324154376983643, + "num_tokens": 479425746.0, + "step": 18525 + }, + { + "epoch": 2.0344827586206895, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4800267219543457, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7409144639968872, + "num_tokens": 479449819.0, + "step": 18526 + }, + { + "epoch": 2.0345925763233033, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3683738708496094, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7242907285690308, + "num_tokens": 479475323.0, + "step": 18527 + }, + { + "epoch": 2.034702394025917, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1694347858428955, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7102915048599243, + "num_tokens": 479505738.0, + "step": 18528 + }, + { + "epoch": 2.0348122117285308, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.585578203201294, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7391015291213989, + "num_tokens": 479527435.0, + "step": 18529 + }, + { + "epoch": 2.034922029431144, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.855306625366211, + "learning_rate": 1e-06, + "loss": 0.7286, + "mean_token_accuracy": 0.7654460668563843, + "num_tokens": 479545792.0, + "step": 18530 + }, + { + "epoch": 2.035031847133758, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2877116203308105, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7373706698417664, + "num_tokens": 479570974.0, + "step": 18531 + }, + { + "epoch": 2.0351416648363716, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7358505725860596, + "learning_rate": 1e-06, + "loss": 0.7367, + "mean_token_accuracy": 0.7620395421981812, + "num_tokens": 479588853.0, + "step": 18532 + }, + { + "epoch": 2.0352514825389854, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7523112297058105, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7413326501846313, + "num_tokens": 479609539.0, + "step": 18533 + }, + { + "epoch": 2.035361300241599, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.304333448410034, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7389490604400635, + "num_tokens": 479635985.0, + "step": 18534 + }, + { + "epoch": 2.0354711179442124, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.261563777923584, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7252737283706665, + "num_tokens": 479662486.0, + "step": 18535 + }, + { + "epoch": 2.035580935646826, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.503695011138916, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7266725897789001, + "num_tokens": 479686596.0, + "step": 18536 + }, + { + "epoch": 2.03569075334944, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0012221336364746, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7394227981567383, + "num_tokens": 479718032.0, + "step": 18537 + }, + { + "epoch": 2.0358005710520537, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.514518976211548, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7310924530029297, + "num_tokens": 479743117.0, + "step": 18538 + }, + { + "epoch": 2.0359103887546675, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.165712594985962, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7530839443206787, + "num_tokens": 479772035.0, + "step": 18539 + }, + { + "epoch": 2.0360202064572808, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.426847457885742, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7324485778808594, + "num_tokens": 479797667.0, + "step": 18540 + }, + { + "epoch": 2.0361300241598945, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.128451108932495, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7031830549240112, + "num_tokens": 479825969.0, + "step": 18541 + }, + { + "epoch": 2.0362398418625083, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.411280393600464, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7325136661529541, + "num_tokens": 479851139.0, + "step": 18542 + }, + { + "epoch": 2.036349659565122, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3479509353637695, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7227245569229126, + "num_tokens": 479876346.0, + "step": 18543 + }, + { + "epoch": 2.0364594772677354, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.8835113048553467, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.731773853302002, + "num_tokens": 479897002.0, + "step": 18544 + }, + { + "epoch": 2.036569294970349, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.522458791732788, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7341938018798828, + "num_tokens": 479919893.0, + "step": 18545 + }, + { + "epoch": 2.036679112672963, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.1812756061553955, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7326103448867798, + "num_tokens": 479949082.0, + "step": 18546 + }, + { + "epoch": 2.0367889303755766, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.5946075916290283, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7276223301887512, + "num_tokens": 479972061.0, + "step": 18547 + }, + { + "epoch": 2.0368987480781904, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.193812131881714, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7582802176475525, + "num_tokens": 480000751.0, + "step": 18548 + }, + { + "epoch": 2.0370085657808037, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.21379017829895, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7291079759597778, + "num_tokens": 480029222.0, + "step": 18549 + }, + { + "epoch": 2.0371183834834174, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4830005168914795, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7496546506881714, + "num_tokens": 480053003.0, + "step": 18550 + }, + { + "epoch": 2.037228201186031, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.191807746887207, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7369896769523621, + "num_tokens": 480080069.0, + "step": 18551 + }, + { + "epoch": 2.037338018888645, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.501695156097412, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7319225072860718, + "num_tokens": 480103242.0, + "step": 18552 + }, + { + "epoch": 2.0374478365912587, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4299509525299072, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.72873854637146, + "num_tokens": 480128365.0, + "step": 18553 + }, + { + "epoch": 2.037557654293872, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.601182222366333, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7503626942634583, + "num_tokens": 480150859.0, + "step": 18554 + }, + { + "epoch": 2.037667471996486, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3931643962860107, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7342854738235474, + "num_tokens": 480176966.0, + "step": 18555 + }, + { + "epoch": 2.0377772896990995, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.7353415489196777, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7452363967895508, + "num_tokens": 480196250.0, + "step": 18556 + }, + { + "epoch": 2.0378871074017133, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.413224458694458, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7509307265281677, + "num_tokens": 480218428.0, + "step": 18557 + }, + { + "epoch": 2.0379969251043266, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1653332710266113, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7442481517791748, + "num_tokens": 480247440.0, + "step": 18558 + }, + { + "epoch": 2.0381067428069404, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.512408971786499, + "learning_rate": 1e-06, + "loss": 0.7699, + "mean_token_accuracy": 0.754002571105957, + "num_tokens": 480270897.0, + "step": 18559 + }, + { + "epoch": 2.038216560509554, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.188300609588623, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7440808415412903, + "num_tokens": 480299102.0, + "step": 18560 + }, + { + "epoch": 2.038326378212168, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6704258918762207, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7197879552841187, + "num_tokens": 480322035.0, + "step": 18561 + }, + { + "epoch": 2.0384361959147816, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3471760749816895, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7462754249572754, + "num_tokens": 480346328.0, + "step": 18562 + }, + { + "epoch": 2.038546013617395, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.525242567062378, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7379041910171509, + "num_tokens": 480370514.0, + "step": 18563 + }, + { + "epoch": 2.0386558313200087, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.472224473953247, + "learning_rate": 1e-06, + "loss": 0.7837, + "mean_token_accuracy": 0.7480214834213257, + "num_tokens": 480391909.0, + "step": 18564 + }, + { + "epoch": 2.0387656490226225, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.511730194091797, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7393606305122375, + "num_tokens": 480415314.0, + "step": 18565 + }, + { + "epoch": 2.0388754667252362, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1726248264312744, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7197663187980652, + "num_tokens": 480444693.0, + "step": 18566 + }, + { + "epoch": 2.03898528442785, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.274622917175293, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.723232626914978, + "num_tokens": 480472164.0, + "step": 18567 + }, + { + "epoch": 2.0390951021304633, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 1.998314619064331, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.756822943687439, + "num_tokens": 480504905.0, + "step": 18568 + }, + { + "epoch": 2.039204919833077, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4655470848083496, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7259770035743713, + "num_tokens": 480530076.0, + "step": 18569 + }, + { + "epoch": 2.039314737535691, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.041067600250244, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7224735021591187, + "num_tokens": 480560966.0, + "step": 18570 + }, + { + "epoch": 2.0394245552383046, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2788660526275635, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7107386589050293, + "num_tokens": 480586988.0, + "step": 18571 + }, + { + "epoch": 2.039534372940918, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.283344030380249, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7064371109008789, + "num_tokens": 480615259.0, + "step": 18572 + }, + { + "epoch": 2.0396441906435316, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.052530527114868, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7173818349838257, + "num_tokens": 480646526.0, + "step": 18573 + }, + { + "epoch": 2.0397540083461454, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.790867567062378, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7267454266548157, + "num_tokens": 480667027.0, + "step": 18574 + }, + { + "epoch": 2.039863826048759, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3236067295074463, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7359812259674072, + "num_tokens": 480690476.0, + "step": 18575 + }, + { + "epoch": 2.039973643751373, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2740981578826904, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7123385667800903, + "num_tokens": 480719978.0, + "step": 18576 + }, + { + "epoch": 2.0400834614539862, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.0821335315704346, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7392297983169556, + "num_tokens": 480749163.0, + "step": 18577 + }, + { + "epoch": 2.0401932791566, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6368284225463867, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7518823146820068, + "num_tokens": 480769114.0, + "step": 18578 + }, + { + "epoch": 2.0403030968592137, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.382023334503174, + "learning_rate": 1e-06, + "loss": 0.7348, + "mean_token_accuracy": 0.764626681804657, + "num_tokens": 480792241.0, + "step": 18579 + }, + { + "epoch": 2.0404129145618275, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4125843048095703, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.714779794216156, + "num_tokens": 480817932.0, + "step": 18580 + }, + { + "epoch": 2.0405227322644413, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.6360881328582764, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7429221272468567, + "num_tokens": 480839867.0, + "step": 18581 + }, + { + "epoch": 2.0406325499670546, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.3870534896850586, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.745704174041748, + "num_tokens": 480863797.0, + "step": 18582 + }, + { + "epoch": 2.0407423676696683, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.148866653442383, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7404038906097412, + "num_tokens": 480894772.0, + "step": 18583 + }, + { + "epoch": 2.040852185372282, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.526393413543701, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7360454797744751, + "num_tokens": 480917888.0, + "step": 18584 + }, + { + "epoch": 2.040962003074896, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5840134620666504, + "learning_rate": 1e-06, + "loss": 0.7444, + "mean_token_accuracy": 0.7565253376960754, + "num_tokens": 480937887.0, + "step": 18585 + }, + { + "epoch": 2.041071820777509, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.584174156188965, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7476032376289368, + "num_tokens": 480959462.0, + "step": 18586 + }, + { + "epoch": 2.041181638480123, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.2343497276306152, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7177901268005371, + "num_tokens": 480987958.0, + "step": 18587 + }, + { + "epoch": 2.0412914561827367, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4333815574645996, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7213767766952515, + "num_tokens": 481011290.0, + "step": 18588 + }, + { + "epoch": 2.0414012738853504, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.41728138923645, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7410932183265686, + "num_tokens": 481036578.0, + "step": 18589 + }, + { + "epoch": 2.041511091587964, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.221247911453247, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7159111499786377, + "num_tokens": 481065866.0, + "step": 18590 + }, + { + "epoch": 2.0416209092905775, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.4315881729125977, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7023826837539673, + "num_tokens": 481092331.0, + "step": 18591 + }, + { + "epoch": 2.0417307269931912, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.5170071125030518, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7241643667221069, + "num_tokens": 481116641.0, + "step": 18592 + }, + { + "epoch": 2.041840544695805, + "ewc_loss": 0.0001583099365234375, + "grad_norm": 2.1126720905303955, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7185631990432739, + "num_tokens": 481147880.0, + "step": 18593 + }, + { + "epoch": 2.0419503623984188, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.1582202911376953, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7291273474693298, + "num_tokens": 481177967.0, + "step": 18594 + }, + { + "epoch": 2.042060180101032, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.607461929321289, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7350685000419617, + "num_tokens": 481199113.0, + "step": 18595 + }, + { + "epoch": 2.042169997803646, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2522597312927246, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.734173059463501, + "num_tokens": 481224513.0, + "step": 18596 + }, + { + "epoch": 2.0422798155062596, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.5661683082580566, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7330073118209839, + "num_tokens": 481247538.0, + "step": 18597 + }, + { + "epoch": 2.0423896332088733, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.7385621070861816, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.741939902305603, + "num_tokens": 481268354.0, + "step": 18598 + }, + { + "epoch": 2.042499450911487, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4064605236053467, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7148103713989258, + "num_tokens": 481292886.0, + "step": 18599 + }, + { + "epoch": 2.0426092686141004, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.530759811401367, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.732663631439209, + "num_tokens": 481318875.0, + "step": 18600 + }, + { + "epoch": 2.042719086316714, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.0113329887390137, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7156092524528503, + "num_tokens": 481350672.0, + "step": 18601 + }, + { + "epoch": 2.042828904019328, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.760129451751709, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7109133005142212, + "num_tokens": 481373624.0, + "step": 18602 + }, + { + "epoch": 2.0429387217219417, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.7676072120666504, + "learning_rate": 1e-06, + "loss": 0.7165, + "mean_token_accuracy": 0.7689472436904907, + "num_tokens": 481392991.0, + "step": 18603 + }, + { + "epoch": 2.0430485394245554, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.7050135135650635, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7338836789131165, + "num_tokens": 481413218.0, + "step": 18604 + }, + { + "epoch": 2.0431583571271688, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.188136577606201, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.712360143661499, + "num_tokens": 481441656.0, + "step": 18605 + }, + { + "epoch": 2.0432681748297825, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.856739044189453, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7435411810874939, + "num_tokens": 481459766.0, + "step": 18606 + }, + { + "epoch": 2.0433779925323963, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.152024507522583, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7374054193496704, + "num_tokens": 481486822.0, + "step": 18607 + }, + { + "epoch": 2.04348781023501, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.5591063499450684, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7469300627708435, + "num_tokens": 481508810.0, + "step": 18608 + }, + { + "epoch": 2.0435976279376233, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4273085594177246, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7196100950241089, + "num_tokens": 481535761.0, + "step": 18609 + }, + { + "epoch": 2.043707445640237, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 1.9928054809570312, + "learning_rate": 1e-06, + "loss": 0.7387, + "mean_token_accuracy": 0.7592048048973083, + "num_tokens": 481567869.0, + "step": 18610 + }, + { + "epoch": 2.043817263342851, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.221991539001465, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7132641077041626, + "num_tokens": 481596756.0, + "step": 18611 + }, + { + "epoch": 2.0439270810454646, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2610862255096436, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6980509757995605, + "num_tokens": 481625764.0, + "step": 18612 + }, + { + "epoch": 2.0440368987480784, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2050979137420654, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7198060750961304, + "num_tokens": 481654419.0, + "step": 18613 + }, + { + "epoch": 2.0441467164506917, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.63983154296875, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7427659034729004, + "num_tokens": 481673690.0, + "step": 18614 + }, + { + "epoch": 2.0442565341533054, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2718067169189453, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.734312117099762, + "num_tokens": 481700537.0, + "step": 18615 + }, + { + "epoch": 2.044366351855919, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.533867120742798, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7482452988624573, + "num_tokens": 481721716.0, + "step": 18616 + }, + { + "epoch": 2.044476169558533, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.5298335552215576, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7351589202880859, + "num_tokens": 481744694.0, + "step": 18617 + }, + { + "epoch": 2.0445859872611467, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.218784809112549, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.732668399810791, + "num_tokens": 481773619.0, + "step": 18618 + }, + { + "epoch": 2.04469580496376, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.524308443069458, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7437364459037781, + "num_tokens": 481796860.0, + "step": 18619 + }, + { + "epoch": 2.044805622666374, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2070729732513428, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7308465242385864, + "num_tokens": 481824971.0, + "step": 18620 + }, + { + "epoch": 2.0449154403689875, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4001309871673584, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7347671985626221, + "num_tokens": 481851380.0, + "step": 18621 + }, + { + "epoch": 2.0450252580716013, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.496922016143799, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7457613945007324, + "num_tokens": 481875181.0, + "step": 18622 + }, + { + "epoch": 2.0451350757742146, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.484381675720215, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7110677361488342, + "num_tokens": 481901916.0, + "step": 18623 + }, + { + "epoch": 2.0452448934768284, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.3515572547912598, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7198532819747925, + "num_tokens": 481929392.0, + "step": 18624 + }, + { + "epoch": 2.045354711179442, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4916582107543945, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7308304309844971, + "num_tokens": 481955024.0, + "step": 18625 + }, + { + "epoch": 2.045464528882056, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2790753841400146, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7359117269515991, + "num_tokens": 481982868.0, + "step": 18626 + }, + { + "epoch": 2.0455743465846696, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2098228931427, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.715367317199707, + "num_tokens": 482010241.0, + "step": 18627 + }, + { + "epoch": 2.045684164287283, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.191500425338745, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7112890481948853, + "num_tokens": 482038554.0, + "step": 18628 + }, + { + "epoch": 2.0457939819898967, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.0600345134735107, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7177287936210632, + "num_tokens": 482071571.0, + "step": 18629 + }, + { + "epoch": 2.0459037996925105, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4600048065185547, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7116052508354187, + "num_tokens": 482094918.0, + "step": 18630 + }, + { + "epoch": 2.046013617395124, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.9016106128692627, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7219299077987671, + "num_tokens": 482113080.0, + "step": 18631 + }, + { + "epoch": 2.046123435097738, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2756588459014893, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7401492595672607, + "num_tokens": 482139261.0, + "step": 18632 + }, + { + "epoch": 2.0462332528003513, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.759571075439453, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7408894300460815, + "num_tokens": 482159933.0, + "step": 18633 + }, + { + "epoch": 2.046343070502965, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.4985063076019287, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7250473499298096, + "num_tokens": 482184511.0, + "step": 18634 + }, + { + "epoch": 2.046452888205579, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.1667122840881348, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7234829664230347, + "num_tokens": 482214353.0, + "step": 18635 + }, + { + "epoch": 2.0465627059081926, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.346299409866333, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7569345235824585, + "num_tokens": 482238019.0, + "step": 18636 + }, + { + "epoch": 2.046672523610806, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.420626401901245, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7305201292037964, + "num_tokens": 482263259.0, + "step": 18637 + }, + { + "epoch": 2.0467823413134196, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.0883684158325195, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7312397360801697, + "num_tokens": 482297651.0, + "step": 18638 + }, + { + "epoch": 2.0468921590160334, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.2837131023406982, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6966623067855835, + "num_tokens": 482326127.0, + "step": 18639 + }, + { + "epoch": 2.047001976718647, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.139453411102295, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7301691174507141, + "num_tokens": 482355648.0, + "step": 18640 + }, + { + "epoch": 2.047111794421261, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 1.8694721460342407, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7378929257392883, + "num_tokens": 482393914.0, + "step": 18641 + }, + { + "epoch": 2.047221612123874, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.313096523284912, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7185556292533875, + "num_tokens": 482419669.0, + "step": 18642 + }, + { + "epoch": 2.047331429826488, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.5701584815979004, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7307666540145874, + "num_tokens": 482441513.0, + "step": 18643 + }, + { + "epoch": 2.0474412475291017, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.432908058166504, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7169507741928101, + "num_tokens": 482467869.0, + "step": 18644 + }, + { + "epoch": 2.0475510652317155, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.236630439758301, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7386566400527954, + "num_tokens": 482496206.0, + "step": 18645 + }, + { + "epoch": 2.047660882934329, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.3885350227355957, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7164770364761353, + "num_tokens": 482521093.0, + "step": 18646 + }, + { + "epoch": 2.0477707006369426, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.513890504837036, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7339982986450195, + "num_tokens": 482545337.0, + "step": 18647 + }, + { + "epoch": 2.0478805183395563, + "ewc_loss": 0.00015926361083984375, + "grad_norm": 2.231158971786499, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7580198645591736, + "num_tokens": 482572245.0, + "step": 18648 + }, + { + "epoch": 2.04799033604217, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1689693927764893, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7418606877326965, + "num_tokens": 482600882.0, + "step": 18649 + }, + { + "epoch": 2.048100153744784, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.387709617614746, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7434079647064209, + "num_tokens": 482626059.0, + "step": 18650 + }, + { + "epoch": 2.048209971447397, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2072150707244873, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.707950234413147, + "num_tokens": 482654852.0, + "step": 18651 + }, + { + "epoch": 2.048319789150011, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.3382835388183594, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7487772107124329, + "num_tokens": 482681155.0, + "step": 18652 + }, + { + "epoch": 2.0484296068526247, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.763700485229492, + "learning_rate": 1e-06, + "loss": 0.7879, + "mean_token_accuracy": 0.7500758171081543, + "num_tokens": 482699738.0, + "step": 18653 + }, + { + "epoch": 2.0485394245552384, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1876285076141357, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7199009656906128, + "num_tokens": 482729568.0, + "step": 18654 + }, + { + "epoch": 2.048649242257852, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4611051082611084, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.745081901550293, + "num_tokens": 482752127.0, + "step": 18655 + }, + { + "epoch": 2.0487590599604655, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.485250949859619, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7359768152236938, + "num_tokens": 482775429.0, + "step": 18656 + }, + { + "epoch": 2.0488688776630792, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2427942752838135, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7162038087844849, + "num_tokens": 482803518.0, + "step": 18657 + }, + { + "epoch": 2.048978695365693, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.334909200668335, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7239159941673279, + "num_tokens": 482829798.0, + "step": 18658 + }, + { + "epoch": 2.0490885130683067, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2835853099823, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7351006269454956, + "num_tokens": 482858683.0, + "step": 18659 + }, + { + "epoch": 2.04919833077092, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2305009365081787, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7244846820831299, + "num_tokens": 482884621.0, + "step": 18660 + }, + { + "epoch": 2.049308148473534, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.3205394744873047, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.734018087387085, + "num_tokens": 482910212.0, + "step": 18661 + }, + { + "epoch": 2.0494179661761476, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1295969486236572, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7463642358779907, + "num_tokens": 482938617.0, + "step": 18662 + }, + { + "epoch": 2.0495277838787613, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1247918605804443, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7321031093597412, + "num_tokens": 482968113.0, + "step": 18663 + }, + { + "epoch": 2.049637601581375, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2245404720306396, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7094023823738098, + "num_tokens": 482998324.0, + "step": 18664 + }, + { + "epoch": 2.0497474192839884, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.257080554962158, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7166097164154053, + "num_tokens": 483025136.0, + "step": 18665 + }, + { + "epoch": 2.049857236986602, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.459362030029297, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7382954359054565, + "num_tokens": 483048026.0, + "step": 18666 + }, + { + "epoch": 2.049967054689216, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.6383566856384277, + "learning_rate": 1e-06, + "loss": 0.7711, + "mean_token_accuracy": 0.7563871741294861, + "num_tokens": 483069609.0, + "step": 18667 + }, + { + "epoch": 2.0500768723918297, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.443190813064575, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7385008335113525, + "num_tokens": 483092995.0, + "step": 18668 + }, + { + "epoch": 2.0501866900944434, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.223863124847412, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7258954048156738, + "num_tokens": 483119666.0, + "step": 18669 + }, + { + "epoch": 2.0502965077970567, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1157851219177246, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7600087523460388, + "num_tokens": 483150595.0, + "step": 18670 + }, + { + "epoch": 2.0504063254996705, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.43843936920166, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7423503398895264, + "num_tokens": 483173866.0, + "step": 18671 + }, + { + "epoch": 2.0505161432022843, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.6601758003234863, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7227846384048462, + "num_tokens": 483195512.0, + "step": 18672 + }, + { + "epoch": 2.050625960904898, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.193556785583496, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7481181025505066, + "num_tokens": 483224278.0, + "step": 18673 + }, + { + "epoch": 2.0507357786075113, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.250333309173584, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7360198497772217, + "num_tokens": 483251606.0, + "step": 18674 + }, + { + "epoch": 2.050845596310125, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4049246311187744, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7235165238380432, + "num_tokens": 483276901.0, + "step": 18675 + }, + { + "epoch": 2.050955414012739, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2157692909240723, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.742387056350708, + "num_tokens": 483304570.0, + "step": 18676 + }, + { + "epoch": 2.0510652317153526, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2382402420043945, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.7563009262084961, + "num_tokens": 483330328.0, + "step": 18677 + }, + { + "epoch": 2.0511750494179664, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.411081075668335, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7042745351791382, + "num_tokens": 483356872.0, + "step": 18678 + }, + { + "epoch": 2.0512848671205797, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.408217191696167, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7213680148124695, + "num_tokens": 483382817.0, + "step": 18679 + }, + { + "epoch": 2.0513946848231934, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.188849449157715, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.7539972066879272, + "num_tokens": 483411726.0, + "step": 18680 + }, + { + "epoch": 2.051504502525807, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.3343536853790283, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7343453764915466, + "num_tokens": 483440186.0, + "step": 18681 + }, + { + "epoch": 2.051614320228421, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.524920701980591, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.738204300403595, + "num_tokens": 483463864.0, + "step": 18682 + }, + { + "epoch": 2.0517241379310347, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.326261043548584, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.732563853263855, + "num_tokens": 483488716.0, + "step": 18683 + }, + { + "epoch": 2.051833955633648, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1986145973205566, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7162929773330688, + "num_tokens": 483516897.0, + "step": 18684 + }, + { + "epoch": 2.0519437733362618, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.604464530944824, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7375988364219666, + "num_tokens": 483539481.0, + "step": 18685 + }, + { + "epoch": 2.0520535910388755, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4402120113372803, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7501766085624695, + "num_tokens": 483562160.0, + "step": 18686 + }, + { + "epoch": 2.0521634087414893, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.505340814590454, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7245667576789856, + "num_tokens": 483587201.0, + "step": 18687 + }, + { + "epoch": 2.0522732264441026, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4947152137756348, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7374073266983032, + "num_tokens": 483610822.0, + "step": 18688 + }, + { + "epoch": 2.0523830441467164, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.587617874145508, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7274676561355591, + "num_tokens": 483632577.0, + "step": 18689 + }, + { + "epoch": 2.05249286184933, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2550504207611084, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7207242846488953, + "num_tokens": 483659643.0, + "step": 18690 + }, + { + "epoch": 2.052602679551944, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.409928798675537, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7324651479721069, + "num_tokens": 483685053.0, + "step": 18691 + }, + { + "epoch": 2.0527124972545576, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.572906732559204, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7284088730812073, + "num_tokens": 483708837.0, + "step": 18692 + }, + { + "epoch": 2.052822314957171, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4183781147003174, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7337190508842468, + "num_tokens": 483732545.0, + "step": 18693 + }, + { + "epoch": 2.0529321326597847, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.439969062805176, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7473046779632568, + "num_tokens": 483755112.0, + "step": 18694 + }, + { + "epoch": 2.0530419503623984, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.5541763305664062, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7409937977790833, + "num_tokens": 483777305.0, + "step": 18695 + }, + { + "epoch": 2.053151768065012, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4248478412628174, + "learning_rate": 1e-06, + "loss": 0.7433, + "mean_token_accuracy": 0.7589316368103027, + "num_tokens": 483800500.0, + "step": 18696 + }, + { + "epoch": 2.0532615857676255, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.315329074859619, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7340649366378784, + "num_tokens": 483826162.0, + "step": 18697 + }, + { + "epoch": 2.0533714034702393, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.0572686195373535, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7334394454956055, + "num_tokens": 483855322.0, + "step": 18698 + }, + { + "epoch": 2.053481221172853, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.6383681297302246, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7379140853881836, + "num_tokens": 483877944.0, + "step": 18699 + }, + { + "epoch": 2.053591038875467, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.255828857421875, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.736135721206665, + "num_tokens": 483904372.0, + "step": 18700 + }, + { + "epoch": 2.0537008565780805, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.413774013519287, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7045751810073853, + "num_tokens": 483933983.0, + "step": 18701 + }, + { + "epoch": 2.053810674280694, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.4410276412963867, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7227013111114502, + "num_tokens": 483960201.0, + "step": 18702 + }, + { + "epoch": 2.0539204919833076, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.1201391220092773, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7148863077163696, + "num_tokens": 483990681.0, + "step": 18703 + }, + { + "epoch": 2.0540303096859214, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.196441411972046, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7361462712287903, + "num_tokens": 484019494.0, + "step": 18704 + }, + { + "epoch": 2.054140127388535, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.395625352859497, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7305318117141724, + "num_tokens": 484044720.0, + "step": 18705 + }, + { + "epoch": 2.054249945091149, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.2427027225494385, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7300084829330444, + "num_tokens": 484071022.0, + "step": 18706 + }, + { + "epoch": 2.054359762793762, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.612605571746826, + "learning_rate": 1e-06, + "loss": 0.7475, + "mean_token_accuracy": 0.7603937983512878, + "num_tokens": 484089798.0, + "step": 18707 + }, + { + "epoch": 2.054469580496376, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.391824960708618, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7406540513038635, + "num_tokens": 484115280.0, + "step": 18708 + }, + { + "epoch": 2.0545793981989897, + "ewc_loss": 0.00016021728515625, + "grad_norm": 2.5230159759521484, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7251696586608887, + "num_tokens": 484139932.0, + "step": 18709 + }, + { + "epoch": 2.0546892159016035, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.038975715637207, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.7524292469024658, + "num_tokens": 484170720.0, + "step": 18710 + }, + { + "epoch": 2.054799033604217, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.151155948638916, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7228150367736816, + "num_tokens": 484199463.0, + "step": 18711 + }, + { + "epoch": 2.0549088513068305, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.5500614643096924, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7451380491256714, + "num_tokens": 484223309.0, + "step": 18712 + }, + { + "epoch": 2.0550186690094443, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.365812301635742, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7129911184310913, + "num_tokens": 484252023.0, + "step": 18713 + }, + { + "epoch": 2.055128486712058, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.498807907104492, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7348660230636597, + "num_tokens": 484276029.0, + "step": 18714 + }, + { + "epoch": 2.055238304414672, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.253692865371704, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7141232490539551, + "num_tokens": 484305587.0, + "step": 18715 + }, + { + "epoch": 2.055348122117285, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 3.0140175819396973, + "learning_rate": 1e-06, + "loss": 0.7604, + "mean_token_accuracy": 0.7577786445617676, + "num_tokens": 484324145.0, + "step": 18716 + }, + { + "epoch": 2.055457939819899, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.0523428916931152, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7249698638916016, + "num_tokens": 484356125.0, + "step": 18717 + }, + { + "epoch": 2.0555677575225126, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.3081648349761963, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7252885103225708, + "num_tokens": 484385600.0, + "step": 18718 + }, + { + "epoch": 2.0556775752251264, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.470787286758423, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7551788687705994, + "num_tokens": 484410154.0, + "step": 18719 + }, + { + "epoch": 2.05578739292774, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.5235297679901123, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7464808225631714, + "num_tokens": 484433103.0, + "step": 18720 + }, + { + "epoch": 2.0558972106303535, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.2511589527130127, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7300399541854858, + "num_tokens": 484461954.0, + "step": 18721 + }, + { + "epoch": 2.0560070283329672, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.376537561416626, + "learning_rate": 1e-06, + "loss": 0.7443, + "mean_token_accuracy": 0.7618734836578369, + "num_tokens": 484484988.0, + "step": 18722 + }, + { + "epoch": 2.056116846035581, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.1122264862060547, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7266941070556641, + "num_tokens": 484516484.0, + "step": 18723 + }, + { + "epoch": 2.0562266637381947, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.438171148300171, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7277594208717346, + "num_tokens": 484539532.0, + "step": 18724 + }, + { + "epoch": 2.056336481440808, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.5795085430145264, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7275600433349609, + "num_tokens": 484564198.0, + "step": 18725 + }, + { + "epoch": 2.056446299143422, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.3526854515075684, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7219587564468384, + "num_tokens": 484590622.0, + "step": 18726 + }, + { + "epoch": 2.0565561168460356, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.235417127609253, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.6987359523773193, + "num_tokens": 484619910.0, + "step": 18727 + }, + { + "epoch": 2.0566659345486493, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.6566109657287598, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7383103370666504, + "num_tokens": 484640093.0, + "step": 18728 + }, + { + "epoch": 2.056775752251263, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.2335121631622314, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7205723524093628, + "num_tokens": 484669591.0, + "step": 18729 + }, + { + "epoch": 2.0568855699538764, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.411388397216797, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7313863039016724, + "num_tokens": 484693323.0, + "step": 18730 + }, + { + "epoch": 2.05699538765649, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.6261322498321533, + "learning_rate": 1e-06, + "loss": 0.753, + "mean_token_accuracy": 0.7615464925765991, + "num_tokens": 484713895.0, + "step": 18731 + }, + { + "epoch": 2.057105205359104, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.6749045848846436, + "learning_rate": 1e-06, + "loss": 0.7179, + "mean_token_accuracy": 0.7701480388641357, + "num_tokens": 484734883.0, + "step": 18732 + }, + { + "epoch": 2.0572150230617177, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.494555711746216, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7368834018707275, + "num_tokens": 484757425.0, + "step": 18733 + }, + { + "epoch": 2.0573248407643314, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.3109304904937744, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.735233724117279, + "num_tokens": 484782516.0, + "step": 18734 + }, + { + "epoch": 2.0574346584669447, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.679441452026367, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7440483570098877, + "num_tokens": 484805520.0, + "step": 18735 + }, + { + "epoch": 2.0575444761695585, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.320779323577881, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.707756757736206, + "num_tokens": 484833321.0, + "step": 18736 + }, + { + "epoch": 2.0576542938721722, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.0760979652404785, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.6960306763648987, + "num_tokens": 484868316.0, + "step": 18737 + }, + { + "epoch": 2.057764111574786, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.4277048110961914, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7377853989601135, + "num_tokens": 484892628.0, + "step": 18738 + }, + { + "epoch": 2.0578739292773993, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.2403154373168945, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7268273234367371, + "num_tokens": 484919840.0, + "step": 18739 + }, + { + "epoch": 2.057983746980013, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.10384202003479, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7387409210205078, + "num_tokens": 484949046.0, + "step": 18740 + }, + { + "epoch": 2.058093564682627, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.4670891761779785, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7213166952133179, + "num_tokens": 484973166.0, + "step": 18741 + }, + { + "epoch": 2.0582033823852406, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.4195330142974854, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7219529747962952, + "num_tokens": 484997745.0, + "step": 18742 + }, + { + "epoch": 2.0583132000878543, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.108091115951538, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7317759394645691, + "num_tokens": 485030912.0, + "step": 18743 + }, + { + "epoch": 2.0584230177904677, + "ewc_loss": 0.00016117095947265625, + "grad_norm": 2.2059943675994873, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7323817014694214, + "num_tokens": 485059141.0, + "step": 18744 + }, + { + "epoch": 2.0585328354930814, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.4013547897338867, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7440706491470337, + "num_tokens": 485083742.0, + "step": 18745 + }, + { + "epoch": 2.058642653195695, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.6831982135772705, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7419610619544983, + "num_tokens": 485105674.0, + "step": 18746 + }, + { + "epoch": 2.058752470898309, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.1956827640533447, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7315394878387451, + "num_tokens": 485132428.0, + "step": 18747 + }, + { + "epoch": 2.0588622886009227, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.4784457683563232, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7427860498428345, + "num_tokens": 485155653.0, + "step": 18748 + }, + { + "epoch": 2.058972106303536, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.3375983238220215, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7291586399078369, + "num_tokens": 485183472.0, + "step": 18749 + }, + { + "epoch": 2.0590819240061498, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.304640293121338, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7258188128471375, + "num_tokens": 485208921.0, + "step": 18750 + }, + { + "epoch": 2.0591917417087635, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.0351386070251465, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7203429937362671, + "num_tokens": 485243201.0, + "step": 18751 + }, + { + "epoch": 2.0593015594113773, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.162982702255249, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7309773564338684, + "num_tokens": 485274135.0, + "step": 18752 + }, + { + "epoch": 2.0594113771139906, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.2973744869232178, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7234681248664856, + "num_tokens": 485303106.0, + "step": 18753 + }, + { + "epoch": 2.0595211948166043, + "ewc_loss": 0.0001621246337890625, + "grad_norm": 2.395341157913208, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7386977672576904, + "num_tokens": 485326573.0, + "step": 18754 + }, + { + "epoch": 2.059631012519218, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.24666166305542, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7343944907188416, + "num_tokens": 485354403.0, + "step": 18755 + }, + { + "epoch": 2.059740830221832, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.399536609649658, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7220249176025391, + "num_tokens": 485381116.0, + "step": 18756 + }, + { + "epoch": 2.0598506479244456, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.440315008163452, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.747140109539032, + "num_tokens": 485404852.0, + "step": 18757 + }, + { + "epoch": 2.059960465627059, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3525214195251465, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.7451536059379578, + "num_tokens": 485429368.0, + "step": 18758 + }, + { + "epoch": 2.0600702833296727, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0611636638641357, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7188713550567627, + "num_tokens": 485462570.0, + "step": 18759 + }, + { + "epoch": 2.0601801010322864, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2943508625030518, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7517503499984741, + "num_tokens": 485486324.0, + "step": 18760 + }, + { + "epoch": 2.0602899187349, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3447656631469727, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7369261980056763, + "num_tokens": 485514079.0, + "step": 18761 + }, + { + "epoch": 2.060399736437514, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3464479446411133, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7143809795379639, + "num_tokens": 485539712.0, + "step": 18762 + }, + { + "epoch": 2.0605095541401273, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3437891006469727, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7181753516197205, + "num_tokens": 485565998.0, + "step": 18763 + }, + { + "epoch": 2.060619371842741, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.553088426589966, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7222309112548828, + "num_tokens": 485590371.0, + "step": 18764 + }, + { + "epoch": 2.060729189545355, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4878880977630615, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7334238886833191, + "num_tokens": 485613661.0, + "step": 18765 + }, + { + "epoch": 2.0608390072479685, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3453352451324463, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.73459392786026, + "num_tokens": 485640582.0, + "step": 18766 + }, + { + "epoch": 2.060948824950582, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.345449209213257, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7163246273994446, + "num_tokens": 485667177.0, + "step": 18767 + }, + { + "epoch": 2.0610586426531956, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1794841289520264, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.738701581954956, + "num_tokens": 485692337.0, + "step": 18768 + }, + { + "epoch": 2.0611684603558094, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.490304708480835, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7214552164077759, + "num_tokens": 485715895.0, + "step": 18769 + }, + { + "epoch": 2.061278278058423, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.436880111694336, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7235932350158691, + "num_tokens": 485740056.0, + "step": 18770 + }, + { + "epoch": 2.061388095761037, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3304660320281982, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7363908290863037, + "num_tokens": 485765531.0, + "step": 18771 + }, + { + "epoch": 2.06149791346365, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.283893585205078, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7341033220291138, + "num_tokens": 485790828.0, + "step": 18772 + }, + { + "epoch": 2.061607731166264, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1725921630859375, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.714548647403717, + "num_tokens": 485821307.0, + "step": 18773 + }, + { + "epoch": 2.0617175488688777, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4905781745910645, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7482446432113647, + "num_tokens": 485842368.0, + "step": 18774 + }, + { + "epoch": 2.0618273665714915, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.265835762023926, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7315759062767029, + "num_tokens": 485869451.0, + "step": 18775 + }, + { + "epoch": 2.0619371842741048, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.359065294265747, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.726647138595581, + "num_tokens": 485894433.0, + "step": 18776 + }, + { + "epoch": 2.0620470019767185, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6782608032226562, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7457794547080994, + "num_tokens": 485915047.0, + "step": 18777 + }, + { + "epoch": 2.0621568196793323, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.297109365463257, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7273343801498413, + "num_tokens": 485942877.0, + "step": 18778 + }, + { + "epoch": 2.062266637381946, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4950785636901855, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7339900732040405, + "num_tokens": 485966534.0, + "step": 18779 + }, + { + "epoch": 2.06237645508456, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.822899580001831, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7470211982727051, + "num_tokens": 485985634.0, + "step": 18780 + }, + { + "epoch": 2.062486272787173, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.591050624847412, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7324777245521545, + "num_tokens": 486012183.0, + "step": 18781 + }, + { + "epoch": 2.062596090489787, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2528927326202393, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7222375273704529, + "num_tokens": 486041381.0, + "step": 18782 + }, + { + "epoch": 2.0627059081924006, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3198161125183105, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.739460825920105, + "num_tokens": 486067899.0, + "step": 18783 + }, + { + "epoch": 2.0628157258950144, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.536768913269043, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7336699366569519, + "num_tokens": 486090785.0, + "step": 18784 + }, + { + "epoch": 2.062925543597628, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.249420404434204, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7279620170593262, + "num_tokens": 486117247.0, + "step": 18785 + }, + { + "epoch": 2.0630353613002415, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.442331314086914, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7334504723548889, + "num_tokens": 486142138.0, + "step": 18786 + }, + { + "epoch": 2.063145179002855, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.927241802215576, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7489004731178284, + "num_tokens": 486162436.0, + "step": 18787 + }, + { + "epoch": 2.063254996705469, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6826558113098145, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7527528405189514, + "num_tokens": 486183170.0, + "step": 18788 + }, + { + "epoch": 2.0633648144080827, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0917751789093018, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7158609628677368, + "num_tokens": 486219250.0, + "step": 18789 + }, + { + "epoch": 2.063474632110696, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.429645299911499, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7205816507339478, + "num_tokens": 486248913.0, + "step": 18790 + }, + { + "epoch": 2.06358444981331, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6223814487457275, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7403491735458374, + "num_tokens": 486270415.0, + "step": 18791 + }, + { + "epoch": 2.0636942675159236, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2917098999023438, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7113205194473267, + "num_tokens": 486298050.0, + "step": 18792 + }, + { + "epoch": 2.0638040852185373, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.423948049545288, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7224945425987244, + "num_tokens": 486321495.0, + "step": 18793 + }, + { + "epoch": 2.063913902921151, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3431248664855957, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7415028810501099, + "num_tokens": 486346931.0, + "step": 18794 + }, + { + "epoch": 2.0640237206237644, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4039909839630127, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7187004089355469, + "num_tokens": 486374258.0, + "step": 18795 + }, + { + "epoch": 2.064133538326378, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3133652210235596, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7220029234886169, + "num_tokens": 486401074.0, + "step": 18796 + }, + { + "epoch": 2.064243356028992, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.390547752380371, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7391701936721802, + "num_tokens": 486425411.0, + "step": 18797 + }, + { + "epoch": 2.0643531737316057, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.434605598449707, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7298839092254639, + "num_tokens": 486450354.0, + "step": 18798 + }, + { + "epoch": 2.0644629914342194, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3038103580474854, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7433123588562012, + "num_tokens": 486479627.0, + "step": 18799 + }, + { + "epoch": 2.0645728091368327, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4673593044281006, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7354353070259094, + "num_tokens": 486502869.0, + "step": 18800 + }, + { + "epoch": 2.0646826268394465, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0527591705322266, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7172263860702515, + "num_tokens": 486534327.0, + "step": 18801 + }, + { + "epoch": 2.0647924445420602, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4056246280670166, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7182012796401978, + "num_tokens": 486559490.0, + "step": 18802 + }, + { + "epoch": 2.064902262244674, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6428768634796143, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7338136434555054, + "num_tokens": 486579601.0, + "step": 18803 + }, + { + "epoch": 2.0650120799472873, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.614347457885742, + "learning_rate": 1e-06, + "loss": 0.783, + "mean_token_accuracy": 0.7569875717163086, + "num_tokens": 486600957.0, + "step": 18804 + }, + { + "epoch": 2.065121897649901, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3832082748413086, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7374534606933594, + "num_tokens": 486627356.0, + "step": 18805 + }, + { + "epoch": 2.065231715352515, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3482930660247803, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7348757982254028, + "num_tokens": 486653002.0, + "step": 18806 + }, + { + "epoch": 2.0653415330551286, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3843111991882324, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.745867133140564, + "num_tokens": 486677465.0, + "step": 18807 + }, + { + "epoch": 2.0654513507577423, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3452863693237305, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7250058650970459, + "num_tokens": 486704398.0, + "step": 18808 + }, + { + "epoch": 2.0655611684603556, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1819562911987305, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7270392775535583, + "num_tokens": 486735319.0, + "step": 18809 + }, + { + "epoch": 2.0656709861629694, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.617732524871826, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7263175845146179, + "num_tokens": 486756914.0, + "step": 18810 + }, + { + "epoch": 2.065780803865583, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.324843645095825, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7408239841461182, + "num_tokens": 486781933.0, + "step": 18811 + }, + { + "epoch": 2.065890621568197, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.31014084815979, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7295637726783752, + "num_tokens": 486808660.0, + "step": 18812 + }, + { + "epoch": 2.0660004392708107, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0837433338165283, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7150890827178955, + "num_tokens": 486844039.0, + "step": 18813 + }, + { + "epoch": 2.066110256973424, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0033950805664062, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7247635126113892, + "num_tokens": 486878633.0, + "step": 18814 + }, + { + "epoch": 2.0662200746760377, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2341582775115967, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7391918301582336, + "num_tokens": 486907779.0, + "step": 18815 + }, + { + "epoch": 2.0663298923786515, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2815370559692383, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7191022038459778, + "num_tokens": 486938039.0, + "step": 18816 + }, + { + "epoch": 2.0664397100812653, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2238240242004395, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7114976644515991, + "num_tokens": 486967462.0, + "step": 18817 + }, + { + "epoch": 2.0665495277838786, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2435197830200195, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7438891530036926, + "num_tokens": 486994806.0, + "step": 18818 + }, + { + "epoch": 2.0666593454864923, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.375441074371338, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7126755714416504, + "num_tokens": 487021312.0, + "step": 18819 + }, + { + "epoch": 2.066769163189106, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2972185611724854, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.733730673789978, + "num_tokens": 487048531.0, + "step": 18820 + }, + { + "epoch": 2.06687898089172, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.389279365539551, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7348726987838745, + "num_tokens": 487074614.0, + "step": 18821 + }, + { + "epoch": 2.0669887985943336, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4869139194488525, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7328207492828369, + "num_tokens": 487097613.0, + "step": 18822 + }, + { + "epoch": 2.067098616296947, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3580996990203857, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7247977256774902, + "num_tokens": 487124192.0, + "step": 18823 + }, + { + "epoch": 2.0672084339995607, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0571415424346924, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.6988351345062256, + "num_tokens": 487157659.0, + "step": 18824 + }, + { + "epoch": 2.0673182517021744, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1780307292938232, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7544592022895813, + "num_tokens": 487186665.0, + "step": 18825 + }, + { + "epoch": 2.067428069404788, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.274218797683716, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.7544772624969482, + "num_tokens": 487213021.0, + "step": 18826 + }, + { + "epoch": 2.0675378871074015, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4727156162261963, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7266228199005127, + "num_tokens": 487236826.0, + "step": 18827 + }, + { + "epoch": 2.0676477048100153, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.516298770904541, + "learning_rate": 1e-06, + "loss": 0.7123, + "mean_token_accuracy": 0.7638180255889893, + "num_tokens": 487257249.0, + "step": 18828 + }, + { + "epoch": 2.067757522512629, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4652535915374756, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7384567260742188, + "num_tokens": 487279719.0, + "step": 18829 + }, + { + "epoch": 2.0678673402152428, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2236812114715576, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7287706732749939, + "num_tokens": 487305998.0, + "step": 18830 + }, + { + "epoch": 2.0679771579178565, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.15596604347229, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7464742064476013, + "num_tokens": 487335168.0, + "step": 18831 + }, + { + "epoch": 2.06808697562047, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.442147970199585, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7352510690689087, + "num_tokens": 487359653.0, + "step": 18832 + }, + { + "epoch": 2.0681967933230836, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6129136085510254, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7487375736236572, + "num_tokens": 487379403.0, + "step": 18833 + }, + { + "epoch": 2.0683066110256974, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.440074920654297, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7323445081710815, + "num_tokens": 487403805.0, + "step": 18834 + }, + { + "epoch": 2.068416428728311, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.318918466567993, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7342143654823303, + "num_tokens": 487429302.0, + "step": 18835 + }, + { + "epoch": 2.068526246430925, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2856032848358154, + "learning_rate": 1e-06, + "loss": 0.7583, + "mean_token_accuracy": 0.7570089101791382, + "num_tokens": 487454107.0, + "step": 18836 + }, + { + "epoch": 2.068636064133538, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2863879203796387, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7386773824691772, + "num_tokens": 487480299.0, + "step": 18837 + }, + { + "epoch": 2.068745881836152, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.35982608795166, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7227827310562134, + "num_tokens": 487507207.0, + "step": 18838 + }, + { + "epoch": 2.0688556995387657, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.398189067840576, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7461733818054199, + "num_tokens": 487532758.0, + "step": 18839 + }, + { + "epoch": 2.0689655172413794, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.5782642364501953, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.747214138507843, + "num_tokens": 487555940.0, + "step": 18840 + }, + { + "epoch": 2.0690753349439928, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.5218513011932373, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7217731475830078, + "num_tokens": 487581254.0, + "step": 18841 + }, + { + "epoch": 2.0691851526466065, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.7836029529571533, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7527987957000732, + "num_tokens": 487600889.0, + "step": 18842 + }, + { + "epoch": 2.0692949703492203, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6882662773132324, + "learning_rate": 1e-06, + "loss": 0.7748, + "mean_token_accuracy": 0.751998245716095, + "num_tokens": 487620106.0, + "step": 18843 + }, + { + "epoch": 2.069404788051834, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0345304012298584, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.751166820526123, + "num_tokens": 487652019.0, + "step": 18844 + }, + { + "epoch": 2.069514605754448, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0940122604370117, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7316756248474121, + "num_tokens": 487682072.0, + "step": 18845 + }, + { + "epoch": 2.069624423457061, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.0889892578125, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7269989252090454, + "num_tokens": 487712303.0, + "step": 18846 + }, + { + "epoch": 2.069734241159675, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.558701753616333, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.744953453540802, + "num_tokens": 487733598.0, + "step": 18847 + }, + { + "epoch": 2.0698440588622886, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3498170375823975, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7381870746612549, + "num_tokens": 487758116.0, + "step": 18848 + }, + { + "epoch": 2.0699538765649024, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3368730545043945, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7168753147125244, + "num_tokens": 487784928.0, + "step": 18849 + }, + { + "epoch": 2.070063694267516, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.488903522491455, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.731788158416748, + "num_tokens": 487806999.0, + "step": 18850 + }, + { + "epoch": 2.0701735119701294, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.42419171333313, + "learning_rate": 1e-06, + "loss": 0.7597, + "mean_token_accuracy": 0.7507326602935791, + "num_tokens": 487829487.0, + "step": 18851 + }, + { + "epoch": 2.070283329672743, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.253321886062622, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7268354892730713, + "num_tokens": 487857701.0, + "step": 18852 + }, + { + "epoch": 2.070393147375357, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.224090099334717, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7250000834465027, + "num_tokens": 487886019.0, + "step": 18853 + }, + { + "epoch": 2.0705029650779707, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.5982277393341064, + "learning_rate": 1e-06, + "loss": 0.7837, + "mean_token_accuracy": 0.7522587776184082, + "num_tokens": 487906054.0, + "step": 18854 + }, + { + "epoch": 2.070612782780584, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2680187225341797, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7406363487243652, + "num_tokens": 487932732.0, + "step": 18855 + }, + { + "epoch": 2.070722600483198, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.528186082839966, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7202845811843872, + "num_tokens": 487954950.0, + "step": 18856 + }, + { + "epoch": 2.0708324181858115, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 1.942960500717163, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7270839214324951, + "num_tokens": 487989789.0, + "step": 18857 + }, + { + "epoch": 2.0709422358884253, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4400031566619873, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7311519384384155, + "num_tokens": 488013438.0, + "step": 18858 + }, + { + "epoch": 2.071052053591039, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.676074266433716, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7399122714996338, + "num_tokens": 488033967.0, + "step": 18859 + }, + { + "epoch": 2.0711618712936524, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.661756753921509, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.735535740852356, + "num_tokens": 488055618.0, + "step": 18860 + }, + { + "epoch": 2.071271688996266, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.525330066680908, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7352296710014343, + "num_tokens": 488077824.0, + "step": 18861 + }, + { + "epoch": 2.07138150669888, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.440678358078003, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.721097469329834, + "num_tokens": 488103374.0, + "step": 18862 + }, + { + "epoch": 2.0714913244014936, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.256223201751709, + "learning_rate": 1e-06, + "loss": 0.7881, + "mean_token_accuracy": 0.7460055947303772, + "num_tokens": 488129776.0, + "step": 18863 + }, + { + "epoch": 2.0716011421041074, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.433842182159424, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7233300805091858, + "num_tokens": 488154567.0, + "step": 18864 + }, + { + "epoch": 2.0717109598067207, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6643478870391846, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7392177581787109, + "num_tokens": 488175092.0, + "step": 18865 + }, + { + "epoch": 2.0718207775093345, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.401247024536133, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7078893780708313, + "num_tokens": 488202329.0, + "step": 18866 + }, + { + "epoch": 2.0719305952119482, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.394911527633667, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7259103655815125, + "num_tokens": 488229448.0, + "step": 18867 + }, + { + "epoch": 2.072040412914562, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.8506243228912354, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7387423515319824, + "num_tokens": 488248341.0, + "step": 18868 + }, + { + "epoch": 2.0721502306171753, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.278982639312744, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7145562171936035, + "num_tokens": 488276403.0, + "step": 18869 + }, + { + "epoch": 2.072260048319789, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3928792476654053, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.718932032585144, + "num_tokens": 488300829.0, + "step": 18870 + }, + { + "epoch": 2.072369866022403, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2276248931884766, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7200361490249634, + "num_tokens": 488330090.0, + "step": 18871 + }, + { + "epoch": 2.0724796837250166, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.196739435195923, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7340357303619385, + "num_tokens": 488357430.0, + "step": 18872 + }, + { + "epoch": 2.0725895014276303, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3940608501434326, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7316205501556396, + "num_tokens": 488382389.0, + "step": 18873 + }, + { + "epoch": 2.0726993191302436, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4452064037323, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7444902658462524, + "num_tokens": 488405819.0, + "step": 18874 + }, + { + "epoch": 2.0728091368328574, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4432766437530518, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7283643484115601, + "num_tokens": 488428902.0, + "step": 18875 + }, + { + "epoch": 2.072918954535471, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.478869915008545, + "learning_rate": 1e-06, + "loss": 0.7798, + "mean_token_accuracy": 0.7495040893554688, + "num_tokens": 488451267.0, + "step": 18876 + }, + { + "epoch": 2.073028772238085, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.22611403465271, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7368865013122559, + "num_tokens": 488477786.0, + "step": 18877 + }, + { + "epoch": 2.073138589940698, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.419355869293213, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7451545596122742, + "num_tokens": 488501779.0, + "step": 18878 + }, + { + "epoch": 2.073248407643312, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.385188102722168, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.728397786617279, + "num_tokens": 488525927.0, + "step": 18879 + }, + { + "epoch": 2.0733582253459257, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.555222749710083, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7320270538330078, + "num_tokens": 488549374.0, + "step": 18880 + }, + { + "epoch": 2.0734680430485395, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1218793392181396, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.716316819190979, + "num_tokens": 488581234.0, + "step": 18881 + }, + { + "epoch": 2.0735778607511532, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4857027530670166, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7376347184181213, + "num_tokens": 488604300.0, + "step": 18882 + }, + { + "epoch": 2.0736876784537666, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.435382604598999, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7263288497924805, + "num_tokens": 488628493.0, + "step": 18883 + }, + { + "epoch": 2.0737974961563803, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.587628126144409, + "learning_rate": 1e-06, + "loss": 0.7669, + "mean_token_accuracy": 0.7554101943969727, + "num_tokens": 488649579.0, + "step": 18884 + }, + { + "epoch": 2.073907313858994, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.22597074508667, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7180080413818359, + "num_tokens": 488679453.0, + "step": 18885 + }, + { + "epoch": 2.074017131561608, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.207582473754883, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7375251054763794, + "num_tokens": 488706913.0, + "step": 18886 + }, + { + "epoch": 2.0741269492642216, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2428059577941895, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7369014024734497, + "num_tokens": 488733613.0, + "step": 18887 + }, + { + "epoch": 2.074236766966835, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2641520500183105, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.721786379814148, + "num_tokens": 488760801.0, + "step": 18888 + }, + { + "epoch": 2.0743465846694487, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2773807048797607, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7223477363586426, + "num_tokens": 488788687.0, + "step": 18889 + }, + { + "epoch": 2.0744564023720624, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.406557559967041, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7342627644538879, + "num_tokens": 488814072.0, + "step": 18890 + }, + { + "epoch": 2.074566220074676, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2439017295837402, + "learning_rate": 1e-06, + "loss": 0.789, + "mean_token_accuracy": 0.7578442096710205, + "num_tokens": 488841485.0, + "step": 18891 + }, + { + "epoch": 2.07467603777729, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2658753395080566, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7398417592048645, + "num_tokens": 488867353.0, + "step": 18892 + }, + { + "epoch": 2.0747858554799032, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3606698513031006, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7026478052139282, + "num_tokens": 488896078.0, + "step": 18893 + }, + { + "epoch": 2.074895673182517, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.737576961517334, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7311776876449585, + "num_tokens": 488917064.0, + "step": 18894 + }, + { + "epoch": 2.0750054908851308, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2001760005950928, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7238684892654419, + "num_tokens": 488945488.0, + "step": 18895 + }, + { + "epoch": 2.0751153085877445, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3511016368865967, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7376889586448669, + "num_tokens": 488968188.0, + "step": 18896 + }, + { + "epoch": 2.075225126290358, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.773250102996826, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7417563796043396, + "num_tokens": 488989368.0, + "step": 18897 + }, + { + "epoch": 2.0753349439929716, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.268582582473755, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7231305837631226, + "num_tokens": 489019061.0, + "step": 18898 + }, + { + "epoch": 2.0754447616955853, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2010185718536377, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7174800634384155, + "num_tokens": 489049187.0, + "step": 18899 + }, + { + "epoch": 2.075554579398199, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.493245840072632, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7423015832901001, + "num_tokens": 489072262.0, + "step": 18900 + }, + { + "epoch": 2.075664397100813, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2909631729125977, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7361459732055664, + "num_tokens": 489098065.0, + "step": 18901 + }, + { + "epoch": 2.075774214803426, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4648118019104004, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7329035997390747, + "num_tokens": 489122738.0, + "step": 18902 + }, + { + "epoch": 2.07588403250604, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3977487087249756, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7193606495857239, + "num_tokens": 489149658.0, + "step": 18903 + }, + { + "epoch": 2.0759938502086537, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.576615571975708, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7438733577728271, + "num_tokens": 489170115.0, + "step": 18904 + }, + { + "epoch": 2.0761036679112674, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.258878469467163, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7090524435043335, + "num_tokens": 489198899.0, + "step": 18905 + }, + { + "epoch": 2.0762134856138807, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6196036338806152, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7306077480316162, + "num_tokens": 489220994.0, + "step": 18906 + }, + { + "epoch": 2.0763233033164945, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.22273850440979, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7251831293106079, + "num_tokens": 489247684.0, + "step": 18907 + }, + { + "epoch": 2.0764331210191083, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2759389877319336, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7415874004364014, + "num_tokens": 489274000.0, + "step": 18908 + }, + { + "epoch": 2.076542938721722, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2234160900115967, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7329205870628357, + "num_tokens": 489301138.0, + "step": 18909 + }, + { + "epoch": 2.0766527564243358, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.574662685394287, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7369987368583679, + "num_tokens": 489323949.0, + "step": 18910 + }, + { + "epoch": 2.076762574126949, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.347079038619995, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7297662496566772, + "num_tokens": 489348177.0, + "step": 18911 + }, + { + "epoch": 2.076872391829563, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.5175771713256836, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7241246700286865, + "num_tokens": 489370653.0, + "step": 18912 + }, + { + "epoch": 2.0769822095321766, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.526902198791504, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7226442694664001, + "num_tokens": 489396076.0, + "step": 18913 + }, + { + "epoch": 2.0770920272347904, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1812286376953125, + "learning_rate": 1e-06, + "loss": 0.7609, + "mean_token_accuracy": 0.7638816833496094, + "num_tokens": 489422322.0, + "step": 18914 + }, + { + "epoch": 2.077201844937404, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.204533815383911, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7181175947189331, + "num_tokens": 489452241.0, + "step": 18915 + }, + { + "epoch": 2.0773116626400174, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.6438562870025635, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7428430318832397, + "num_tokens": 489471717.0, + "step": 18916 + }, + { + "epoch": 2.077421480342631, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.555908203125, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7531957626342773, + "num_tokens": 489493599.0, + "step": 18917 + }, + { + "epoch": 2.077531298045245, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4243857860565186, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7138742208480835, + "num_tokens": 489517991.0, + "step": 18918 + }, + { + "epoch": 2.0776411157478587, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.067412853240967, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7063150405883789, + "num_tokens": 489550331.0, + "step": 18919 + }, + { + "epoch": 2.077750933450472, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.2452595233917236, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7210248112678528, + "num_tokens": 489576755.0, + "step": 18920 + }, + { + "epoch": 2.0778607511530858, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3008599281311035, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7398941516876221, + "num_tokens": 489603232.0, + "step": 18921 + }, + { + "epoch": 2.0779705688556995, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.1640641689300537, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7090363502502441, + "num_tokens": 489633861.0, + "step": 18922 + }, + { + "epoch": 2.0780803865583133, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.5741684436798096, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7290608882904053, + "num_tokens": 489656715.0, + "step": 18923 + }, + { + "epoch": 2.078190204260927, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.7044546604156494, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7345774173736572, + "num_tokens": 489678745.0, + "step": 18924 + }, + { + "epoch": 2.0783000219635404, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.306598663330078, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.725764811038971, + "num_tokens": 489705042.0, + "step": 18925 + }, + { + "epoch": 2.078409839666154, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.3222906589508057, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7291857004165649, + "num_tokens": 489732389.0, + "step": 18926 + }, + { + "epoch": 2.078519657368768, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.4368858337402344, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7041151523590088, + "num_tokens": 489756253.0, + "step": 18927 + }, + { + "epoch": 2.0786294750713816, + "ewc_loss": 0.00016307830810546875, + "grad_norm": 2.48866605758667, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7491761445999146, + "num_tokens": 489779374.0, + "step": 18928 + }, + { + "epoch": 2.078739292773995, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.522881031036377, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7251260876655579, + "num_tokens": 489802145.0, + "step": 18929 + }, + { + "epoch": 2.0788491104766087, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.098510265350342, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7148454785346985, + "num_tokens": 489832481.0, + "step": 18930 + }, + { + "epoch": 2.0789589281792225, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.459963321685791, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7210980653762817, + "num_tokens": 489855325.0, + "step": 18931 + }, + { + "epoch": 2.079068745881836, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.366567373275757, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7295554876327515, + "num_tokens": 489880338.0, + "step": 18932 + }, + { + "epoch": 2.07917856358445, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.4661409854888916, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7395086884498596, + "num_tokens": 489904091.0, + "step": 18933 + }, + { + "epoch": 2.0792883812870633, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.518622398376465, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7274011969566345, + "num_tokens": 489928676.0, + "step": 18934 + }, + { + "epoch": 2.079398198989677, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.2157344818115234, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7270936965942383, + "num_tokens": 489956846.0, + "step": 18935 + }, + { + "epoch": 2.079508016692291, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.4760215282440186, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7270114421844482, + "num_tokens": 489981568.0, + "step": 18936 + }, + { + "epoch": 2.0796178343949046, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.2705562114715576, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.744458019733429, + "num_tokens": 490007835.0, + "step": 18937 + }, + { + "epoch": 2.0797276520975183, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.5602025985717773, + "learning_rate": 1e-06, + "loss": 0.66, + "mean_token_accuracy": 0.7847926616668701, + "num_tokens": 490027937.0, + "step": 18938 + }, + { + "epoch": 2.0798374698001316, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.1481704711914062, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7153212428092957, + "num_tokens": 490063001.0, + "step": 18939 + }, + { + "epoch": 2.0799472875027454, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.33437442779541, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7421067953109741, + "num_tokens": 490088863.0, + "step": 18940 + }, + { + "epoch": 2.080057105205359, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.534977436065674, + "learning_rate": 1e-06, + "loss": 0.7446, + "mean_token_accuracy": 0.7599437236785889, + "num_tokens": 490110820.0, + "step": 18941 + }, + { + "epoch": 2.080166922907973, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.159287214279175, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7194195985794067, + "num_tokens": 490140847.0, + "step": 18942 + }, + { + "epoch": 2.0802767406105866, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.338684320449829, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.7054808139801025, + "num_tokens": 490166679.0, + "step": 18943 + }, + { + "epoch": 2.0803865583132, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.667003870010376, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7423096299171448, + "num_tokens": 490187894.0, + "step": 18944 + }, + { + "epoch": 2.0804963760158137, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 1.9169905185699463, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7268498539924622, + "num_tokens": 490221796.0, + "step": 18945 + }, + { + "epoch": 2.0806061937184275, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.5423223972320557, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7308148145675659, + "num_tokens": 490243502.0, + "step": 18946 + }, + { + "epoch": 2.0807160114210412, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.2106237411499023, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7454789280891418, + "num_tokens": 490271915.0, + "step": 18947 + }, + { + "epoch": 2.0808258291236545, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.2651822566986084, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7343088388442993, + "num_tokens": 490300156.0, + "step": 18948 + }, + { + "epoch": 2.0809356468262683, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.245403528213501, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.702377438545227, + "num_tokens": 490330308.0, + "step": 18949 + }, + { + "epoch": 2.081045464528882, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.217909097671509, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.7134102582931519, + "num_tokens": 490358907.0, + "step": 18950 + }, + { + "epoch": 2.081155282231496, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4736671447753906, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7213045358657837, + "num_tokens": 490383541.0, + "step": 18951 + }, + { + "epoch": 2.0812650999341096, + "ewc_loss": 0.00016498565673828125, + "grad_norm": 2.2932021617889404, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7351478338241577, + "num_tokens": 490409125.0, + "step": 18952 + }, + { + "epoch": 2.081374917636723, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.443650245666504, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7394766211509705, + "num_tokens": 490432306.0, + "step": 18953 + }, + { + "epoch": 2.0814847353393366, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.5858545303344727, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.723537027835846, + "num_tokens": 490454009.0, + "step": 18954 + }, + { + "epoch": 2.0815945530419504, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.2682619094848633, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7170801758766174, + "num_tokens": 490482565.0, + "step": 18955 + }, + { + "epoch": 2.081704370744564, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 1.963095784187317, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7056493759155273, + "num_tokens": 490519834.0, + "step": 18956 + }, + { + "epoch": 2.0818141884471775, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3377525806427, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.7555522322654724, + "num_tokens": 490546074.0, + "step": 18957 + }, + { + "epoch": 2.0819240061497912, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.529118299484253, + "learning_rate": 1e-06, + "loss": 0.729, + "mean_token_accuracy": 0.7605735659599304, + "num_tokens": 490568911.0, + "step": 18958 + }, + { + "epoch": 2.082033823852405, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.312944173812866, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7174269556999207, + "num_tokens": 490594591.0, + "step": 18959 + }, + { + "epoch": 2.0821436415550187, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3685383796691895, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.735238790512085, + "num_tokens": 490619546.0, + "step": 18960 + }, + { + "epoch": 2.0822534592576325, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.340299367904663, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7326886057853699, + "num_tokens": 490648240.0, + "step": 18961 + }, + { + "epoch": 2.082363276960246, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.328629970550537, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7274801731109619, + "num_tokens": 490672451.0, + "step": 18962 + }, + { + "epoch": 2.0824730946628596, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3667824268341064, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7435916066169739, + "num_tokens": 490698015.0, + "step": 18963 + }, + { + "epoch": 2.0825829123654733, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.390310049057007, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7481977939605713, + "num_tokens": 490721407.0, + "step": 18964 + }, + { + "epoch": 2.082692730068087, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.4951677322387695, + "learning_rate": 1e-06, + "loss": 0.7685, + "mean_token_accuracy": 0.754024863243103, + "num_tokens": 490743600.0, + "step": 18965 + }, + { + "epoch": 2.082802547770701, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.4410881996154785, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7357279062271118, + "num_tokens": 490766922.0, + "step": 18966 + }, + { + "epoch": 2.082912365473314, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.720209836959839, + "learning_rate": 1e-06, + "loss": 0.7841, + "mean_token_accuracy": 0.748041033744812, + "num_tokens": 490787071.0, + "step": 18967 + }, + { + "epoch": 2.083022183175928, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.020555019378662, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7325164675712585, + "num_tokens": 490818928.0, + "step": 18968 + }, + { + "epoch": 2.0831320008785417, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.299722194671631, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7222509384155273, + "num_tokens": 490845146.0, + "step": 18969 + }, + { + "epoch": 2.0832418185811554, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.413137912750244, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7233690619468689, + "num_tokens": 490870347.0, + "step": 18970 + }, + { + "epoch": 2.0833516362837687, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3897807598114014, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7243770360946655, + "num_tokens": 490897708.0, + "step": 18971 + }, + { + "epoch": 2.0834614539863825, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.2325494289398193, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.719852089881897, + "num_tokens": 490926444.0, + "step": 18972 + }, + { + "epoch": 2.0835712716889963, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.375955581665039, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7446590662002563, + "num_tokens": 490951897.0, + "step": 18973 + }, + { + "epoch": 2.08368108939161, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 1.9776734113693237, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7053797245025635, + "num_tokens": 490987745.0, + "step": 18974 + }, + { + "epoch": 2.0837909070942238, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.133342742919922, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7141594886779785, + "num_tokens": 491018623.0, + "step": 18975 + }, + { + "epoch": 2.083900724796837, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.4269602298736572, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7343944311141968, + "num_tokens": 491042741.0, + "step": 18976 + }, + { + "epoch": 2.084010542499451, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.2736825942993164, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7303770184516907, + "num_tokens": 491068750.0, + "step": 18977 + }, + { + "epoch": 2.0841203602020646, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.428154945373535, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7132159471511841, + "num_tokens": 491092239.0, + "step": 18978 + }, + { + "epoch": 2.0842301779046783, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.51772141456604, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7500219345092773, + "num_tokens": 491114475.0, + "step": 18979 + }, + { + "epoch": 2.084339995607292, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.6931936740875244, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7430113554000854, + "num_tokens": 491134611.0, + "step": 18980 + }, + { + "epoch": 2.0844498133099054, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.645052909851074, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7360897064208984, + "num_tokens": 491155132.0, + "step": 18981 + }, + { + "epoch": 2.084559631012519, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.520735025405884, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7423866391181946, + "num_tokens": 491178243.0, + "step": 18982 + }, + { + "epoch": 2.084669448715133, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.1433310508728027, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.7060044407844543, + "num_tokens": 491209504.0, + "step": 18983 + }, + { + "epoch": 2.0847792664177467, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.5848119258880615, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7282328605651855, + "num_tokens": 491234330.0, + "step": 18984 + }, + { + "epoch": 2.08488908412036, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.231088161468506, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7334036827087402, + "num_tokens": 491262831.0, + "step": 18985 + }, + { + "epoch": 2.0849989018229738, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3268368244171143, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7104207277297974, + "num_tokens": 491290913.0, + "step": 18986 + }, + { + "epoch": 2.0851087195255875, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.518937826156616, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7164428234100342, + "num_tokens": 491314614.0, + "step": 18987 + }, + { + "epoch": 2.0852185372282013, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.209362030029297, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.716550350189209, + "num_tokens": 491344133.0, + "step": 18988 + }, + { + "epoch": 2.085328354930815, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.375972032546997, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7311025857925415, + "num_tokens": 491370174.0, + "step": 18989 + }, + { + "epoch": 2.0854381726334283, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.403407335281372, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7389957904815674, + "num_tokens": 491396116.0, + "step": 18990 + }, + { + "epoch": 2.085547990336042, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.1239051818847656, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7028782367706299, + "num_tokens": 491430346.0, + "step": 18991 + }, + { + "epoch": 2.085657808038656, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.296125650405884, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7310032248497009, + "num_tokens": 491458979.0, + "step": 18992 + }, + { + "epoch": 2.0857676257412696, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.457393169403076, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7464966773986816, + "num_tokens": 491484176.0, + "step": 18993 + }, + { + "epoch": 2.0858774434438834, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.377763032913208, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7306310534477234, + "num_tokens": 491509964.0, + "step": 18994 + }, + { + "epoch": 2.0859872611464967, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.5932083129882812, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.730876088142395, + "num_tokens": 491531663.0, + "step": 18995 + }, + { + "epoch": 2.0860970788491104, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.758763313293457, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7204837799072266, + "num_tokens": 491554738.0, + "step": 18996 + }, + { + "epoch": 2.086206896551724, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.4301717281341553, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.740801215171814, + "num_tokens": 491578738.0, + "step": 18997 + }, + { + "epoch": 2.086316714254338, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.204392433166504, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7120686173439026, + "num_tokens": 491606169.0, + "step": 18998 + }, + { + "epoch": 2.0864265319569513, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.2041175365448, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7392486333847046, + "num_tokens": 491634519.0, + "step": 18999 + }, + { + "epoch": 2.086536349659565, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3242223262786865, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.706299901008606, + "num_tokens": 491664802.0, + "step": 19000 + }, + { + "epoch": 2.086646167362179, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 1.9589303731918335, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7407540678977966, + "num_tokens": 491699769.0, + "step": 19001 + }, + { + "epoch": 2.0867559850647925, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3771095275878906, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7316493988037109, + "num_tokens": 491726122.0, + "step": 19002 + }, + { + "epoch": 2.0868658027674063, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.4566752910614014, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7409114241600037, + "num_tokens": 491751062.0, + "step": 19003 + }, + { + "epoch": 2.0869756204700196, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.3514585494995117, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7365204095840454, + "num_tokens": 491777880.0, + "step": 19004 + }, + { + "epoch": 2.0870854381726334, + "ewc_loss": 0.0001659393310546875, + "grad_norm": 2.316913366317749, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7285176515579224, + "num_tokens": 491803602.0, + "step": 19005 + }, + { + "epoch": 2.087195255875247, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.1292498111724854, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7308802008628845, + "num_tokens": 491833549.0, + "step": 19006 + }, + { + "epoch": 2.087305073577861, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5904674530029297, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7069764137268066, + "num_tokens": 491855771.0, + "step": 19007 + }, + { + "epoch": 2.087414891280474, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2055304050445557, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7300224304199219, + "num_tokens": 491883597.0, + "step": 19008 + }, + { + "epoch": 2.087524708983088, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.1476919651031494, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.718746542930603, + "num_tokens": 491912746.0, + "step": 19009 + }, + { + "epoch": 2.0876345266857017, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4239680767059326, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7350487112998962, + "num_tokens": 491937029.0, + "step": 19010 + }, + { + "epoch": 2.0877443443883155, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4877209663391113, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.742205798625946, + "num_tokens": 491959920.0, + "step": 19011 + }, + { + "epoch": 2.087854162090929, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.1554036140441895, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7435925006866455, + "num_tokens": 491989044.0, + "step": 19012 + }, + { + "epoch": 2.0879639797935425, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4224696159362793, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7309218645095825, + "num_tokens": 492013024.0, + "step": 19013 + }, + { + "epoch": 2.0880737974961563, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.1755824089050293, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7277899384498596, + "num_tokens": 492044906.0, + "step": 19014 + }, + { + "epoch": 2.08818361519877, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3644163608551025, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7358916997909546, + "num_tokens": 492070927.0, + "step": 19015 + }, + { + "epoch": 2.088293432901384, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.455881118774414, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.726886510848999, + "num_tokens": 492095050.0, + "step": 19016 + }, + { + "epoch": 2.0884032506039976, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.6513259410858154, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7316532731056213, + "num_tokens": 492115775.0, + "step": 19017 + }, + { + "epoch": 2.088513068306611, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.51615834236145, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7281301021575928, + "num_tokens": 492139040.0, + "step": 19018 + }, + { + "epoch": 2.0886228860092246, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3882405757904053, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7235369682312012, + "num_tokens": 492167280.0, + "step": 19019 + }, + { + "epoch": 2.0887327037118384, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.419313430786133, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7247627973556519, + "num_tokens": 492193315.0, + "step": 19020 + }, + { + "epoch": 2.088842521414452, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4807655811309814, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7265480160713196, + "num_tokens": 492217710.0, + "step": 19021 + }, + { + "epoch": 2.0889523391170655, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.769071578979492, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7219395637512207, + "num_tokens": 492238561.0, + "step": 19022 + }, + { + "epoch": 2.089062156819679, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5369808673858643, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7287306785583496, + "num_tokens": 492261352.0, + "step": 19023 + }, + { + "epoch": 2.089171974522293, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3649158477783203, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.728624701499939, + "num_tokens": 492286699.0, + "step": 19024 + }, + { + "epoch": 2.0892817922249067, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5291872024536133, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7435160875320435, + "num_tokens": 492309374.0, + "step": 19025 + }, + { + "epoch": 2.0893916099275205, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.0446479320526123, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7194663882255554, + "num_tokens": 492342546.0, + "step": 19026 + }, + { + "epoch": 2.089501427630134, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3089940547943115, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7497276067733765, + "num_tokens": 492369335.0, + "step": 19027 + }, + { + "epoch": 2.0896112453327476, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5377447605133057, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7235730886459351, + "num_tokens": 492393485.0, + "step": 19028 + }, + { + "epoch": 2.0897210630353613, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.692826986312866, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7300296425819397, + "num_tokens": 492417059.0, + "step": 19029 + }, + { + "epoch": 2.089830880737975, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.7282767295837402, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7314627170562744, + "num_tokens": 492438264.0, + "step": 19030 + }, + { + "epoch": 2.089940698440589, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4280130863189697, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7155916690826416, + "num_tokens": 492463019.0, + "step": 19031 + }, + { + "epoch": 2.090050516143202, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2948386669158936, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7282770276069641, + "num_tokens": 492489882.0, + "step": 19032 + }, + { + "epoch": 2.090160333845816, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3854639530181885, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.745771050453186, + "num_tokens": 492514387.0, + "step": 19033 + }, + { + "epoch": 2.0902701515484297, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.6078531742095947, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7518665790557861, + "num_tokens": 492535709.0, + "step": 19034 + }, + { + "epoch": 2.0903799692510434, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.4598162174224854, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7316253185272217, + "num_tokens": 492558357.0, + "step": 19035 + }, + { + "epoch": 2.0904897869536567, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.404498815536499, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7217158079147339, + "num_tokens": 492585376.0, + "step": 19036 + }, + { + "epoch": 2.0905996046562705, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.173656463623047, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7122247219085693, + "num_tokens": 492614595.0, + "step": 19037 + }, + { + "epoch": 2.0907094223588842, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.7828636169433594, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7461130619049072, + "num_tokens": 492632727.0, + "step": 19038 + }, + { + "epoch": 2.090819240061498, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.119725465774536, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7129522562026978, + "num_tokens": 492662994.0, + "step": 19039 + }, + { + "epoch": 2.0909290577641118, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.312584161758423, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7277443408966064, + "num_tokens": 492690017.0, + "step": 19040 + }, + { + "epoch": 2.091038875466725, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2729976177215576, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7321589589118958, + "num_tokens": 492718107.0, + "step": 19041 + }, + { + "epoch": 2.091148693169339, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5591211318969727, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7336391806602478, + "num_tokens": 492741305.0, + "step": 19042 + }, + { + "epoch": 2.0912585108719526, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2923529148101807, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7307174205780029, + "num_tokens": 492766787.0, + "step": 19043 + }, + { + "epoch": 2.0913683285745663, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.313868284225464, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7321852445602417, + "num_tokens": 492793789.0, + "step": 19044 + }, + { + "epoch": 2.09147814627718, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2669334411621094, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7428446412086487, + "num_tokens": 492820917.0, + "step": 19045 + }, + { + "epoch": 2.0915879639797934, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.1771938800811768, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7123836278915405, + "num_tokens": 492849776.0, + "step": 19046 + }, + { + "epoch": 2.091697781682407, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.2023863792419434, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7413734197616577, + "num_tokens": 492877582.0, + "step": 19047 + }, + { + "epoch": 2.091807599385021, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.3662807941436768, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7052534818649292, + "num_tokens": 492906048.0, + "step": 19048 + }, + { + "epoch": 2.0919174170876347, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.5303404331207275, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7274675369262695, + "num_tokens": 492928817.0, + "step": 19049 + }, + { + "epoch": 2.092027234790248, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.552619218826294, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7357585430145264, + "num_tokens": 492951138.0, + "step": 19050 + }, + { + "epoch": 2.0921370524928617, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 1.9806197881698608, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7391957640647888, + "num_tokens": 492985746.0, + "step": 19051 + }, + { + "epoch": 2.0922468701954755, + "ewc_loss": 0.00016689300537109375, + "grad_norm": 2.414776086807251, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7378416061401367, + "num_tokens": 493008945.0, + "step": 19052 + }, + { + "epoch": 2.0923566878980893, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4813663959503174, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7376551032066345, + "num_tokens": 493032172.0, + "step": 19053 + }, + { + "epoch": 2.092466505600703, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.0299479961395264, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7196553945541382, + "num_tokens": 493064895.0, + "step": 19054 + }, + { + "epoch": 2.0925763233033163, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2825963497161865, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7015075087547302, + "num_tokens": 493094111.0, + "step": 19055 + }, + { + "epoch": 2.09268614100593, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.207780599594116, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7353894710540771, + "num_tokens": 493120831.0, + "step": 19056 + }, + { + "epoch": 2.092795958708544, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.286803960800171, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.732953667640686, + "num_tokens": 493148749.0, + "step": 19057 + }, + { + "epoch": 2.0929057764111576, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.0477120876312256, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7213461399078369, + "num_tokens": 493187015.0, + "step": 19058 + }, + { + "epoch": 2.093015594113771, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.587372064590454, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7245916724205017, + "num_tokens": 493211688.0, + "step": 19059 + }, + { + "epoch": 2.0931254118163847, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2961952686309814, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7355225086212158, + "num_tokens": 493238035.0, + "step": 19060 + }, + { + "epoch": 2.0932352295189984, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2374045848846436, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7461098432540894, + "num_tokens": 493265714.0, + "step": 19061 + }, + { + "epoch": 2.093345047221612, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1953470706939697, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7507959604263306, + "num_tokens": 493294297.0, + "step": 19062 + }, + { + "epoch": 2.093454864924226, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4993793964385986, + "learning_rate": 1e-06, + "loss": 0.793, + "mean_token_accuracy": 0.7533696889877319, + "num_tokens": 493315677.0, + "step": 19063 + }, + { + "epoch": 2.0935646826268393, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.41207218170166, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7146005630493164, + "num_tokens": 493342213.0, + "step": 19064 + }, + { + "epoch": 2.093674500329453, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4983298778533936, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7306387424468994, + "num_tokens": 493365307.0, + "step": 19065 + }, + { + "epoch": 2.0937843180320668, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.111433982849121, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7207155227661133, + "num_tokens": 493396239.0, + "step": 19066 + }, + { + "epoch": 2.0938941357346805, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.854851722717285, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.731998085975647, + "num_tokens": 493416150.0, + "step": 19067 + }, + { + "epoch": 2.0940039534372943, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4461729526519775, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7104039192199707, + "num_tokens": 493440906.0, + "step": 19068 + }, + { + "epoch": 2.0941137711399076, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2689948081970215, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7452749609947205, + "num_tokens": 493467869.0, + "step": 19069 + }, + { + "epoch": 2.0942235888425214, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.212259531021118, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.723859965801239, + "num_tokens": 493496899.0, + "step": 19070 + }, + { + "epoch": 2.094333406545135, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.7049803733825684, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7430459260940552, + "num_tokens": 493518423.0, + "step": 19071 + }, + { + "epoch": 2.094443224247749, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3864495754241943, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.727949857711792, + "num_tokens": 493543772.0, + "step": 19072 + }, + { + "epoch": 2.0945530419503626, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.342465877532959, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7442789673805237, + "num_tokens": 493569302.0, + "step": 19073 + }, + { + "epoch": 2.094662859652976, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.7432467937469482, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7358970642089844, + "num_tokens": 493590965.0, + "step": 19074 + }, + { + "epoch": 2.0947726773555897, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3426854610443115, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7115668058395386, + "num_tokens": 493618059.0, + "step": 19075 + }, + { + "epoch": 2.0948824950582035, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3027729988098145, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7382901906967163, + "num_tokens": 493644737.0, + "step": 19076 + }, + { + "epoch": 2.094992312760817, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.378169298171997, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7178426384925842, + "num_tokens": 493670418.0, + "step": 19077 + }, + { + "epoch": 2.0951021304634305, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.0869579315185547, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7437958717346191, + "num_tokens": 493700392.0, + "step": 19078 + }, + { + "epoch": 2.0952119481660443, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1815502643585205, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7463229298591614, + "num_tokens": 493727488.0, + "step": 19079 + }, + { + "epoch": 2.095321765868658, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.430292844772339, + "learning_rate": 1e-06, + "loss": 0.7814, + "mean_token_accuracy": 0.7525944113731384, + "num_tokens": 493750634.0, + "step": 19080 + }, + { + "epoch": 2.095431583571272, + "ewc_loss": 0.0001678466796875, + "grad_norm": 3.213517665863037, + "learning_rate": 1e-06, + "loss": 0.7461, + "mean_token_accuracy": 0.7527985572814941, + "num_tokens": 493766122.0, + "step": 19081 + }, + { + "epoch": 2.0955414012738856, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2898802757263184, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7053981423377991, + "num_tokens": 493792108.0, + "step": 19082 + }, + { + "epoch": 2.095651218976499, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2699105739593506, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7330962419509888, + "num_tokens": 493818761.0, + "step": 19083 + }, + { + "epoch": 2.0957610366791126, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.509413242340088, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7409487962722778, + "num_tokens": 493842579.0, + "step": 19084 + }, + { + "epoch": 2.0958708543817264, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.505072832107544, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7248411178588867, + "num_tokens": 493866585.0, + "step": 19085 + }, + { + "epoch": 2.09598067208434, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.453141689300537, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7305623292922974, + "num_tokens": 493891549.0, + "step": 19086 + }, + { + "epoch": 2.0960904897869534, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3805713653564453, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7112065553665161, + "num_tokens": 493917048.0, + "step": 19087 + }, + { + "epoch": 2.096200307489567, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.765153408050537, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7313899397850037, + "num_tokens": 493936071.0, + "step": 19088 + }, + { + "epoch": 2.096310125192181, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.580075263977051, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7334998846054077, + "num_tokens": 493957816.0, + "step": 19089 + }, + { + "epoch": 2.0964199428947947, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2686350345611572, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7146222591400146, + "num_tokens": 493983395.0, + "step": 19090 + }, + { + "epoch": 2.0965297605974085, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3315606117248535, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7107568979263306, + "num_tokens": 494012211.0, + "step": 19091 + }, + { + "epoch": 2.096639578300022, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3074910640716553, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7198688983917236, + "num_tokens": 494038658.0, + "step": 19092 + }, + { + "epoch": 2.0967493960026355, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.375432252883911, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7324914336204529, + "num_tokens": 494064052.0, + "step": 19093 + }, + { + "epoch": 2.0968592137052493, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.393232822418213, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7331737279891968, + "num_tokens": 494089375.0, + "step": 19094 + }, + { + "epoch": 2.096969031407863, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.198819398880005, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7282209396362305, + "num_tokens": 494118483.0, + "step": 19095 + }, + { + "epoch": 2.097078849110477, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5519094467163086, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7112377285957336, + "num_tokens": 494141958.0, + "step": 19096 + }, + { + "epoch": 2.09718866681309, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1634607315063477, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.726325511932373, + "num_tokens": 494170703.0, + "step": 19097 + }, + { + "epoch": 2.097298484515704, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.488576889038086, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7197660207748413, + "num_tokens": 494194926.0, + "step": 19098 + }, + { + "epoch": 2.0974083022183176, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.358488082885742, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7505927681922913, + "num_tokens": 494220067.0, + "step": 19099 + }, + { + "epoch": 2.0975181199209314, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.268112897872925, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7435963153839111, + "num_tokens": 494246743.0, + "step": 19100 + }, + { + "epoch": 2.0976279376235447, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.630962371826172, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7221698760986328, + "num_tokens": 494267627.0, + "step": 19101 + }, + { + "epoch": 2.0977377553261585, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.082115888595581, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7207844257354736, + "num_tokens": 494298287.0, + "step": 19102 + }, + { + "epoch": 2.0978475730287722, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3412301540374756, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7248525023460388, + "num_tokens": 494323848.0, + "step": 19103 + }, + { + "epoch": 2.097957390731386, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4690845012664795, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7379553914070129, + "num_tokens": 494348007.0, + "step": 19104 + }, + { + "epoch": 2.0980672084339997, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5035171508789062, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7321630716323853, + "num_tokens": 494372682.0, + "step": 19105 + }, + { + "epoch": 2.098177026136613, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.279787540435791, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7442659735679626, + "num_tokens": 494398639.0, + "step": 19106 + }, + { + "epoch": 2.098286843839227, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.698385000228882, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7292548418045044, + "num_tokens": 494420579.0, + "step": 19107 + }, + { + "epoch": 2.0983966615418406, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5689308643341064, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7240689992904663, + "num_tokens": 494444564.0, + "step": 19108 + }, + { + "epoch": 2.0985064792444543, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.475898265838623, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.722730278968811, + "num_tokens": 494468730.0, + "step": 19109 + }, + { + "epoch": 2.0986162969470676, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.448225975036621, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7267199754714966, + "num_tokens": 494494472.0, + "step": 19110 + }, + { + "epoch": 2.0987261146496814, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.216634511947632, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7175832986831665, + "num_tokens": 494523174.0, + "step": 19111 + }, + { + "epoch": 2.098835932352295, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.548727512359619, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7355538010597229, + "num_tokens": 494545274.0, + "step": 19112 + }, + { + "epoch": 2.098945750054909, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.504699468612671, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.737153947353363, + "num_tokens": 494566831.0, + "step": 19113 + }, + { + "epoch": 2.0990555677575227, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.0590271949768066, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7225365042686462, + "num_tokens": 494598634.0, + "step": 19114 + }, + { + "epoch": 2.099165385460136, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3731753826141357, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7208214998245239, + "num_tokens": 494624358.0, + "step": 19115 + }, + { + "epoch": 2.0992752031627497, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2711634635925293, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7353041172027588, + "num_tokens": 494652414.0, + "step": 19116 + }, + { + "epoch": 2.0993850208653635, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.324605703353882, + "learning_rate": 1e-06, + "loss": 0.7863, + "mean_token_accuracy": 0.7489033341407776, + "num_tokens": 494677253.0, + "step": 19117 + }, + { + "epoch": 2.0994948385679773, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.6106526851654053, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7479865550994873, + "num_tokens": 494697568.0, + "step": 19118 + }, + { + "epoch": 2.099604656270591, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3153371810913086, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7291483879089355, + "num_tokens": 494723570.0, + "step": 19119 + }, + { + "epoch": 2.0997144739732043, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.6337108612060547, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7412907481193542, + "num_tokens": 494744675.0, + "step": 19120 + }, + { + "epoch": 2.099824291675818, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1693921089172363, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7305365204811096, + "num_tokens": 494771771.0, + "step": 19121 + }, + { + "epoch": 2.099934109378432, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5339131355285645, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7219892144203186, + "num_tokens": 494796506.0, + "step": 19122 + }, + { + "epoch": 2.1000439270810456, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.0560545921325684, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7391034364700317, + "num_tokens": 494829970.0, + "step": 19123 + }, + { + "epoch": 2.1001537447836593, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4183578491210938, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.719223141670227, + "num_tokens": 494856312.0, + "step": 19124 + }, + { + "epoch": 2.1002635624862727, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1698925495147705, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7433230876922607, + "num_tokens": 494887483.0, + "step": 19125 + }, + { + "epoch": 2.1003733801888864, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4726672172546387, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7383564710617065, + "num_tokens": 494910484.0, + "step": 19126 + }, + { + "epoch": 2.1004831978915, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.188343048095703, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7489377856254578, + "num_tokens": 494942311.0, + "step": 19127 + }, + { + "epoch": 2.100593015594114, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.651810884475708, + "learning_rate": 1e-06, + "loss": 0.7504, + "mean_token_accuracy": 0.7591357231140137, + "num_tokens": 494963926.0, + "step": 19128 + }, + { + "epoch": 2.1007028332967272, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.1446008682250977, + "learning_rate": 1e-06, + "loss": 0.7492, + "mean_token_accuracy": 0.7643975019454956, + "num_tokens": 494992019.0, + "step": 19129 + }, + { + "epoch": 2.100812650999341, + "ewc_loss": 0.0001678466796875, + "grad_norm": 1.9590747356414795, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7087419629096985, + "num_tokens": 495028204.0, + "step": 19130 + }, + { + "epoch": 2.1009224687019548, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.7644941806793213, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7404472231864929, + "num_tokens": 495048069.0, + "step": 19131 + }, + { + "epoch": 2.1010322864045685, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2102527618408203, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.7522950172424316, + "num_tokens": 495075347.0, + "step": 19132 + }, + { + "epoch": 2.1011421041071823, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.489713430404663, + "learning_rate": 1e-06, + "loss": 0.7801, + "mean_token_accuracy": 0.7565253376960754, + "num_tokens": 495099759.0, + "step": 19133 + }, + { + "epoch": 2.1012519218097956, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.241879463195801, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7280621528625488, + "num_tokens": 495128353.0, + "step": 19134 + }, + { + "epoch": 2.1013617395124093, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.775852918624878, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7473365068435669, + "num_tokens": 495148321.0, + "step": 19135 + }, + { + "epoch": 2.101471557215023, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4384028911590576, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7268891930580139, + "num_tokens": 495173397.0, + "step": 19136 + }, + { + "epoch": 2.101581374917637, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.446324586868286, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7178525328636169, + "num_tokens": 495197877.0, + "step": 19137 + }, + { + "epoch": 2.10169119262025, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.213106155395508, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7241120338439941, + "num_tokens": 495227587.0, + "step": 19138 + }, + { + "epoch": 2.101801010322864, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.328397512435913, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7237644195556641, + "num_tokens": 495253735.0, + "step": 19139 + }, + { + "epoch": 2.1019108280254777, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.534963607788086, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7342091202735901, + "num_tokens": 495275860.0, + "step": 19140 + }, + { + "epoch": 2.1020206457280914, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.668090343475342, + "learning_rate": 1e-06, + "loss": 0.7683, + "mean_token_accuracy": 0.7576262950897217, + "num_tokens": 495295552.0, + "step": 19141 + }, + { + "epoch": 2.102130463430705, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.549952983856201, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7361328601837158, + "num_tokens": 495318799.0, + "step": 19142 + }, + { + "epoch": 2.1022402811333185, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.6522889137268066, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.741671085357666, + "num_tokens": 495340919.0, + "step": 19143 + }, + { + "epoch": 2.1023500988359323, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.293532371520996, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7288038730621338, + "num_tokens": 495366914.0, + "step": 19144 + }, + { + "epoch": 2.102459916538546, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.885686159133911, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7388700842857361, + "num_tokens": 495384879.0, + "step": 19145 + }, + { + "epoch": 2.10256973424116, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.509246349334717, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7234576940536499, + "num_tokens": 495407837.0, + "step": 19146 + }, + { + "epoch": 2.1026795519437735, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.302527666091919, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7435532212257385, + "num_tokens": 495434520.0, + "step": 19147 + }, + { + "epoch": 2.102789369646387, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.92913818359375, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7457329034805298, + "num_tokens": 495452239.0, + "step": 19148 + }, + { + "epoch": 2.1028991873490006, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3298234939575195, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7258408069610596, + "num_tokens": 495478247.0, + "step": 19149 + }, + { + "epoch": 2.1030090050516144, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2669665813446045, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7198970317840576, + "num_tokens": 495507686.0, + "step": 19150 + }, + { + "epoch": 2.103118822754228, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.069178342819214, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7220646142959595, + "num_tokens": 495542033.0, + "step": 19151 + }, + { + "epoch": 2.1032286404568414, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.3078317642211914, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7527976632118225, + "num_tokens": 495568759.0, + "step": 19152 + }, + { + "epoch": 2.103338458159455, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4547767639160156, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.729634165763855, + "num_tokens": 495591660.0, + "step": 19153 + }, + { + "epoch": 2.103448275862069, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2419192790985107, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7426671981811523, + "num_tokens": 495618271.0, + "step": 19154 + }, + { + "epoch": 2.1035580935646827, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.529071569442749, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.716415524482727, + "num_tokens": 495642256.0, + "step": 19155 + }, + { + "epoch": 2.1036679112672965, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.614220142364502, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7568280696868896, + "num_tokens": 495662454.0, + "step": 19156 + }, + { + "epoch": 2.1037777289699098, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2570347785949707, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7212221026420593, + "num_tokens": 495690054.0, + "step": 19157 + }, + { + "epoch": 2.1038875466725235, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.4462153911590576, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7425469160079956, + "num_tokens": 495713834.0, + "step": 19158 + }, + { + "epoch": 2.1039973643751373, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.220243453979492, + "learning_rate": 1e-06, + "loss": 0.7606, + "mean_token_accuracy": 0.7563984990119934, + "num_tokens": 495740116.0, + "step": 19159 + }, + { + "epoch": 2.104107182077751, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5929388999938965, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7374800443649292, + "num_tokens": 495762199.0, + "step": 19160 + }, + { + "epoch": 2.104216999780365, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2771389484405518, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7444056868553162, + "num_tokens": 495787374.0, + "step": 19161 + }, + { + "epoch": 2.104326817482978, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.077279567718506, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7233889102935791, + "num_tokens": 495820269.0, + "step": 19162 + }, + { + "epoch": 2.104436635185592, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.2774147987365723, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.72760009765625, + "num_tokens": 495848752.0, + "step": 19163 + }, + { + "epoch": 2.1045464528882056, + "ewc_loss": 0.0001678466796875, + "grad_norm": 2.5652060508728027, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7285534143447876, + "num_tokens": 495871746.0, + "step": 19164 + }, + { + "epoch": 2.1046562705908194, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.655320882797241, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7198783159255981, + "num_tokens": 495894868.0, + "step": 19165 + }, + { + "epoch": 2.1047660882934327, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.195704221725464, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7151263356208801, + "num_tokens": 495923860.0, + "step": 19166 + }, + { + "epoch": 2.1048759059960465, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.0421435832977295, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7358646988868713, + "num_tokens": 495957197.0, + "step": 19167 + }, + { + "epoch": 2.10498572369866, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.278149366378784, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7195441722869873, + "num_tokens": 495985904.0, + "step": 19168 + }, + { + "epoch": 2.105095541401274, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.172677755355835, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.692104160785675, + "num_tokens": 496015293.0, + "step": 19169 + }, + { + "epoch": 2.1052053591038877, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.321406602859497, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7278591990470886, + "num_tokens": 496042205.0, + "step": 19170 + }, + { + "epoch": 2.105315176806501, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.3499960899353027, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7257906794548035, + "num_tokens": 496067972.0, + "step": 19171 + }, + { + "epoch": 2.105424994509115, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.4226937294006348, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7487431764602661, + "num_tokens": 496092607.0, + "step": 19172 + }, + { + "epoch": 2.1055348122117286, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.1890244483947754, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7274839282035828, + "num_tokens": 496122886.0, + "step": 19173 + }, + { + "epoch": 2.1056446299143423, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.5426979064941406, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7156656384468079, + "num_tokens": 496146924.0, + "step": 19174 + }, + { + "epoch": 2.105754447616956, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.600533962249756, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7370491027832031, + "num_tokens": 496166899.0, + "step": 19175 + }, + { + "epoch": 2.1058642653195694, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.246504783630371, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7190055847167969, + "num_tokens": 496196377.0, + "step": 19176 + }, + { + "epoch": 2.105974083022183, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.3237972259521484, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7255094647407532, + "num_tokens": 496223225.0, + "step": 19177 + }, + { + "epoch": 2.106083900724797, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.2962677478790283, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7149380445480347, + "num_tokens": 496250112.0, + "step": 19178 + }, + { + "epoch": 2.1061937184274107, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.070192813873291, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7405276298522949, + "num_tokens": 496279922.0, + "step": 19179 + }, + { + "epoch": 2.106303536130024, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.4834375381469727, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7117272019386292, + "num_tokens": 496304436.0, + "step": 19180 + }, + { + "epoch": 2.1064133538326377, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.171966791152954, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7195835113525391, + "num_tokens": 496334696.0, + "step": 19181 + }, + { + "epoch": 2.1065231715352515, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.5245532989501953, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7354437112808228, + "num_tokens": 496358059.0, + "step": 19182 + }, + { + "epoch": 2.1066329892378652, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.0727946758270264, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7248767614364624, + "num_tokens": 496387740.0, + "step": 19183 + }, + { + "epoch": 2.106742806940479, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.2547953128814697, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7185439467430115, + "num_tokens": 496415509.0, + "step": 19184 + }, + { + "epoch": 2.1068526246430923, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.727783441543579, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7314222455024719, + "num_tokens": 496437011.0, + "step": 19185 + }, + { + "epoch": 2.106962442345706, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.0466549396514893, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7227349877357483, + "num_tokens": 496469315.0, + "step": 19186 + }, + { + "epoch": 2.10707226004832, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.3947739601135254, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7221660614013672, + "num_tokens": 496496872.0, + "step": 19187 + }, + { + "epoch": 2.1071820777509336, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.2082571983337402, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7435107231140137, + "num_tokens": 496526371.0, + "step": 19188 + }, + { + "epoch": 2.107291895453547, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.3678667545318604, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7262144088745117, + "num_tokens": 496552103.0, + "step": 19189 + }, + { + "epoch": 2.1074017131561606, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.2726621627807617, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7187622785568237, + "num_tokens": 496579357.0, + "step": 19190 + }, + { + "epoch": 2.1075115308587744, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.4310638904571533, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7132619023323059, + "num_tokens": 496605039.0, + "step": 19191 + }, + { + "epoch": 2.107621348561388, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.9261691570281982, + "learning_rate": 1e-06, + "loss": 0.7477, + "mean_token_accuracy": 0.757597029209137, + "num_tokens": 496621766.0, + "step": 19192 + }, + { + "epoch": 2.107731166264002, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.2786359786987305, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7385607957839966, + "num_tokens": 496649137.0, + "step": 19193 + }, + { + "epoch": 2.1078409839666152, + "ewc_loss": 0.00016880035400390625, + "grad_norm": 2.381592273712158, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7331926822662354, + "num_tokens": 496675600.0, + "step": 19194 + }, + { + "epoch": 2.107950801669229, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1933367252349854, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7314586043357849, + "num_tokens": 496702492.0, + "step": 19195 + }, + { + "epoch": 2.1080606193718427, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.356375217437744, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7474057078361511, + "num_tokens": 496729085.0, + "step": 19196 + }, + { + "epoch": 2.1081704370744565, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.439460277557373, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7287089824676514, + "num_tokens": 496752094.0, + "step": 19197 + }, + { + "epoch": 2.1082802547770703, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.0663211345672607, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7022063732147217, + "num_tokens": 496785434.0, + "step": 19198 + }, + { + "epoch": 2.1083900724796836, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1702849864959717, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.718103289604187, + "num_tokens": 496814915.0, + "step": 19199 + }, + { + "epoch": 2.1084998901822973, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.132830858230591, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.736957311630249, + "num_tokens": 496842686.0, + "step": 19200 + }, + { + "epoch": 2.108609707884911, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.62556791305542, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7330421209335327, + "num_tokens": 496863892.0, + "step": 19201 + }, + { + "epoch": 2.108719525587525, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3980720043182373, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.733731746673584, + "num_tokens": 496886522.0, + "step": 19202 + }, + { + "epoch": 2.108829343290138, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4779884815216064, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7424963712692261, + "num_tokens": 496909244.0, + "step": 19203 + }, + { + "epoch": 2.108939160992752, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.487189769744873, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7212866544723511, + "num_tokens": 496935758.0, + "step": 19204 + }, + { + "epoch": 2.1090489786953657, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4185311794281006, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.737522542476654, + "num_tokens": 496960666.0, + "step": 19205 + }, + { + "epoch": 2.1091587963979794, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2158210277557373, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7259024381637573, + "num_tokens": 496987818.0, + "step": 19206 + }, + { + "epoch": 2.109268614100593, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3895363807678223, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7329129576683044, + "num_tokens": 497013415.0, + "step": 19207 + }, + { + "epoch": 2.1093784318032065, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6212568283081055, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7385462522506714, + "num_tokens": 497034957.0, + "step": 19208 + }, + { + "epoch": 2.1094882495058203, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.528865337371826, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7397316098213196, + "num_tokens": 497057299.0, + "step": 19209 + }, + { + "epoch": 2.109598067208434, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5698325634002686, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7226316928863525, + "num_tokens": 497083160.0, + "step": 19210 + }, + { + "epoch": 2.1097078849110478, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4080963134765625, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7274075746536255, + "num_tokens": 497110314.0, + "step": 19211 + }, + { + "epoch": 2.1098177026136615, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.424253225326538, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7483077645301819, + "num_tokens": 497133958.0, + "step": 19212 + }, + { + "epoch": 2.109927520316275, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4496963024139404, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7355719804763794, + "num_tokens": 497156841.0, + "step": 19213 + }, + { + "epoch": 2.1100373380188886, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4742722511291504, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7509976029396057, + "num_tokens": 497181270.0, + "step": 19214 + }, + { + "epoch": 2.1101471557215024, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3797554969787598, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7333529591560364, + "num_tokens": 497206080.0, + "step": 19215 + }, + { + "epoch": 2.110256973424116, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1644160747528076, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7175078988075256, + "num_tokens": 497236514.0, + "step": 19216 + }, + { + "epoch": 2.1103667911267294, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1730854511260986, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7179601788520813, + "num_tokens": 497265054.0, + "step": 19217 + }, + { + "epoch": 2.110476608829343, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6939377784729004, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7468170523643494, + "num_tokens": 497284782.0, + "step": 19218 + }, + { + "epoch": 2.110586426531957, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.109553575515747, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7324028015136719, + "num_tokens": 497315170.0, + "step": 19219 + }, + { + "epoch": 2.1106962442345707, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4063594341278076, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.715585470199585, + "num_tokens": 497342680.0, + "step": 19220 + }, + { + "epoch": 2.1108060619371845, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6574387550354004, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7195297479629517, + "num_tokens": 497365246.0, + "step": 19221 + }, + { + "epoch": 2.1109158796397978, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6460320949554443, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7224520444869995, + "num_tokens": 497386614.0, + "step": 19222 + }, + { + "epoch": 2.1110256973424115, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.576103925704956, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.733992338180542, + "num_tokens": 497408631.0, + "step": 19223 + }, + { + "epoch": 2.1111355150450253, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2937111854553223, + "learning_rate": 1e-06, + "loss": 0.7755, + "mean_token_accuracy": 0.7517480850219727, + "num_tokens": 497435262.0, + "step": 19224 + }, + { + "epoch": 2.111245332747639, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.268388271331787, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7340495586395264, + "num_tokens": 497462587.0, + "step": 19225 + }, + { + "epoch": 2.111355150450253, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.459746837615967, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7376260757446289, + "num_tokens": 497486300.0, + "step": 19226 + }, + { + "epoch": 2.111464968152866, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5482306480407715, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7416425347328186, + "num_tokens": 497507945.0, + "step": 19227 + }, + { + "epoch": 2.11157478585548, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1290974617004395, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7059115171432495, + "num_tokens": 497540400.0, + "step": 19228 + }, + { + "epoch": 2.1116846035580936, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6533596515655518, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7371805906295776, + "num_tokens": 497562841.0, + "step": 19229 + }, + { + "epoch": 2.1117944212607074, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.446507453918457, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7184381484985352, + "num_tokens": 497586063.0, + "step": 19230 + }, + { + "epoch": 2.1119042389633207, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3584513664245605, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.739828884601593, + "num_tokens": 497611524.0, + "step": 19231 + }, + { + "epoch": 2.1120140566659344, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.357097625732422, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7369375228881836, + "num_tokens": 497636834.0, + "step": 19232 + }, + { + "epoch": 2.112123874368548, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4515225887298584, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7378629446029663, + "num_tokens": 497662050.0, + "step": 19233 + }, + { + "epoch": 2.112233692071162, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.138214588165283, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7316160798072815, + "num_tokens": 497691458.0, + "step": 19234 + }, + { + "epoch": 2.1123435097737757, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.541740655899048, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7369039058685303, + "num_tokens": 497713683.0, + "step": 19235 + }, + { + "epoch": 2.112453327476389, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.385892391204834, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7301879525184631, + "num_tokens": 497740788.0, + "step": 19236 + }, + { + "epoch": 2.112563145179003, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2921769618988037, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7237817645072937, + "num_tokens": 497767673.0, + "step": 19237 + }, + { + "epoch": 2.1126729628816165, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.717505931854248, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7484682202339172, + "num_tokens": 497787761.0, + "step": 19238 + }, + { + "epoch": 2.1127827805842303, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2359540462493896, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7251543402671814, + "num_tokens": 497818542.0, + "step": 19239 + }, + { + "epoch": 2.1128925982868436, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.390467643737793, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7182537913322449, + "num_tokens": 497845783.0, + "step": 19240 + }, + { + "epoch": 2.1130024159894574, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1447927951812744, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7257633209228516, + "num_tokens": 497876838.0, + "step": 19241 + }, + { + "epoch": 2.113112233692071, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.164949417114258, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7196508049964905, + "num_tokens": 497906561.0, + "step": 19242 + }, + { + "epoch": 2.113222051394685, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2495903968811035, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7071704864501953, + "num_tokens": 497936046.0, + "step": 19243 + }, + { + "epoch": 2.1133318690972986, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5649352073669434, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7279213666915894, + "num_tokens": 497959831.0, + "step": 19244 + }, + { + "epoch": 2.113441686799912, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3962082862854004, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7317934036254883, + "num_tokens": 497986435.0, + "step": 19245 + }, + { + "epoch": 2.1135515045025257, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2541346549987793, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7226582169532776, + "num_tokens": 498016562.0, + "step": 19246 + }, + { + "epoch": 2.1136613222051395, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2977993488311768, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7404648065567017, + "num_tokens": 498042669.0, + "step": 19247 + }, + { + "epoch": 2.1137711399077532, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.828032970428467, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7264323830604553, + "num_tokens": 498062798.0, + "step": 19248 + }, + { + "epoch": 2.113880957610367, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.249602794647217, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7196027040481567, + "num_tokens": 498090075.0, + "step": 19249 + }, + { + "epoch": 2.1139907753129803, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.132819890975952, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7202979922294617, + "num_tokens": 498119591.0, + "step": 19250 + }, + { + "epoch": 2.114100593015594, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 1.9705184698104858, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7256143093109131, + "num_tokens": 498154592.0, + "step": 19251 + }, + { + "epoch": 2.114210410718208, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1441335678100586, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7156617641448975, + "num_tokens": 498185943.0, + "step": 19252 + }, + { + "epoch": 2.1143202284208216, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4345614910125732, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7184780836105347, + "num_tokens": 498212773.0, + "step": 19253 + }, + { + "epoch": 2.1144300461234353, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1582467555999756, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7356420755386353, + "num_tokens": 498240260.0, + "step": 19254 + }, + { + "epoch": 2.1145398638260486, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.476706027984619, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7305675745010376, + "num_tokens": 498264336.0, + "step": 19255 + }, + { + "epoch": 2.1146496815286624, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2290165424346924, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7206939458847046, + "num_tokens": 498293781.0, + "step": 19256 + }, + { + "epoch": 2.114759499231276, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2058751583099365, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7227936387062073, + "num_tokens": 498322343.0, + "step": 19257 + }, + { + "epoch": 2.11486931693389, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.254211664199829, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7417824268341064, + "num_tokens": 498352780.0, + "step": 19258 + }, + { + "epoch": 2.114979134636503, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2614245414733887, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7395495176315308, + "num_tokens": 498377441.0, + "step": 19259 + }, + { + "epoch": 2.115088952339117, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.750486373901367, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7308979034423828, + "num_tokens": 498397880.0, + "step": 19260 + }, + { + "epoch": 2.1151987700417307, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2304131984710693, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7168422341346741, + "num_tokens": 498426227.0, + "step": 19261 + }, + { + "epoch": 2.1153085877443445, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3691136837005615, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.720058798789978, + "num_tokens": 498453888.0, + "step": 19262 + }, + { + "epoch": 2.1154184054469582, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3841025829315186, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7290540933609009, + "num_tokens": 498482431.0, + "step": 19263 + }, + { + "epoch": 2.1155282231495716, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1638946533203125, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7153239250183105, + "num_tokens": 498517225.0, + "step": 19264 + }, + { + "epoch": 2.1156380408521853, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3478970527648926, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7352786064147949, + "num_tokens": 498542827.0, + "step": 19265 + }, + { + "epoch": 2.115747858554799, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.090191125869751, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6992141008377075, + "num_tokens": 498574854.0, + "step": 19266 + }, + { + "epoch": 2.115857676257413, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5129616260528564, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7276132702827454, + "num_tokens": 498600075.0, + "step": 19267 + }, + { + "epoch": 2.115967493960026, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2548961639404297, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7214280366897583, + "num_tokens": 498627775.0, + "step": 19268 + }, + { + "epoch": 2.11607731166264, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3635201454162598, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7247018814086914, + "num_tokens": 498652204.0, + "step": 19269 + }, + { + "epoch": 2.1161871293652537, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6980462074279785, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7441961169242859, + "num_tokens": 498672745.0, + "step": 19270 + }, + { + "epoch": 2.1162969470678674, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.515834093093872, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7356321811676025, + "num_tokens": 498695867.0, + "step": 19271 + }, + { + "epoch": 2.116406764770481, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.7795870304107666, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7331221699714661, + "num_tokens": 498715297.0, + "step": 19272 + }, + { + "epoch": 2.1165165824730945, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.416168689727783, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7305428981781006, + "num_tokens": 498739354.0, + "step": 19273 + }, + { + "epoch": 2.1166264001757082, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.407118320465088, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7448660135269165, + "num_tokens": 498762108.0, + "step": 19274 + }, + { + "epoch": 2.116736217878322, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.313039779663086, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7359887361526489, + "num_tokens": 498789026.0, + "step": 19275 + }, + { + "epoch": 2.1168460355809358, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.28895902633667, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7273175716400146, + "num_tokens": 498814879.0, + "step": 19276 + }, + { + "epoch": 2.1169558532835495, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5991275310516357, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7429099678993225, + "num_tokens": 498836829.0, + "step": 19277 + }, + { + "epoch": 2.117065670986163, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.525883913040161, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7249730825424194, + "num_tokens": 498861658.0, + "step": 19278 + }, + { + "epoch": 2.1171754886887766, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.9037601947784424, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7282696962356567, + "num_tokens": 498880637.0, + "step": 19279 + }, + { + "epoch": 2.1172853063913903, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.419269561767578, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7169994115829468, + "num_tokens": 498907284.0, + "step": 19280 + }, + { + "epoch": 2.117395124094004, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.291348457336426, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7292790412902832, + "num_tokens": 498936942.0, + "step": 19281 + }, + { + "epoch": 2.1175049417966174, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.091660499572754, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7506133317947388, + "num_tokens": 498966621.0, + "step": 19282 + }, + { + "epoch": 2.117614759499231, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4217872619628906, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7356657981872559, + "num_tokens": 498990140.0, + "step": 19283 + }, + { + "epoch": 2.117724577201845, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.629929542541504, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.730027437210083, + "num_tokens": 499011949.0, + "step": 19284 + }, + { + "epoch": 2.1178343949044587, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.252678155899048, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7333457469940186, + "num_tokens": 499038072.0, + "step": 19285 + }, + { + "epoch": 2.1179442126070724, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.558190107345581, + "learning_rate": 1e-06, + "loss": 0.7296, + "mean_token_accuracy": 0.7645730972290039, + "num_tokens": 499058058.0, + "step": 19286 + }, + { + "epoch": 2.1180540303096858, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1370620727539062, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7172874212265015, + "num_tokens": 499090487.0, + "step": 19287 + }, + { + "epoch": 2.1181638480122995, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3807713985443115, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7181606292724609, + "num_tokens": 499117531.0, + "step": 19288 + }, + { + "epoch": 2.1182736657149133, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.273609161376953, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7436158657073975, + "num_tokens": 499146062.0, + "step": 19289 + }, + { + "epoch": 2.118383483417527, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.223313093185425, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7116749882698059, + "num_tokens": 499178273.0, + "step": 19290 + }, + { + "epoch": 2.118493301120141, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.122053384780884, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7309509515762329, + "num_tokens": 499209831.0, + "step": 19291 + }, + { + "epoch": 2.118603118822754, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.303615093231201, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7233572006225586, + "num_tokens": 499236931.0, + "step": 19292 + }, + { + "epoch": 2.118712936525368, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.082963705062866, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7203561067581177, + "num_tokens": 499268058.0, + "step": 19293 + }, + { + "epoch": 2.1188227542279816, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3301055431365967, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7495342493057251, + "num_tokens": 499294168.0, + "step": 19294 + }, + { + "epoch": 2.1189325719305954, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.335364818572998, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7259089350700378, + "num_tokens": 499321024.0, + "step": 19295 + }, + { + "epoch": 2.1190423896332087, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1105809211730957, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7497034668922424, + "num_tokens": 499352640.0, + "step": 19296 + }, + { + "epoch": 2.1191522073358224, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.374096393585205, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7269256114959717, + "num_tokens": 499379641.0, + "step": 19297 + }, + { + "epoch": 2.119262025038436, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1446585655212402, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.720822274684906, + "num_tokens": 499411356.0, + "step": 19298 + }, + { + "epoch": 2.11937184274105, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.707195997238159, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7439817190170288, + "num_tokens": 499431949.0, + "step": 19299 + }, + { + "epoch": 2.1194816604436637, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.932310104370117, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7369405031204224, + "num_tokens": 499448969.0, + "step": 19300 + }, + { + "epoch": 2.119591478146277, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3515686988830566, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7293140888214111, + "num_tokens": 499474514.0, + "step": 19301 + }, + { + "epoch": 2.1197012958488908, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5004446506500244, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7349925637245178, + "num_tokens": 499497833.0, + "step": 19302 + }, + { + "epoch": 2.1198111135515045, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1795365810394287, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7421106100082397, + "num_tokens": 499527147.0, + "step": 19303 + }, + { + "epoch": 2.1199209312541183, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3700473308563232, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7215566635131836, + "num_tokens": 499553289.0, + "step": 19304 + }, + { + "epoch": 2.120030748956732, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5981178283691406, + "learning_rate": 1e-06, + "loss": 0.749, + "mean_token_accuracy": 0.7583242058753967, + "num_tokens": 499575098.0, + "step": 19305 + }, + { + "epoch": 2.1201405666593454, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.166794776916504, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7133288979530334, + "num_tokens": 499605573.0, + "step": 19306 + }, + { + "epoch": 2.120250384361959, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.414307117462158, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7327394485473633, + "num_tokens": 499630419.0, + "step": 19307 + }, + { + "epoch": 2.120360202064573, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3610498905181885, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7380088567733765, + "num_tokens": 499655863.0, + "step": 19308 + }, + { + "epoch": 2.1204700197671866, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.0343871116638184, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7142056226730347, + "num_tokens": 499689025.0, + "step": 19309 + }, + { + "epoch": 2.1205798374698, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1827776432037354, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.735088586807251, + "num_tokens": 499717713.0, + "step": 19310 + }, + { + "epoch": 2.1206896551724137, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.578385829925537, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7519903779029846, + "num_tokens": 499739851.0, + "step": 19311 + }, + { + "epoch": 2.1207994728750275, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.091514825820923, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7351912260055542, + "num_tokens": 499770583.0, + "step": 19312 + }, + { + "epoch": 2.120909290577641, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3059847354888916, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7488613128662109, + "num_tokens": 499798119.0, + "step": 19313 + }, + { + "epoch": 2.121019108280255, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.181724786758423, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7014834880828857, + "num_tokens": 499828541.0, + "step": 19314 + }, + { + "epoch": 2.1211289259828683, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5780344009399414, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7540270090103149, + "num_tokens": 499849585.0, + "step": 19315 + }, + { + "epoch": 2.121238743685482, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4251017570495605, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7407690286636353, + "num_tokens": 499871411.0, + "step": 19316 + }, + { + "epoch": 2.121348561388096, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.6064932346343994, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7283427119255066, + "num_tokens": 499895292.0, + "step": 19317 + }, + { + "epoch": 2.1214583790907096, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.225503444671631, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7199833393096924, + "num_tokens": 499925480.0, + "step": 19318 + }, + { + "epoch": 2.121568196793323, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.7486655712127686, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7236732244491577, + "num_tokens": 499948452.0, + "step": 19319 + }, + { + "epoch": 2.1216780144959366, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3900325298309326, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7361449003219604, + "num_tokens": 499972771.0, + "step": 19320 + }, + { + "epoch": 2.1217878321985504, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.734884262084961, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7361534833908081, + "num_tokens": 499993166.0, + "step": 19321 + }, + { + "epoch": 2.121897649901164, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.1812310218811035, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7426036596298218, + "num_tokens": 500024026.0, + "step": 19322 + }, + { + "epoch": 2.122007467603778, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.346114158630371, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7322862148284912, + "num_tokens": 500052901.0, + "step": 19323 + }, + { + "epoch": 2.122117285306391, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.469670534133911, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7274461388587952, + "num_tokens": 500076201.0, + "step": 19324 + }, + { + "epoch": 2.122227103009005, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2615370750427246, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7144694924354553, + "num_tokens": 500102261.0, + "step": 19325 + }, + { + "epoch": 2.1223369207116187, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.764956474304199, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.7527947425842285, + "num_tokens": 500121653.0, + "step": 19326 + }, + { + "epoch": 2.1224467384142325, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2371206283569336, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7345322370529175, + "num_tokens": 500150513.0, + "step": 19327 + }, + { + "epoch": 2.1225565561168462, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3425662517547607, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7242966890335083, + "num_tokens": 500176494.0, + "step": 19328 + }, + { + "epoch": 2.1226663738194596, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.696816921234131, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7503790259361267, + "num_tokens": 500197664.0, + "step": 19329 + }, + { + "epoch": 2.1227761915220733, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2913596630096436, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7162413001060486, + "num_tokens": 500225618.0, + "step": 19330 + }, + { + "epoch": 2.122886009224687, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.389716863632202, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7332065105438232, + "num_tokens": 500250309.0, + "step": 19331 + }, + { + "epoch": 2.122995826927301, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4490573406219482, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7389907836914062, + "num_tokens": 500274450.0, + "step": 19332 + }, + { + "epoch": 2.123105644629914, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.137322425842285, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7224903702735901, + "num_tokens": 500305517.0, + "step": 19333 + }, + { + "epoch": 2.123215462332528, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.387754201889038, + "learning_rate": 1e-06, + "loss": 0.7352, + "mean_token_accuracy": 0.7669209241867065, + "num_tokens": 500329715.0, + "step": 19334 + }, + { + "epoch": 2.1233252800351416, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4438412189483643, + "learning_rate": 1e-06, + "loss": 0.7881, + "mean_token_accuracy": 0.7466716170310974, + "num_tokens": 500352933.0, + "step": 19335 + }, + { + "epoch": 2.1234350977377554, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.502448081970215, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7294619083404541, + "num_tokens": 500374961.0, + "step": 19336 + }, + { + "epoch": 2.123544915440369, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2473559379577637, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7387992143630981, + "num_tokens": 500404202.0, + "step": 19337 + }, + { + "epoch": 2.1236547331429825, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.258831739425659, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7327320575714111, + "num_tokens": 500431137.0, + "step": 19338 + }, + { + "epoch": 2.1237645508455962, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3319742679595947, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.745063066482544, + "num_tokens": 500456704.0, + "step": 19339 + }, + { + "epoch": 2.12387436854821, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5975959300994873, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.739772617816925, + "num_tokens": 500479314.0, + "step": 19340 + }, + { + "epoch": 2.1239841862508237, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4175961017608643, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7286472320556641, + "num_tokens": 500502601.0, + "step": 19341 + }, + { + "epoch": 2.1240940039534375, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.61069393157959, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7208343744277954, + "num_tokens": 500525637.0, + "step": 19342 + }, + { + "epoch": 2.124203821656051, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2795708179473877, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7242189645767212, + "num_tokens": 500551665.0, + "step": 19343 + }, + { + "epoch": 2.1243136393586646, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3429346084594727, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7210239171981812, + "num_tokens": 500580210.0, + "step": 19344 + }, + { + "epoch": 2.1244234570612783, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3956692218780518, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7185366153717041, + "num_tokens": 500605288.0, + "step": 19345 + }, + { + "epoch": 2.124533274763892, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.0876047611236572, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7195701599121094, + "num_tokens": 500634790.0, + "step": 19346 + }, + { + "epoch": 2.1246430924665054, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.281848907470703, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7087653875350952, + "num_tokens": 500661686.0, + "step": 19347 + }, + { + "epoch": 2.124752910169119, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.7149057388305664, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7441263198852539, + "num_tokens": 500682856.0, + "step": 19348 + }, + { + "epoch": 2.124862727871733, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.151355504989624, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7220889329910278, + "num_tokens": 500713087.0, + "step": 19349 + }, + { + "epoch": 2.1249725455743467, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5471572875976562, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7349854707717896, + "num_tokens": 500736800.0, + "step": 19350 + }, + { + "epoch": 2.1250823632769604, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.475555896759033, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7434202432632446, + "num_tokens": 500760036.0, + "step": 19351 + }, + { + "epoch": 2.1251921809795737, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.472473621368408, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7320242524147034, + "num_tokens": 500786040.0, + "step": 19352 + }, + { + "epoch": 2.1253019986821875, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2719478607177734, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7308584451675415, + "num_tokens": 500813124.0, + "step": 19353 + }, + { + "epoch": 2.1254118163848013, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.4172892570495605, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7392467856407166, + "num_tokens": 500838213.0, + "step": 19354 + }, + { + "epoch": 2.125521634087415, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.428239345550537, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7364863753318787, + "num_tokens": 500863071.0, + "step": 19355 + }, + { + "epoch": 2.1256314517900288, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.3814122676849365, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7283092737197876, + "num_tokens": 500889884.0, + "step": 19356 + }, + { + "epoch": 2.125741269492642, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.255211114883423, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7487246990203857, + "num_tokens": 500917979.0, + "step": 19357 + }, + { + "epoch": 2.125851087195256, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.356022834777832, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7259984016418457, + "num_tokens": 500944795.0, + "step": 19358 + }, + { + "epoch": 2.1259609048978696, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2098023891448975, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7254314422607422, + "num_tokens": 500973310.0, + "step": 19359 + }, + { + "epoch": 2.1260707226004834, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.2207095623016357, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7330325245857239, + "num_tokens": 501001309.0, + "step": 19360 + }, + { + "epoch": 2.1261805403030967, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.21284818649292, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7127536535263062, + "num_tokens": 501031692.0, + "step": 19361 + }, + { + "epoch": 2.1262903580057104, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.53503155708313, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7315841913223267, + "num_tokens": 501055722.0, + "step": 19362 + }, + { + "epoch": 2.126400175708324, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.422916889190674, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7377275228500366, + "num_tokens": 501079954.0, + "step": 19363 + }, + { + "epoch": 2.126509993410938, + "ewc_loss": 0.00017070770263671875, + "grad_norm": 2.5276870727539062, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7398492693901062, + "num_tokens": 501102979.0, + "step": 19364 + }, + { + "epoch": 2.1266198111135517, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3222553730010986, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7338082790374756, + "num_tokens": 501128962.0, + "step": 19365 + }, + { + "epoch": 2.126729628816165, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.503805637359619, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7355135679244995, + "num_tokens": 501153358.0, + "step": 19366 + }, + { + "epoch": 2.1268394465187788, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.091445207595825, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7320820093154907, + "num_tokens": 501183138.0, + "step": 19367 + }, + { + "epoch": 2.1269492642213925, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.8019092082977295, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7372152805328369, + "num_tokens": 501202590.0, + "step": 19368 + }, + { + "epoch": 2.1270590819240063, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5652847290039062, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7431426048278809, + "num_tokens": 501224546.0, + "step": 19369 + }, + { + "epoch": 2.1271688996266196, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5355422496795654, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7442103028297424, + "num_tokens": 501246890.0, + "step": 19370 + }, + { + "epoch": 2.1272787173292333, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5188355445861816, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7421345710754395, + "num_tokens": 501271218.0, + "step": 19371 + }, + { + "epoch": 2.127388535031847, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2637784481048584, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7203026413917542, + "num_tokens": 501298334.0, + "step": 19372 + }, + { + "epoch": 2.127498352734461, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.6950604915618896, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7419862151145935, + "num_tokens": 501318225.0, + "step": 19373 + }, + { + "epoch": 2.1276081704370746, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.614607095718384, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7488062977790833, + "num_tokens": 501337679.0, + "step": 19374 + }, + { + "epoch": 2.127717988139688, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.6388823986053467, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7556548118591309, + "num_tokens": 501359800.0, + "step": 19375 + }, + { + "epoch": 2.1278278058423017, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3765759468078613, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7198827266693115, + "num_tokens": 501384640.0, + "step": 19376 + }, + { + "epoch": 2.1279376235449154, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3968026638031006, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7396645545959473, + "num_tokens": 501408346.0, + "step": 19377 + }, + { + "epoch": 2.128047441247529, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.386263132095337, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7374189496040344, + "num_tokens": 501434397.0, + "step": 19378 + }, + { + "epoch": 2.128157258950143, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5938029289245605, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7294625043869019, + "num_tokens": 501458928.0, + "step": 19379 + }, + { + "epoch": 2.1282670766527563, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5460948944091797, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7312734127044678, + "num_tokens": 501482911.0, + "step": 19380 + }, + { + "epoch": 2.12837689435537, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2899417877197266, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7284607887268066, + "num_tokens": 501511662.0, + "step": 19381 + }, + { + "epoch": 2.128486712057984, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.168264150619507, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7282840013504028, + "num_tokens": 501540540.0, + "step": 19382 + }, + { + "epoch": 2.1285965297605975, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2140347957611084, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7071694135665894, + "num_tokens": 501570794.0, + "step": 19383 + }, + { + "epoch": 2.1287063474632113, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3202438354492188, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7140322327613831, + "num_tokens": 501598342.0, + "step": 19384 + }, + { + "epoch": 2.1288161651658246, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.6860244274139404, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7480603456497192, + "num_tokens": 501617992.0, + "step": 19385 + }, + { + "epoch": 2.1289259828684384, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5265746116638184, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7260189056396484, + "num_tokens": 501640359.0, + "step": 19386 + }, + { + "epoch": 2.129035800571052, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3349928855895996, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.736215353012085, + "num_tokens": 501666252.0, + "step": 19387 + }, + { + "epoch": 2.129145618273666, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.1780691146850586, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7073314785957336, + "num_tokens": 501696625.0, + "step": 19388 + }, + { + "epoch": 2.129255435976279, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.42008113861084, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7077200412750244, + "num_tokens": 501721953.0, + "step": 19389 + }, + { + "epoch": 2.129365253678893, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.084804058074951, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7468639016151428, + "num_tokens": 501751758.0, + "step": 19390 + }, + { + "epoch": 2.1294750713815067, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.298569917678833, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.7563093900680542, + "num_tokens": 501777933.0, + "step": 19391 + }, + { + "epoch": 2.1295848890841205, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3045384883880615, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7040981650352478, + "num_tokens": 501805384.0, + "step": 19392 + }, + { + "epoch": 2.1296947067867342, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.8032984733581543, + "learning_rate": 1e-06, + "loss": 0.7722, + "mean_token_accuracy": 0.7528904676437378, + "num_tokens": 501825156.0, + "step": 19393 + }, + { + "epoch": 2.1298045244893475, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5024423599243164, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7546218633651733, + "num_tokens": 501847690.0, + "step": 19394 + }, + { + "epoch": 2.1299143421919613, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5843794345855713, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7368090152740479, + "num_tokens": 501871064.0, + "step": 19395 + }, + { + "epoch": 2.130024159894575, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2526328563690186, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.716688334941864, + "num_tokens": 501898573.0, + "step": 19396 + }, + { + "epoch": 2.130133977597189, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2760279178619385, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7299045324325562, + "num_tokens": 501926272.0, + "step": 19397 + }, + { + "epoch": 2.130243795299802, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2957160472869873, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7018283009529114, + "num_tokens": 501954207.0, + "step": 19398 + }, + { + "epoch": 2.130353613002416, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.7996678352355957, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.730140745639801, + "num_tokens": 501977226.0, + "step": 19399 + }, + { + "epoch": 2.1304634307050296, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3375086784362793, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7474889755249023, + "num_tokens": 502001304.0, + "step": 19400 + }, + { + "epoch": 2.1305732484076434, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5769197940826416, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7387241125106812, + "num_tokens": 502023302.0, + "step": 19401 + }, + { + "epoch": 2.130683066110257, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3224432468414307, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7324832677841187, + "num_tokens": 502048565.0, + "step": 19402 + }, + { + "epoch": 2.1307928838128705, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.566588878631592, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7218486666679382, + "num_tokens": 502071693.0, + "step": 19403 + }, + { + "epoch": 2.130902701515484, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.171660900115967, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7032487392425537, + "num_tokens": 502102292.0, + "step": 19404 + }, + { + "epoch": 2.131012519218098, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5820858478546143, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.744860827922821, + "num_tokens": 502124153.0, + "step": 19405 + }, + { + "epoch": 2.1311223369207117, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.388460636138916, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7456644773483276, + "num_tokens": 502149127.0, + "step": 19406 + }, + { + "epoch": 2.1312321546233255, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2914814949035645, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7404366135597229, + "num_tokens": 502175688.0, + "step": 19407 + }, + { + "epoch": 2.131341972325939, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.886326789855957, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.736685037612915, + "num_tokens": 502194729.0, + "step": 19408 + }, + { + "epoch": 2.1314517900285526, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2360880374908447, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7318880558013916, + "num_tokens": 502223616.0, + "step": 19409 + }, + { + "epoch": 2.1315616077311663, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.5832624435424805, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7355835437774658, + "num_tokens": 502246377.0, + "step": 19410 + }, + { + "epoch": 2.13167142543378, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3176355361938477, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7143256664276123, + "num_tokens": 502274348.0, + "step": 19411 + }, + { + "epoch": 2.1317812431363934, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.1580255031585693, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7384325265884399, + "num_tokens": 502304932.0, + "step": 19412 + }, + { + "epoch": 2.131891060839007, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.121039390563965, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7250785827636719, + "num_tokens": 502334295.0, + "step": 19413 + }, + { + "epoch": 2.132000878541621, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2459287643432617, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.731135368347168, + "num_tokens": 502361123.0, + "step": 19414 + }, + { + "epoch": 2.1321106962442347, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4289207458496094, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7407124638557434, + "num_tokens": 502385868.0, + "step": 19415 + }, + { + "epoch": 2.1322205139468484, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3487441539764404, + "learning_rate": 1e-06, + "loss": 0.7408, + "mean_token_accuracy": 0.7644071578979492, + "num_tokens": 502411193.0, + "step": 19416 + }, + { + "epoch": 2.1323303316494617, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4442689418792725, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7327659130096436, + "num_tokens": 502436624.0, + "step": 19417 + }, + { + "epoch": 2.1324401493520755, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3537259101867676, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7127158641815186, + "num_tokens": 502463703.0, + "step": 19418 + }, + { + "epoch": 2.1325499670546892, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3476405143737793, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7265398502349854, + "num_tokens": 502490220.0, + "step": 19419 + }, + { + "epoch": 2.132659784757303, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.0049664974212646, + "learning_rate": 1e-06, + "loss": 0.7725, + "mean_token_accuracy": 0.7585366368293762, + "num_tokens": 502520829.0, + "step": 19420 + }, + { + "epoch": 2.1327696024599163, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2283267974853516, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7164062857627869, + "num_tokens": 502548809.0, + "step": 19421 + }, + { + "epoch": 2.13287942016253, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.334712028503418, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7365782260894775, + "num_tokens": 502572556.0, + "step": 19422 + }, + { + "epoch": 2.132989237865144, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.247535228729248, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7189409732818604, + "num_tokens": 502601085.0, + "step": 19423 + }, + { + "epoch": 2.1330990555677576, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3790414333343506, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7278761863708496, + "num_tokens": 502626949.0, + "step": 19424 + }, + { + "epoch": 2.1332088732703713, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.6037261486053467, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7522276639938354, + "num_tokens": 502648130.0, + "step": 19425 + }, + { + "epoch": 2.1333186909729847, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.420236349105835, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7040061354637146, + "num_tokens": 502672859.0, + "step": 19426 + }, + { + "epoch": 2.1334285086755984, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.144298553466797, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7105820775032043, + "num_tokens": 502704382.0, + "step": 19427 + }, + { + "epoch": 2.133538326378212, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.2099461555480957, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7437976598739624, + "num_tokens": 502734070.0, + "step": 19428 + }, + { + "epoch": 2.133648144080826, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 3.0042731761932373, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7462842464447021, + "num_tokens": 502752470.0, + "step": 19429 + }, + { + "epoch": 2.1337579617834397, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4372634887695312, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7224923372268677, + "num_tokens": 502777691.0, + "step": 19430 + }, + { + "epoch": 2.133867779486053, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4888906478881836, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7245562076568604, + "num_tokens": 502802280.0, + "step": 19431 + }, + { + "epoch": 2.1339775971886668, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.046090841293335, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7219462394714355, + "num_tokens": 502836765.0, + "step": 19432 + }, + { + "epoch": 2.1340874148912805, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.147571325302124, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.710634171962738, + "num_tokens": 502865359.0, + "step": 19433 + }, + { + "epoch": 2.1341972325938943, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3597636222839355, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7282484769821167, + "num_tokens": 502890557.0, + "step": 19434 + }, + { + "epoch": 2.134307050296508, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.172508478164673, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7326239943504333, + "num_tokens": 502919872.0, + "step": 19435 + }, + { + "epoch": 2.1344168679991213, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.578655481338501, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7079290747642517, + "num_tokens": 502945931.0, + "step": 19436 + }, + { + "epoch": 2.134526685701735, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3060171604156494, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7285003662109375, + "num_tokens": 502973651.0, + "step": 19437 + }, + { + "epoch": 2.134636503404349, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4263088703155518, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7226347327232361, + "num_tokens": 502999893.0, + "step": 19438 + }, + { + "epoch": 2.1347463211069626, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.859426498413086, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7473658323287964, + "num_tokens": 503019470.0, + "step": 19439 + }, + { + "epoch": 2.134856138809576, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.4936635494232178, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7282156348228455, + "num_tokens": 503043003.0, + "step": 19440 + }, + { + "epoch": 2.1349659565121897, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1510703563690186, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7171183824539185, + "num_tokens": 503072935.0, + "step": 19441 + }, + { + "epoch": 2.1350757742148034, + "ewc_loss": 0.00017261505126953125, + "grad_norm": 2.3843588829040527, + "learning_rate": 1e-06, + "loss": 0.7439, + "mean_token_accuracy": 0.7595099210739136, + "num_tokens": 503097451.0, + "step": 19442 + }, + { + "epoch": 2.135185591917417, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4819631576538086, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7325823307037354, + "num_tokens": 503122261.0, + "step": 19443 + }, + { + "epoch": 2.135295409620031, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.233431816101074, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.715100109577179, + "num_tokens": 503152427.0, + "step": 19444 + }, + { + "epoch": 2.1354052273226443, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4499218463897705, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7118939161300659, + "num_tokens": 503177395.0, + "step": 19445 + }, + { + "epoch": 2.135515045025258, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.51033878326416, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7332863807678223, + "num_tokens": 503200906.0, + "step": 19446 + }, + { + "epoch": 2.1356248627278718, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3233511447906494, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7251182794570923, + "num_tokens": 503224808.0, + "step": 19447 + }, + { + "epoch": 2.1357346804304855, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2971904277801514, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7169235944747925, + "num_tokens": 503252145.0, + "step": 19448 + }, + { + "epoch": 2.135844498133099, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.300945520401001, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7465025782585144, + "num_tokens": 503280156.0, + "step": 19449 + }, + { + "epoch": 2.1359543158357126, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.443932294845581, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7517886757850647, + "num_tokens": 503305478.0, + "step": 19450 + }, + { + "epoch": 2.1360641335383264, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4326446056365967, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7238768935203552, + "num_tokens": 503328600.0, + "step": 19451 + }, + { + "epoch": 2.13617395124094, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6627228260040283, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.724469780921936, + "num_tokens": 503351610.0, + "step": 19452 + }, + { + "epoch": 2.136283768943554, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.8313546180725098, + "learning_rate": 1e-06, + "loss": 0.7644, + "mean_token_accuracy": 0.755748987197876, + "num_tokens": 503370742.0, + "step": 19453 + }, + { + "epoch": 2.136393586646167, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3322296142578125, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7365286350250244, + "num_tokens": 503397469.0, + "step": 19454 + }, + { + "epoch": 2.136503404348781, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.0070815086364746, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7366030216217041, + "num_tokens": 503431983.0, + "step": 19455 + }, + { + "epoch": 2.1366132220513947, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.275381088256836, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7399669289588928, + "num_tokens": 503460751.0, + "step": 19456 + }, + { + "epoch": 2.1367230397540085, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3844852447509766, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7293620109558105, + "num_tokens": 503485362.0, + "step": 19457 + }, + { + "epoch": 2.136832857456622, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4220478534698486, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7521460652351379, + "num_tokens": 503508116.0, + "step": 19458 + }, + { + "epoch": 2.1369426751592355, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.327773094177246, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7275999188423157, + "num_tokens": 503535310.0, + "step": 19459 + }, + { + "epoch": 2.1370524928618493, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.358096122741699, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6963920593261719, + "num_tokens": 503565033.0, + "step": 19460 + }, + { + "epoch": 2.137162310564463, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6470789909362793, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7535147666931152, + "num_tokens": 503585580.0, + "step": 19461 + }, + { + "epoch": 2.137272128267077, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.402024745941162, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7373536825180054, + "num_tokens": 503610320.0, + "step": 19462 + }, + { + "epoch": 2.13738194596969, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3345611095428467, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7392500042915344, + "num_tokens": 503636171.0, + "step": 19463 + }, + { + "epoch": 2.137491763672304, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.136906385421753, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7233911752700806, + "num_tokens": 503667294.0, + "step": 19464 + }, + { + "epoch": 2.1376015813749176, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.473379373550415, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7455753087997437, + "num_tokens": 503693320.0, + "step": 19465 + }, + { + "epoch": 2.1377113990775314, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.405719041824341, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7427369356155396, + "num_tokens": 503718333.0, + "step": 19466 + }, + { + "epoch": 2.137821216780145, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.039881467819214, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7344434857368469, + "num_tokens": 503750800.0, + "step": 19467 + }, + { + "epoch": 2.1379310344827585, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2937185764312744, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7488142251968384, + "num_tokens": 503776981.0, + "step": 19468 + }, + { + "epoch": 2.138040852185372, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1030471324920654, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.738358736038208, + "num_tokens": 503805743.0, + "step": 19469 + }, + { + "epoch": 2.138150669887986, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4823856353759766, + "learning_rate": 1e-06, + "loss": 0.7798, + "mean_token_accuracy": 0.7557199001312256, + "num_tokens": 503828450.0, + "step": 19470 + }, + { + "epoch": 2.1382604875905997, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.60036039352417, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7395053505897522, + "num_tokens": 503851585.0, + "step": 19471 + }, + { + "epoch": 2.138370305293213, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3392527103424072, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7391780614852905, + "num_tokens": 503877200.0, + "step": 19472 + }, + { + "epoch": 2.138480122995827, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6406280994415283, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7272865772247314, + "num_tokens": 503901070.0, + "step": 19473 + }, + { + "epoch": 2.1385899406984406, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4742188453674316, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7207782864570618, + "num_tokens": 503924246.0, + "step": 19474 + }, + { + "epoch": 2.1386997584010543, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1657915115356445, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7170094847679138, + "num_tokens": 503954041.0, + "step": 19475 + }, + { + "epoch": 2.138809576103668, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.217153787612915, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7327441573143005, + "num_tokens": 503981062.0, + "step": 19476 + }, + { + "epoch": 2.1389193938062814, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7148854732513428, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7352843284606934, + "num_tokens": 504002281.0, + "step": 19477 + }, + { + "epoch": 2.139029211508895, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1521034240722656, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7406386137008667, + "num_tokens": 504032604.0, + "step": 19478 + }, + { + "epoch": 2.139139029211509, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.150334358215332, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7199403643608093, + "num_tokens": 504061406.0, + "step": 19479 + }, + { + "epoch": 2.1392488469141226, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.587843418121338, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7400874495506287, + "num_tokens": 504082777.0, + "step": 19480 + }, + { + "epoch": 2.1393586646167364, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5097641944885254, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7246561050415039, + "num_tokens": 504107983.0, + "step": 19481 + }, + { + "epoch": 2.1394684823193497, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.211519718170166, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.7033413052558899, + "num_tokens": 504139751.0, + "step": 19482 + }, + { + "epoch": 2.1395783000219635, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4329583644866943, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7301186919212341, + "num_tokens": 504163631.0, + "step": 19483 + }, + { + "epoch": 2.1396881177245772, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4469079971313477, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7194843888282776, + "num_tokens": 504188728.0, + "step": 19484 + }, + { + "epoch": 2.139797935427191, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4098691940307617, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7264183759689331, + "num_tokens": 504214656.0, + "step": 19485 + }, + { + "epoch": 2.1399077531298047, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5721652507781982, + "learning_rate": 1e-06, + "loss": 0.7126, + "mean_token_accuracy": 0.7644040584564209, + "num_tokens": 504236576.0, + "step": 19486 + }, + { + "epoch": 2.140017570832418, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.254300355911255, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7359164953231812, + "num_tokens": 504263587.0, + "step": 19487 + }, + { + "epoch": 2.140127388535032, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.464402198791504, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7394105195999146, + "num_tokens": 504286875.0, + "step": 19488 + }, + { + "epoch": 2.1402372062376456, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3903348445892334, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7450997829437256, + "num_tokens": 504310774.0, + "step": 19489 + }, + { + "epoch": 2.1403470239402593, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.656137704849243, + "learning_rate": 1e-06, + "loss": 0.7725, + "mean_token_accuracy": 0.7531813383102417, + "num_tokens": 504332683.0, + "step": 19490 + }, + { + "epoch": 2.1404568416428726, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4060490131378174, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7010774612426758, + "num_tokens": 504359380.0, + "step": 19491 + }, + { + "epoch": 2.1405666593454864, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.205066442489624, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7296221852302551, + "num_tokens": 504389134.0, + "step": 19492 + }, + { + "epoch": 2.1406764770481, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.416264295578003, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7205427885055542, + "num_tokens": 504414439.0, + "step": 19493 + }, + { + "epoch": 2.140786294750714, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.601066827774048, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7207293510437012, + "num_tokens": 504437166.0, + "step": 19494 + }, + { + "epoch": 2.1408961124533277, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4415743350982666, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7310030460357666, + "num_tokens": 504460330.0, + "step": 19495 + }, + { + "epoch": 2.141005930155941, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.282963514328003, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7134393453598022, + "num_tokens": 504487957.0, + "step": 19496 + }, + { + "epoch": 2.1411157478585547, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.18208909034729, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.731572687625885, + "num_tokens": 504516567.0, + "step": 19497 + }, + { + "epoch": 2.1412255655611685, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.254159450531006, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7279103994369507, + "num_tokens": 504545127.0, + "step": 19498 + }, + { + "epoch": 2.1413353832637823, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.925464630126953, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7461334466934204, + "num_tokens": 504563497.0, + "step": 19499 + }, + { + "epoch": 2.1414452009663956, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4163081645965576, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.7471306324005127, + "num_tokens": 504587343.0, + "step": 19500 + }, + { + "epoch": 2.1415550186690093, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.091083288192749, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.72648686170578, + "num_tokens": 504620265.0, + "step": 19501 + }, + { + "epoch": 2.141664836371623, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.735766649246216, + "learning_rate": 1e-06, + "loss": 0.7142, + "mean_token_accuracy": 0.7681792974472046, + "num_tokens": 504638830.0, + "step": 19502 + }, + { + "epoch": 2.141774654074237, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5832419395446777, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7302901744842529, + "num_tokens": 504661914.0, + "step": 19503 + }, + { + "epoch": 2.1418844717768506, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2881412506103516, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7261776924133301, + "num_tokens": 504689335.0, + "step": 19504 + }, + { + "epoch": 2.141994289479464, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1294195652008057, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7265273332595825, + "num_tokens": 504718675.0, + "step": 19505 + }, + { + "epoch": 2.1421041071820777, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4962222576141357, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.741546630859375, + "num_tokens": 504741613.0, + "step": 19506 + }, + { + "epoch": 2.1422139248846914, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.270127058029175, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.73036128282547, + "num_tokens": 504769859.0, + "step": 19507 + }, + { + "epoch": 2.142323742587305, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2499442100524902, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7465313673019409, + "num_tokens": 504797627.0, + "step": 19508 + }, + { + "epoch": 2.142433560289919, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4613285064697266, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7450765371322632, + "num_tokens": 504820414.0, + "step": 19509 + }, + { + "epoch": 2.1425433779925322, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.389362096786499, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7266467809677124, + "num_tokens": 504844428.0, + "step": 19510 + }, + { + "epoch": 2.142653195695146, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.445059061050415, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7009844779968262, + "num_tokens": 504871808.0, + "step": 19511 + }, + { + "epoch": 2.1427630133977598, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4177286624908447, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7271644473075867, + "num_tokens": 504897352.0, + "step": 19512 + }, + { + "epoch": 2.1428728311003735, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.532827854156494, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7400418519973755, + "num_tokens": 504920529.0, + "step": 19513 + }, + { + "epoch": 2.142982648802987, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5478904247283936, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7348067760467529, + "num_tokens": 504943383.0, + "step": 19514 + }, + { + "epoch": 2.1430924665056006, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.242018222808838, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7318992018699646, + "num_tokens": 504972413.0, + "step": 19515 + }, + { + "epoch": 2.1432022842082143, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7433812618255615, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7374064326286316, + "num_tokens": 504994458.0, + "step": 19516 + }, + { + "epoch": 2.143312101910828, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3835270404815674, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7369124889373779, + "num_tokens": 505022774.0, + "step": 19517 + }, + { + "epoch": 2.143421919613442, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4816176891326904, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7300206422805786, + "num_tokens": 505046288.0, + "step": 19518 + }, + { + "epoch": 2.143531737316055, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3370962142944336, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7335019707679749, + "num_tokens": 505073053.0, + "step": 19519 + }, + { + "epoch": 2.143641555018669, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2576727867126465, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7271280288696289, + "num_tokens": 505101192.0, + "step": 19520 + }, + { + "epoch": 2.1437513727212827, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4232072830200195, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.726137638092041, + "num_tokens": 505127222.0, + "step": 19521 + }, + { + "epoch": 2.1438611904238964, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.270573854446411, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7012224793434143, + "num_tokens": 505155887.0, + "step": 19522 + }, + { + "epoch": 2.1439710081265098, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.310011386871338, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.724896252155304, + "num_tokens": 505184600.0, + "step": 19523 + }, + { + "epoch": 2.1440808258291235, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5812041759490967, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7220492362976074, + "num_tokens": 505208945.0, + "step": 19524 + }, + { + "epoch": 2.1441906435317373, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.601543426513672, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7564364671707153, + "num_tokens": 505232164.0, + "step": 19525 + }, + { + "epoch": 2.144300461234351, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2700531482696533, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7140436172485352, + "num_tokens": 505260252.0, + "step": 19526 + }, + { + "epoch": 2.144410278936965, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5025391578674316, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7472668290138245, + "num_tokens": 505282177.0, + "step": 19527 + }, + { + "epoch": 2.144520096639578, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5714519023895264, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7335919141769409, + "num_tokens": 505305190.0, + "step": 19528 + }, + { + "epoch": 2.144629914342192, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.499507427215576, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7394787669181824, + "num_tokens": 505329180.0, + "step": 19529 + }, + { + "epoch": 2.1447397320448056, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.134945869445801, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7414793372154236, + "num_tokens": 505358431.0, + "step": 19530 + }, + { + "epoch": 2.1448495497474194, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1759486198425293, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.716422438621521, + "num_tokens": 505388397.0, + "step": 19531 + }, + { + "epoch": 2.144959367450033, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2229220867156982, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7313774824142456, + "num_tokens": 505420726.0, + "step": 19532 + }, + { + "epoch": 2.1450691851526464, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3727383613586426, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7471087574958801, + "num_tokens": 505447288.0, + "step": 19533 + }, + { + "epoch": 2.14517900285526, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.0242342948913574, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7242021560668945, + "num_tokens": 505481821.0, + "step": 19534 + }, + { + "epoch": 2.145288820557874, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.175713062286377, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7257476449012756, + "num_tokens": 505510728.0, + "step": 19535 + }, + { + "epoch": 2.1453986382604877, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6126530170440674, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7171547412872314, + "num_tokens": 505533522.0, + "step": 19536 + }, + { + "epoch": 2.1455084559631015, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.252887725830078, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7126466035842896, + "num_tokens": 505561078.0, + "step": 19537 + }, + { + "epoch": 2.145618273665715, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.037275791168213, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7083538770675659, + "num_tokens": 505593858.0, + "step": 19538 + }, + { + "epoch": 2.1457280913683285, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.492983818054199, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7270105481147766, + "num_tokens": 505619141.0, + "step": 19539 + }, + { + "epoch": 2.1458379090709423, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.270134925842285, + "learning_rate": 1e-06, + "loss": 0.7921, + "mean_token_accuracy": 0.7518050670623779, + "num_tokens": 505645105.0, + "step": 19540 + }, + { + "epoch": 2.145947726773556, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2696404457092285, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7121039628982544, + "num_tokens": 505675618.0, + "step": 19541 + }, + { + "epoch": 2.1460575444761694, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.425069570541382, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7205216288566589, + "num_tokens": 505701316.0, + "step": 19542 + }, + { + "epoch": 2.146167362178783, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.271688938140869, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7195161581039429, + "num_tokens": 505728786.0, + "step": 19543 + }, + { + "epoch": 2.146277179881397, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3491477966308594, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7455006241798401, + "num_tokens": 505754729.0, + "step": 19544 + }, + { + "epoch": 2.1463869975840106, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1427249908447266, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7092968225479126, + "num_tokens": 505786922.0, + "step": 19545 + }, + { + "epoch": 2.1464968152866244, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2342031002044678, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7275010347366333, + "num_tokens": 505816400.0, + "step": 19546 + }, + { + "epoch": 2.1466066329892377, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.327272415161133, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7637948989868164, + "num_tokens": 505839740.0, + "step": 19547 + }, + { + "epoch": 2.1467164506918515, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2083630561828613, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7362024784088135, + "num_tokens": 505868143.0, + "step": 19548 + }, + { + "epoch": 2.146826268394465, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3771376609802246, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.724535346031189, + "num_tokens": 505894277.0, + "step": 19549 + }, + { + "epoch": 2.146936086097079, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4241721630096436, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7341727614402771, + "num_tokens": 505919504.0, + "step": 19550 + }, + { + "epoch": 2.1470459037996923, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.583359718322754, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7315713763237, + "num_tokens": 505942117.0, + "step": 19551 + }, + { + "epoch": 2.147155721502306, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.564768075942993, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7241945862770081, + "num_tokens": 505966337.0, + "step": 19552 + }, + { + "epoch": 2.14726553920492, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4954888820648193, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.71824049949646, + "num_tokens": 505991024.0, + "step": 19553 + }, + { + "epoch": 2.1473753569075336, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.20151948928833, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7260069251060486, + "num_tokens": 506018116.0, + "step": 19554 + }, + { + "epoch": 2.1474851746101473, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.357354164123535, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7252206206321716, + "num_tokens": 506042599.0, + "step": 19555 + }, + { + "epoch": 2.1475949923127606, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2240614891052246, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6961404085159302, + "num_tokens": 506071190.0, + "step": 19556 + }, + { + "epoch": 2.1477048100153744, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2240724563598633, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7342370748519897, + "num_tokens": 506101070.0, + "step": 19557 + }, + { + "epoch": 2.147814627717988, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4977762699127197, + "learning_rate": 1e-06, + "loss": 0.7573, + "mean_token_accuracy": 0.7547787427902222, + "num_tokens": 506123347.0, + "step": 19558 + }, + { + "epoch": 2.147924445420602, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5616087913513184, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7259331941604614, + "num_tokens": 506146118.0, + "step": 19559 + }, + { + "epoch": 2.1480342631232157, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.346839427947998, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7171931862831116, + "num_tokens": 506174054.0, + "step": 19560 + }, + { + "epoch": 2.148144080825829, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4651005268096924, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7434703707695007, + "num_tokens": 506197243.0, + "step": 19561 + }, + { + "epoch": 2.1482538985284427, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.0671885013580322, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7149421572685242, + "num_tokens": 506229973.0, + "step": 19562 + }, + { + "epoch": 2.1483637162310565, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.683742046356201, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.73515784740448, + "num_tokens": 506250745.0, + "step": 19563 + }, + { + "epoch": 2.1484735339336702, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.103825330734253, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7205677032470703, + "num_tokens": 506282315.0, + "step": 19564 + }, + { + "epoch": 2.148583351636284, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.051518201828003, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7188775539398193, + "num_tokens": 506316308.0, + "step": 19565 + }, + { + "epoch": 2.1486931693388973, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1950178146362305, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7222952842712402, + "num_tokens": 506345180.0, + "step": 19566 + }, + { + "epoch": 2.148802987041511, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.433828115463257, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7386796474456787, + "num_tokens": 506369076.0, + "step": 19567 + }, + { + "epoch": 2.148912804744125, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4602015018463135, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.717353105545044, + "num_tokens": 506393479.0, + "step": 19568 + }, + { + "epoch": 2.1490226224467386, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.335373878479004, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7323147654533386, + "num_tokens": 506419932.0, + "step": 19569 + }, + { + "epoch": 2.149132440149352, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2556118965148926, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7233423590660095, + "num_tokens": 506448953.0, + "step": 19570 + }, + { + "epoch": 2.1492422578519657, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1941754817962646, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7316504120826721, + "num_tokens": 506478305.0, + "step": 19571 + }, + { + "epoch": 2.1493520755545794, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2319648265838623, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7327452898025513, + "num_tokens": 506503441.0, + "step": 19572 + }, + { + "epoch": 2.149461893257193, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4732859134674072, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7157657146453857, + "num_tokens": 506527930.0, + "step": 19573 + }, + { + "epoch": 2.149571710959807, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.229595899581909, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7215006351470947, + "num_tokens": 506557797.0, + "step": 19574 + }, + { + "epoch": 2.1496815286624202, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2894277572631836, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7433954477310181, + "num_tokens": 506584377.0, + "step": 19575 + }, + { + "epoch": 2.149791346365034, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.990509510040283, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7320585250854492, + "num_tokens": 506603088.0, + "step": 19576 + }, + { + "epoch": 2.1499011640676478, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3868308067321777, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7136045694351196, + "num_tokens": 506630461.0, + "step": 19577 + }, + { + "epoch": 2.1500109817702615, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.356563091278076, + "learning_rate": 1e-06, + "loss": 0.7376, + "mean_token_accuracy": 0.7632365226745605, + "num_tokens": 506654651.0, + "step": 19578 + }, + { + "epoch": 2.150120799472875, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.110917091369629, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7111873626708984, + "num_tokens": 506685387.0, + "step": 19579 + }, + { + "epoch": 2.1502306171754886, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.246601104736328, + "learning_rate": 1e-06, + "loss": 0.7399, + "mean_token_accuracy": 0.7566526532173157, + "num_tokens": 506710343.0, + "step": 19580 + }, + { + "epoch": 2.1503404348781023, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2499916553497314, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7546621561050415, + "num_tokens": 506736132.0, + "step": 19581 + }, + { + "epoch": 2.150450252580716, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6448557376861572, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7317928075790405, + "num_tokens": 506757653.0, + "step": 19582 + }, + { + "epoch": 2.15056007028333, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7003068923950195, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7411668300628662, + "num_tokens": 506777449.0, + "step": 19583 + }, + { + "epoch": 2.150669887985943, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.320284843444824, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7125378251075745, + "num_tokens": 506804917.0, + "step": 19584 + }, + { + "epoch": 2.150779705688557, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2475388050079346, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7440958023071289, + "num_tokens": 506831845.0, + "step": 19585 + }, + { + "epoch": 2.1508895233911707, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4993162155151367, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7265857458114624, + "num_tokens": 506858898.0, + "step": 19586 + }, + { + "epoch": 2.1509993410937844, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6509037017822266, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7351652979850769, + "num_tokens": 506881919.0, + "step": 19587 + }, + { + "epoch": 2.151109158796398, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3652546405792236, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7423022985458374, + "num_tokens": 506907384.0, + "step": 19588 + }, + { + "epoch": 2.1512189764990115, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4815921783447266, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7212328910827637, + "num_tokens": 506931758.0, + "step": 19589 + }, + { + "epoch": 2.1513287942016253, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.665677547454834, + "learning_rate": 1e-06, + "loss": 0.7692, + "mean_token_accuracy": 0.7550114989280701, + "num_tokens": 506950804.0, + "step": 19590 + }, + { + "epoch": 2.151438611904239, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5917413234710693, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7355821132659912, + "num_tokens": 506973831.0, + "step": 19591 + }, + { + "epoch": 2.1515484296068528, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.312929391860962, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7104676961898804, + "num_tokens": 507002361.0, + "step": 19592 + }, + { + "epoch": 2.151658247309466, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1867337226867676, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7337455749511719, + "num_tokens": 507031774.0, + "step": 19593 + }, + { + "epoch": 2.15176806501208, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3469510078430176, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7082756161689758, + "num_tokens": 507056790.0, + "step": 19594 + }, + { + "epoch": 2.1518778827146936, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4266750812530518, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.719541072845459, + "num_tokens": 507082694.0, + "step": 19595 + }, + { + "epoch": 2.1519877004173074, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.850858688354492, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7447495460510254, + "num_tokens": 507102802.0, + "step": 19596 + }, + { + "epoch": 2.152097518119921, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3367245197296143, + "learning_rate": 1e-06, + "loss": 0.7793, + "mean_token_accuracy": 0.7524728775024414, + "num_tokens": 507126400.0, + "step": 19597 + }, + { + "epoch": 2.1522073358225344, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.289839506149292, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7377797961235046, + "num_tokens": 507153282.0, + "step": 19598 + }, + { + "epoch": 2.152317153525148, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7197043895721436, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7506433725357056, + "num_tokens": 507174702.0, + "step": 19599 + }, + { + "epoch": 2.152426971227762, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2363510131835938, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7183108329772949, + "num_tokens": 507202298.0, + "step": 19600 + }, + { + "epoch": 2.1525367889303757, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4164812564849854, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7154940366744995, + "num_tokens": 507227837.0, + "step": 19601 + }, + { + "epoch": 2.152646606632989, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.235677480697632, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7235555052757263, + "num_tokens": 507255427.0, + "step": 19602 + }, + { + "epoch": 2.1527564243356028, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.254399299621582, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7302820682525635, + "num_tokens": 507282890.0, + "step": 19603 + }, + { + "epoch": 2.1528662420382165, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.543111562728882, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7457232475280762, + "num_tokens": 507306909.0, + "step": 19604 + }, + { + "epoch": 2.1529760597408303, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 1.9880210161209106, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7465861439704895, + "num_tokens": 507338753.0, + "step": 19605 + }, + { + "epoch": 2.153085877443444, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.6078836917877197, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7198362350463867, + "num_tokens": 507360979.0, + "step": 19606 + }, + { + "epoch": 2.1531956951460574, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.101356029510498, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7339210510253906, + "num_tokens": 507390687.0, + "step": 19607 + }, + { + "epoch": 2.153305512848671, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3616371154785156, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.703353762626648, + "num_tokens": 507415851.0, + "step": 19608 + }, + { + "epoch": 2.153415330551285, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7894504070281982, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7333284616470337, + "num_tokens": 507436547.0, + "step": 19609 + }, + { + "epoch": 2.1535251482538986, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5894153118133545, + "learning_rate": 1e-06, + "loss": 0.7308, + "mean_token_accuracy": 0.768983781337738, + "num_tokens": 507456399.0, + "step": 19610 + }, + { + "epoch": 2.1536349659565124, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3363399505615234, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7082037329673767, + "num_tokens": 507483644.0, + "step": 19611 + }, + { + "epoch": 2.1537447836591257, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4343528747558594, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7409242391586304, + "num_tokens": 507507416.0, + "step": 19612 + }, + { + "epoch": 2.1538546013617395, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.312687873840332, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.6987955570220947, + "num_tokens": 507534168.0, + "step": 19613 + }, + { + "epoch": 2.153964419064353, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2601981163024902, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7345916032791138, + "num_tokens": 507560904.0, + "step": 19614 + }, + { + "epoch": 2.154074236766967, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.622767686843872, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7538057565689087, + "num_tokens": 507583246.0, + "step": 19615 + }, + { + "epoch": 2.1541840544695807, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.355139970779419, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7419295907020569, + "num_tokens": 507608366.0, + "step": 19616 + }, + { + "epoch": 2.154293872172194, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3243956565856934, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7211101651191711, + "num_tokens": 507637581.0, + "step": 19617 + }, + { + "epoch": 2.154403689874808, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.736156463623047, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7380311489105225, + "num_tokens": 507657453.0, + "step": 19618 + }, + { + "epoch": 2.1545135075774215, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.308817148208618, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7117224335670471, + "num_tokens": 507685092.0, + "step": 19619 + }, + { + "epoch": 2.1546233252800353, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3684792518615723, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7373021245002747, + "num_tokens": 507710248.0, + "step": 19620 + }, + { + "epoch": 2.1547331429826486, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4047329425811768, + "learning_rate": 1e-06, + "loss": 0.7607, + "mean_token_accuracy": 0.7582950592041016, + "num_tokens": 507735881.0, + "step": 19621 + }, + { + "epoch": 2.1548429606852624, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.389190912246704, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7216975688934326, + "num_tokens": 507762092.0, + "step": 19622 + }, + { + "epoch": 2.154952778387876, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.228130578994751, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7147190570831299, + "num_tokens": 507789685.0, + "step": 19623 + }, + { + "epoch": 2.15506259609049, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4968674182891846, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7141945362091064, + "num_tokens": 507815097.0, + "step": 19624 + }, + { + "epoch": 2.1551724137931036, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3901803493499756, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.737251877784729, + "num_tokens": 507839109.0, + "step": 19625 + }, + { + "epoch": 2.155282231495717, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4503848552703857, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7036342620849609, + "num_tokens": 507864532.0, + "step": 19626 + }, + { + "epoch": 2.1553920491983307, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.7296271324157715, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7311648726463318, + "num_tokens": 507886816.0, + "step": 19627 + }, + { + "epoch": 2.1555018669009445, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.380527973175049, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7454087734222412, + "num_tokens": 507910570.0, + "step": 19628 + }, + { + "epoch": 2.1556116846035582, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.644606590270996, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7517238259315491, + "num_tokens": 507930748.0, + "step": 19629 + }, + { + "epoch": 2.1557215023061715, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.542996883392334, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7201862335205078, + "num_tokens": 507954800.0, + "step": 19630 + }, + { + "epoch": 2.1558313200087853, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.803659677505493, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7392300963401794, + "num_tokens": 507974285.0, + "step": 19631 + }, + { + "epoch": 2.155941137711399, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.247459888458252, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7424052953720093, + "num_tokens": 508001138.0, + "step": 19632 + }, + { + "epoch": 2.156050955414013, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2890450954437256, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7338330745697021, + "num_tokens": 508027055.0, + "step": 19633 + }, + { + "epoch": 2.1561607731166266, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2507379055023193, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7142559289932251, + "num_tokens": 508055242.0, + "step": 19634 + }, + { + "epoch": 2.15627059081924, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2656517028808594, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7108232975006104, + "num_tokens": 508084001.0, + "step": 19635 + }, + { + "epoch": 2.1563804085218536, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2446937561035156, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.733905553817749, + "num_tokens": 508111763.0, + "step": 19636 + }, + { + "epoch": 2.1564902262244674, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3721487522125244, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7412787675857544, + "num_tokens": 508137475.0, + "step": 19637 + }, + { + "epoch": 2.156600043927081, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2481963634490967, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7281590104103088, + "num_tokens": 508165164.0, + "step": 19638 + }, + { + "epoch": 2.156709861629695, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.8934006690979004, + "learning_rate": 1e-06, + "loss": 0.7586, + "mean_token_accuracy": 0.7551496624946594, + "num_tokens": 508183366.0, + "step": 19639 + }, + { + "epoch": 2.1568196793323082, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.358611583709717, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7172516584396362, + "num_tokens": 508210922.0, + "step": 19640 + }, + { + "epoch": 2.156929497034922, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.475088596343994, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7253432869911194, + "num_tokens": 508233612.0, + "step": 19641 + }, + { + "epoch": 2.1570393147375357, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.202533721923828, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7168701887130737, + "num_tokens": 508262678.0, + "step": 19642 + }, + { + "epoch": 2.1571491324401495, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.291012763977051, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7190327048301697, + "num_tokens": 508290696.0, + "step": 19643 + }, + { + "epoch": 2.157258950142763, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.1338536739349365, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7413758039474487, + "num_tokens": 508321748.0, + "step": 19644 + }, + { + "epoch": 2.1573687678453766, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3918964862823486, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7331984043121338, + "num_tokens": 508345438.0, + "step": 19645 + }, + { + "epoch": 2.1574785855479903, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3177614212036133, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7168493270874023, + "num_tokens": 508372099.0, + "step": 19646 + }, + { + "epoch": 2.157588403250604, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.3487982749938965, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7353727221488953, + "num_tokens": 508398974.0, + "step": 19647 + }, + { + "epoch": 2.157698220953218, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.2648298740386963, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7261208891868591, + "num_tokens": 508428562.0, + "step": 19648 + }, + { + "epoch": 2.157808038655831, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.382894515991211, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7119132280349731, + "num_tokens": 508456487.0, + "step": 19649 + }, + { + "epoch": 2.157917856358445, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5161092281341553, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7391649484634399, + "num_tokens": 508481014.0, + "step": 19650 + }, + { + "epoch": 2.1580276740610587, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.4087507724761963, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7333296537399292, + "num_tokens": 508506398.0, + "step": 19651 + }, + { + "epoch": 2.1581374917636724, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.603153944015503, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7400245070457458, + "num_tokens": 508530440.0, + "step": 19652 + }, + { + "epoch": 2.1582473094662857, + "ewc_loss": 0.0001735687255859375, + "grad_norm": 2.5688588619232178, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7218765616416931, + "num_tokens": 508553552.0, + "step": 19653 + }, + { + "epoch": 2.1583571271688995, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.29953932762146, + "learning_rate": 1e-06, + "loss": 0.7935, + "mean_token_accuracy": 0.7469388246536255, + "num_tokens": 508579793.0, + "step": 19654 + }, + { + "epoch": 2.1584669448715132, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.017963409423828, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7152676582336426, + "num_tokens": 508614825.0, + "step": 19655 + }, + { + "epoch": 2.158576762574127, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.6401405334472656, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7466360926628113, + "num_tokens": 508636198.0, + "step": 19656 + }, + { + "epoch": 2.1586865802767408, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1880505084991455, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7230952978134155, + "num_tokens": 508665107.0, + "step": 19657 + }, + { + "epoch": 2.158796397979354, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2477149963378906, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7313486337661743, + "num_tokens": 508692624.0, + "step": 19658 + }, + { + "epoch": 2.158906215681968, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.42785382270813, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7348558902740479, + "num_tokens": 508716756.0, + "step": 19659 + }, + { + "epoch": 2.1590160333845816, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2899630069732666, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7208403944969177, + "num_tokens": 508745627.0, + "step": 19660 + }, + { + "epoch": 2.1591258510871953, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5202677249908447, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7168207168579102, + "num_tokens": 508770712.0, + "step": 19661 + }, + { + "epoch": 2.159235668789809, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1602227687835693, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.6999454498291016, + "num_tokens": 508800724.0, + "step": 19662 + }, + { + "epoch": 2.1593454864924224, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2910168170928955, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7342692613601685, + "num_tokens": 508826714.0, + "step": 19663 + }, + { + "epoch": 2.159455304195036, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3837549686431885, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7275256514549255, + "num_tokens": 508852337.0, + "step": 19664 + }, + { + "epoch": 2.15956512189765, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.8304386138916016, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7274657487869263, + "num_tokens": 508872900.0, + "step": 19665 + }, + { + "epoch": 2.1596749396002637, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.406764030456543, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7395257353782654, + "num_tokens": 508898461.0, + "step": 19666 + }, + { + "epoch": 2.1597847573028774, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2632577419281006, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7333566546440125, + "num_tokens": 508925970.0, + "step": 19667 + }, + { + "epoch": 2.1598945750054908, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1814093589782715, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7214324474334717, + "num_tokens": 508954639.0, + "step": 19668 + }, + { + "epoch": 2.1600043927081045, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.085055351257324, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7144602537155151, + "num_tokens": 508987944.0, + "step": 19669 + }, + { + "epoch": 2.1601142104107183, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.997603416442871, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7395557165145874, + "num_tokens": 509006625.0, + "step": 19670 + }, + { + "epoch": 2.160224028113332, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1473212242126465, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7224011421203613, + "num_tokens": 509036342.0, + "step": 19671 + }, + { + "epoch": 2.1603338458159453, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.214712381362915, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.716896653175354, + "num_tokens": 509064156.0, + "step": 19672 + }, + { + "epoch": 2.160443663518559, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.441527843475342, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.731468915939331, + "num_tokens": 509088825.0, + "step": 19673 + }, + { + "epoch": 2.160553481221173, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1141843795776367, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7227694988250732, + "num_tokens": 509120868.0, + "step": 19674 + }, + { + "epoch": 2.1606632989237866, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1560792922973633, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7081621289253235, + "num_tokens": 509150995.0, + "step": 19675 + }, + { + "epoch": 2.1607731166264004, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2589685916900635, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7454777956008911, + "num_tokens": 509177385.0, + "step": 19676 + }, + { + "epoch": 2.1608829343290137, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.520521640777588, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.69586181640625, + "num_tokens": 509203424.0, + "step": 19677 + }, + { + "epoch": 2.1609927520316274, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3499889373779297, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7169764637947083, + "num_tokens": 509228899.0, + "step": 19678 + }, + { + "epoch": 2.161102569734241, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5137217044830322, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7385356426239014, + "num_tokens": 509256290.0, + "step": 19679 + }, + { + "epoch": 2.161212387436855, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.289612054824829, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7307769656181335, + "num_tokens": 509283170.0, + "step": 19680 + }, + { + "epoch": 2.1613222051394683, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2410378456115723, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7172138094902039, + "num_tokens": 509312279.0, + "step": 19681 + }, + { + "epoch": 2.161432022842082, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3340389728546143, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7417619824409485, + "num_tokens": 509338350.0, + "step": 19682 + }, + { + "epoch": 2.161541840544696, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.180901050567627, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7369101047515869, + "num_tokens": 509366366.0, + "step": 19683 + }, + { + "epoch": 2.1616516582473095, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.7812421321868896, + "learning_rate": 1e-06, + "loss": 0.6898, + "mean_token_accuracy": 0.7785063982009888, + "num_tokens": 509383927.0, + "step": 19684 + }, + { + "epoch": 2.1617614759499233, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.404268503189087, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7157818078994751, + "num_tokens": 509411164.0, + "step": 19685 + }, + { + "epoch": 2.1618712936525366, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.486720323562622, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7197505235671997, + "num_tokens": 509436736.0, + "step": 19686 + }, + { + "epoch": 2.1619811113551504, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.474299669265747, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7078050374984741, + "num_tokens": 509462765.0, + "step": 19687 + }, + { + "epoch": 2.162090929057764, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5632898807525635, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7465100288391113, + "num_tokens": 509485486.0, + "step": 19688 + }, + { + "epoch": 2.162200746760378, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.385173797607422, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7540667057037354, + "num_tokens": 509511127.0, + "step": 19689 + }, + { + "epoch": 2.1623105644629916, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.316091299057007, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7414201498031616, + "num_tokens": 509536626.0, + "step": 19690 + }, + { + "epoch": 2.162420382165605, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.39550518989563, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7502884864807129, + "num_tokens": 509560467.0, + "step": 19691 + }, + { + "epoch": 2.1625301998682187, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.464729070663452, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7145241498947144, + "num_tokens": 509587425.0, + "step": 19692 + }, + { + "epoch": 2.1626400175708325, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.538221597671509, + "learning_rate": 1e-06, + "loss": 0.7879, + "mean_token_accuracy": 0.7432836890220642, + "num_tokens": 509610781.0, + "step": 19693 + }, + { + "epoch": 2.162749835273446, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5105185508728027, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7295696139335632, + "num_tokens": 509635032.0, + "step": 19694 + }, + { + "epoch": 2.16285965297606, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.238102436065674, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7259789705276489, + "num_tokens": 509663336.0, + "step": 19695 + }, + { + "epoch": 2.1629694706786733, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.683367967605591, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7106629014015198, + "num_tokens": 509687814.0, + "step": 19696 + }, + { + "epoch": 2.163079288381287, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3838694095611572, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7176735401153564, + "num_tokens": 509713456.0, + "step": 19697 + }, + { + "epoch": 2.163189106083901, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.7681543827056885, + "learning_rate": 1e-06, + "loss": 0.7647, + "mean_token_accuracy": 0.7550007104873657, + "num_tokens": 509735769.0, + "step": 19698 + }, + { + "epoch": 2.1632989237865146, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5017404556274414, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7235105633735657, + "num_tokens": 509761707.0, + "step": 19699 + }, + { + "epoch": 2.163408741489128, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.044081926345825, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7328306436538696, + "num_tokens": 509793233.0, + "step": 19700 + }, + { + "epoch": 2.1635185591917416, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1635894775390625, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7255971431732178, + "num_tokens": 509820851.0, + "step": 19701 + }, + { + "epoch": 2.1636283768943554, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1924571990966797, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7274677753448486, + "num_tokens": 509848138.0, + "step": 19702 + }, + { + "epoch": 2.163738194596969, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3619394302368164, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.723429262638092, + "num_tokens": 509875119.0, + "step": 19703 + }, + { + "epoch": 2.1638480122995825, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3684182167053223, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7298316359519958, + "num_tokens": 509902039.0, + "step": 19704 + }, + { + "epoch": 2.163957830002196, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.342653274536133, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7359617948532104, + "num_tokens": 509926147.0, + "step": 19705 + }, + { + "epoch": 2.16406764770481, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.7219860553741455, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.721538782119751, + "num_tokens": 509949592.0, + "step": 19706 + }, + { + "epoch": 2.1641774654074237, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.318929672241211, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7355331778526306, + "num_tokens": 509975177.0, + "step": 19707 + }, + { + "epoch": 2.1642872831100375, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.428542375564575, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7376574277877808, + "num_tokens": 509998908.0, + "step": 19708 + }, + { + "epoch": 2.164397100812651, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.4353084564208984, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7384250164031982, + "num_tokens": 510024977.0, + "step": 19709 + }, + { + "epoch": 2.1645069185152646, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.229705572128296, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7374458909034729, + "num_tokens": 510054957.0, + "step": 19710 + }, + { + "epoch": 2.1646167362178783, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3329415321350098, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7357662916183472, + "num_tokens": 510081631.0, + "step": 19711 + }, + { + "epoch": 2.164726553920492, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.376878023147583, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7416186332702637, + "num_tokens": 510107994.0, + "step": 19712 + }, + { + "epoch": 2.164836371623106, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.249188184738159, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7067958116531372, + "num_tokens": 510136020.0, + "step": 19713 + }, + { + "epoch": 2.164946189325719, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.4663474559783936, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7401867508888245, + "num_tokens": 510161525.0, + "step": 19714 + }, + { + "epoch": 2.165056007028333, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2870872020721436, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7132344245910645, + "num_tokens": 510188589.0, + "step": 19715 + }, + { + "epoch": 2.1651658247309467, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1606719493865967, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.734258770942688, + "num_tokens": 510216497.0, + "step": 19716 + }, + { + "epoch": 2.1652756424335604, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3379971981048584, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7379900217056274, + "num_tokens": 510240815.0, + "step": 19717 + }, + { + "epoch": 2.165385460136174, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5142087936401367, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7313008904457092, + "num_tokens": 510263707.0, + "step": 19718 + }, + { + "epoch": 2.1654952778387875, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.351026773452759, + "learning_rate": 1e-06, + "loss": 0.7427, + "mean_token_accuracy": 0.7604498863220215, + "num_tokens": 510287514.0, + "step": 19719 + }, + { + "epoch": 2.1656050955414012, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.1931092739105225, + "learning_rate": 1e-06, + "loss": 0.7651, + "mean_token_accuracy": 0.7504574656486511, + "num_tokens": 510315369.0, + "step": 19720 + }, + { + "epoch": 2.165714913244015, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5118215084075928, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7427034378051758, + "num_tokens": 510339585.0, + "step": 19721 + }, + { + "epoch": 2.1658247309466288, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.3221516609191895, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7131983041763306, + "num_tokens": 510367976.0, + "step": 19722 + }, + { + "epoch": 2.165934548649242, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.589935302734375, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7514036297798157, + "num_tokens": 510389909.0, + "step": 19723 + }, + { + "epoch": 2.166044366351856, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.529330015182495, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7229639291763306, + "num_tokens": 510412666.0, + "step": 19724 + }, + { + "epoch": 2.1661541840544696, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.797419548034668, + "learning_rate": 1e-06, + "loss": 0.7573, + "mean_token_accuracy": 0.7553360462188721, + "num_tokens": 510431793.0, + "step": 19725 + }, + { + "epoch": 2.1662640017570833, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.429105043411255, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7385865449905396, + "num_tokens": 510457084.0, + "step": 19726 + }, + { + "epoch": 2.166373819459697, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.370544672012329, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7285509705543518, + "num_tokens": 510483651.0, + "step": 19727 + }, + { + "epoch": 2.1664836371623104, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.5195841789245605, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7409929633140564, + "num_tokens": 510508228.0, + "step": 19728 + }, + { + "epoch": 2.166593454864924, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.2982630729675293, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7161519527435303, + "num_tokens": 510535082.0, + "step": 19729 + }, + { + "epoch": 2.166703272567538, + "ewc_loss": 0.00017452239990234375, + "grad_norm": 2.311495780944824, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.749312162399292, + "num_tokens": 510561521.0, + "step": 19730 + }, + { + "epoch": 2.1668130902701517, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4825403690338135, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7384501695632935, + "num_tokens": 510586160.0, + "step": 19731 + }, + { + "epoch": 2.166922907972765, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4750125408172607, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7352750301361084, + "num_tokens": 510611053.0, + "step": 19732 + }, + { + "epoch": 2.1670327256753787, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.332097053527832, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7208734750747681, + "num_tokens": 510639136.0, + "step": 19733 + }, + { + "epoch": 2.1671425433779925, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.270373582839966, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6984779238700867, + "num_tokens": 510667618.0, + "step": 19734 + }, + { + "epoch": 2.1672523610806063, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.567051410675049, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.7540056705474854, + "num_tokens": 510690341.0, + "step": 19735 + }, + { + "epoch": 2.16736217878322, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.21134614944458, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7336133718490601, + "num_tokens": 510717531.0, + "step": 19736 + }, + { + "epoch": 2.1674719964858333, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.026355504989624, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7014793157577515, + "num_tokens": 510750234.0, + "step": 19737 + }, + { + "epoch": 2.167581814188447, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2175188064575195, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7285193204879761, + "num_tokens": 510779786.0, + "step": 19738 + }, + { + "epoch": 2.167691631891061, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5427722930908203, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7394579648971558, + "num_tokens": 510802666.0, + "step": 19739 + }, + { + "epoch": 2.1678014495936746, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.47441029548645, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7322129607200623, + "num_tokens": 510825954.0, + "step": 19740 + }, + { + "epoch": 2.1679112672962884, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.6851720809936523, + "learning_rate": 1e-06, + "loss": 0.7618, + "mean_token_accuracy": 0.7594552636146545, + "num_tokens": 510847338.0, + "step": 19741 + }, + { + "epoch": 2.1680210849989017, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.481362819671631, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7174875736236572, + "num_tokens": 510873730.0, + "step": 19742 + }, + { + "epoch": 2.1681309027015154, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2593328952789307, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7139869332313538, + "num_tokens": 510903789.0, + "step": 19743 + }, + { + "epoch": 2.168240720404129, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2734363079071045, + "learning_rate": 1e-06, + "loss": 0.7072, + "mean_token_accuracy": 0.7755014896392822, + "num_tokens": 510929730.0, + "step": 19744 + }, + { + "epoch": 2.168350538106743, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4248244762420654, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7426081895828247, + "num_tokens": 510955077.0, + "step": 19745 + }, + { + "epoch": 2.1684603558093567, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4145021438598633, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7458037734031677, + "num_tokens": 510979271.0, + "step": 19746 + }, + { + "epoch": 2.16857017351197, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.029477834701538, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7125897407531738, + "num_tokens": 511012254.0, + "step": 19747 + }, + { + "epoch": 2.1686799912145838, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.7178072929382324, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7413380146026611, + "num_tokens": 511033036.0, + "step": 19748 + }, + { + "epoch": 2.1687898089171975, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3973071575164795, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7493312954902649, + "num_tokens": 511058387.0, + "step": 19749 + }, + { + "epoch": 2.1688996266198113, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2413837909698486, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7316392660140991, + "num_tokens": 511084828.0, + "step": 19750 + }, + { + "epoch": 2.1690094443224246, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.319599151611328, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7305542230606079, + "num_tokens": 511112041.0, + "step": 19751 + }, + { + "epoch": 2.1691192620250384, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.146291971206665, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.717093825340271, + "num_tokens": 511143885.0, + "step": 19752 + }, + { + "epoch": 2.169229079727652, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4485464096069336, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7244834899902344, + "num_tokens": 511168625.0, + "step": 19753 + }, + { + "epoch": 2.169338897430266, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3244128227233887, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7300028800964355, + "num_tokens": 511194843.0, + "step": 19754 + }, + { + "epoch": 2.1694487151328796, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.6477417945861816, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.723677396774292, + "num_tokens": 511216030.0, + "step": 19755 + }, + { + "epoch": 2.169558532835493, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.414332866668701, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.717628002166748, + "num_tokens": 511241673.0, + "step": 19756 + }, + { + "epoch": 2.1696683505381067, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4268839359283447, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7263762950897217, + "num_tokens": 511265301.0, + "step": 19757 + }, + { + "epoch": 2.1697781682407205, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.7305455207824707, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7379878759384155, + "num_tokens": 511284891.0, + "step": 19758 + }, + { + "epoch": 2.169887985943334, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.373335123062134, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7183817625045776, + "num_tokens": 511310615.0, + "step": 19759 + }, + { + "epoch": 2.1699978036459475, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1264114379882812, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7389850616455078, + "num_tokens": 511339053.0, + "step": 19760 + }, + { + "epoch": 2.1701076213485613, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.025191307067871, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7117944955825806, + "num_tokens": 511371657.0, + "step": 19761 + }, + { + "epoch": 2.170217439051175, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2475578784942627, + "learning_rate": 1e-06, + "loss": 0.799, + "mean_token_accuracy": 0.7460858225822449, + "num_tokens": 511400108.0, + "step": 19762 + }, + { + "epoch": 2.170327256753789, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4635813236236572, + "learning_rate": 1e-06, + "loss": 0.7423, + "mean_token_accuracy": 0.7578464150428772, + "num_tokens": 511422134.0, + "step": 19763 + }, + { + "epoch": 2.1704370744564025, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1398463249206543, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7322919964790344, + "num_tokens": 511452620.0, + "step": 19764 + }, + { + "epoch": 2.170546892159016, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4937024116516113, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7305354475975037, + "num_tokens": 511477119.0, + "step": 19765 + }, + { + "epoch": 2.1706567098616296, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1854376792907715, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7256644368171692, + "num_tokens": 511506474.0, + "step": 19766 + }, + { + "epoch": 2.1707665275642434, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5236568450927734, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7557864189147949, + "num_tokens": 511529501.0, + "step": 19767 + }, + { + "epoch": 2.170876345266857, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2893736362457275, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7398959398269653, + "num_tokens": 511558257.0, + "step": 19768 + }, + { + "epoch": 2.170986162969471, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3953933715820312, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.727739691734314, + "num_tokens": 511583088.0, + "step": 19769 + }, + { + "epoch": 2.171095980672084, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1829137802124023, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7243009805679321, + "num_tokens": 511612614.0, + "step": 19770 + }, + { + "epoch": 2.171205798374698, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.6183106899261475, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7425267100334167, + "num_tokens": 511633909.0, + "step": 19771 + }, + { + "epoch": 2.1713156160773117, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3345906734466553, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7242843508720398, + "num_tokens": 511659968.0, + "step": 19772 + }, + { + "epoch": 2.1714254337799255, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5813751220703125, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7417606115341187, + "num_tokens": 511682095.0, + "step": 19773 + }, + { + "epoch": 2.171535251482539, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.231708526611328, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7200429439544678, + "num_tokens": 511710990.0, + "step": 19774 + }, + { + "epoch": 2.1716450691851525, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3037009239196777, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7464432716369629, + "num_tokens": 511739624.0, + "step": 19775 + }, + { + "epoch": 2.1717548868877663, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.4280197620391846, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7118628025054932, + "num_tokens": 511764140.0, + "step": 19776 + }, + { + "epoch": 2.17186470459038, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2843565940856934, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7386810779571533, + "num_tokens": 511791977.0, + "step": 19777 + }, + { + "epoch": 2.171974522292994, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5327279567718506, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7338901162147522, + "num_tokens": 511815727.0, + "step": 19778 + }, + { + "epoch": 2.172084339995607, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.171741485595703, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7154597640037537, + "num_tokens": 511844926.0, + "step": 19779 + }, + { + "epoch": 2.172194157698221, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.0959644317626953, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.7433847784996033, + "num_tokens": 511877010.0, + "step": 19780 + }, + { + "epoch": 2.1723039754008346, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5918056964874268, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7348130941390991, + "num_tokens": 511897821.0, + "step": 19781 + }, + { + "epoch": 2.1724137931034484, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.333714485168457, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7235769033432007, + "num_tokens": 511925977.0, + "step": 19782 + }, + { + "epoch": 2.1725236108060617, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.385145425796509, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7371995449066162, + "num_tokens": 511949715.0, + "step": 19783 + }, + { + "epoch": 2.1726334285086755, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.300467014312744, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7520708441734314, + "num_tokens": 511977205.0, + "step": 19784 + }, + { + "epoch": 2.1727432462112892, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.369701385498047, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7355680465698242, + "num_tokens": 512002980.0, + "step": 19785 + }, + { + "epoch": 2.172853063913903, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5728156566619873, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7462130784988403, + "num_tokens": 512023265.0, + "step": 19786 + }, + { + "epoch": 2.1729628816165167, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.234567880630493, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7310282588005066, + "num_tokens": 512049915.0, + "step": 19787 + }, + { + "epoch": 2.17307269931913, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1677939891815186, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7322113513946533, + "num_tokens": 512078755.0, + "step": 19788 + }, + { + "epoch": 2.173182517021744, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.0796775817871094, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7409645318984985, + "num_tokens": 512109347.0, + "step": 19789 + }, + { + "epoch": 2.1732923347243576, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2150721549987793, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7379441857337952, + "num_tokens": 512137702.0, + "step": 19790 + }, + { + "epoch": 2.1734021524269713, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3287389278411865, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7339047193527222, + "num_tokens": 512163873.0, + "step": 19791 + }, + { + "epoch": 2.173511970129585, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5196518898010254, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.735502302646637, + "num_tokens": 512187004.0, + "step": 19792 + }, + { + "epoch": 2.1736217878321984, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.8911354541778564, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7472001314163208, + "num_tokens": 512206001.0, + "step": 19793 + }, + { + "epoch": 2.173731605534812, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.086578845977783, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.714762806892395, + "num_tokens": 512239300.0, + "step": 19794 + }, + { + "epoch": 2.173841423237426, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.185115337371826, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7383956909179688, + "num_tokens": 512267649.0, + "step": 19795 + }, + { + "epoch": 2.1739512409400397, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5901148319244385, + "learning_rate": 1e-06, + "loss": 0.7326, + "mean_token_accuracy": 0.7659696936607361, + "num_tokens": 512287428.0, + "step": 19796 + }, + { + "epoch": 2.1740610586426534, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3858301639556885, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7378160953521729, + "num_tokens": 512312435.0, + "step": 19797 + }, + { + "epoch": 2.1741708763452667, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3822383880615234, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7436227202415466, + "num_tokens": 512336657.0, + "step": 19798 + }, + { + "epoch": 2.1742806940478805, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3107893466949463, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7147396206855774, + "num_tokens": 512363372.0, + "step": 19799 + }, + { + "epoch": 2.1743905117504942, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.36525297164917, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7267324328422546, + "num_tokens": 512391054.0, + "step": 19800 + }, + { + "epoch": 2.174500329453108, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.614957332611084, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7405948042869568, + "num_tokens": 512413378.0, + "step": 19801 + }, + { + "epoch": 2.1746101471557213, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.556055784225464, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7356500029563904, + "num_tokens": 512436309.0, + "step": 19802 + }, + { + "epoch": 2.174719964858335, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.559842109680176, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7379382252693176, + "num_tokens": 512458612.0, + "step": 19803 + }, + { + "epoch": 2.174829782560949, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.47879958152771, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7158210873603821, + "num_tokens": 512483455.0, + "step": 19804 + }, + { + "epoch": 2.1749396002635626, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3424904346466064, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7302988171577454, + "num_tokens": 512509908.0, + "step": 19805 + }, + { + "epoch": 2.1750494179661763, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.487323045730591, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7229397296905518, + "num_tokens": 512534717.0, + "step": 19806 + }, + { + "epoch": 2.1751592356687897, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.565537929534912, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7436282634735107, + "num_tokens": 512557093.0, + "step": 19807 + }, + { + "epoch": 2.1752690533714034, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.341931104660034, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7142919898033142, + "num_tokens": 512584806.0, + "step": 19808 + }, + { + "epoch": 2.175378871074017, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.3371334075927734, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7103763818740845, + "num_tokens": 512610521.0, + "step": 19809 + }, + { + "epoch": 2.175488688776631, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.109408140182495, + "learning_rate": 1e-06, + "loss": 0.7703, + "mean_token_accuracy": 0.753717303276062, + "num_tokens": 512638067.0, + "step": 19810 + }, + { + "epoch": 2.1755985064792442, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.6192922592163086, + "learning_rate": 1e-06, + "loss": 0.7248, + "mean_token_accuracy": 0.7634685635566711, + "num_tokens": 512657400.0, + "step": 19811 + }, + { + "epoch": 2.175708324181858, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.391658067703247, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7403428554534912, + "num_tokens": 512683831.0, + "step": 19812 + }, + { + "epoch": 2.1758181418844718, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2929041385650635, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7311035990715027, + "num_tokens": 512711437.0, + "step": 19813 + }, + { + "epoch": 2.1759279595870855, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.2218706607818604, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7197527885437012, + "num_tokens": 512739807.0, + "step": 19814 + }, + { + "epoch": 2.1760377772896993, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.042426586151123, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7411840558052063, + "num_tokens": 512772090.0, + "step": 19815 + }, + { + "epoch": 2.1761475949923126, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5466952323913574, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7239284515380859, + "num_tokens": 512796315.0, + "step": 19816 + }, + { + "epoch": 2.1762574126949263, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.287099838256836, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7163605690002441, + "num_tokens": 512823980.0, + "step": 19817 + }, + { + "epoch": 2.17636723039754, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.1478424072265625, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7442684173583984, + "num_tokens": 512853809.0, + "step": 19818 + }, + { + "epoch": 2.176477048100154, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.5181264877319336, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7433140277862549, + "num_tokens": 512877992.0, + "step": 19819 + }, + { + "epoch": 2.1765868658027676, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.572880744934082, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.7462403774261475, + "num_tokens": 512900783.0, + "step": 19820 + }, + { + "epoch": 2.176696683505381, + "ewc_loss": 0.00017547607421875, + "grad_norm": 2.345818281173706, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7268144488334656, + "num_tokens": 512926616.0, + "step": 19821 + }, + { + "epoch": 2.1768065012079947, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4679505825042725, + "learning_rate": 1e-06, + "loss": 0.794, + "mean_token_accuracy": 0.7535113096237183, + "num_tokens": 512950669.0, + "step": 19822 + }, + { + "epoch": 2.1769163189106084, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4139933586120605, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7209776639938354, + "num_tokens": 512975107.0, + "step": 19823 + }, + { + "epoch": 2.177026136613222, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2965078353881836, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7457501888275146, + "num_tokens": 513001713.0, + "step": 19824 + }, + { + "epoch": 2.1771359543158355, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6179051399230957, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7127493619918823, + "num_tokens": 513026129.0, + "step": 19825 + }, + { + "epoch": 2.1772457720184493, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.436943769454956, + "learning_rate": 1e-06, + "loss": 0.7557, + "mean_token_accuracy": 0.7565644979476929, + "num_tokens": 513049616.0, + "step": 19826 + }, + { + "epoch": 2.177355589721063, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.427062749862671, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7221555709838867, + "num_tokens": 513076361.0, + "step": 19827 + }, + { + "epoch": 2.177465407423677, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3481593132019043, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7254785895347595, + "num_tokens": 513103893.0, + "step": 19828 + }, + { + "epoch": 2.1775752251262905, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6685941219329834, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7311485409736633, + "num_tokens": 513125256.0, + "step": 19829 + }, + { + "epoch": 2.177685042828904, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2765939235687256, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7053164839744568, + "num_tokens": 513152505.0, + "step": 19830 + }, + { + "epoch": 2.1777948605315176, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.0271902084350586, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7406737804412842, + "num_tokens": 513182326.0, + "step": 19831 + }, + { + "epoch": 2.1779046782341314, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.478590488433838, + "learning_rate": 1e-06, + "loss": 0.7824, + "mean_token_accuracy": 0.7604200839996338, + "num_tokens": 513204562.0, + "step": 19832 + }, + { + "epoch": 2.178014495936745, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.136822462081909, + "learning_rate": 1e-06, + "loss": 0.7635, + "mean_token_accuracy": 0.7575564384460449, + "num_tokens": 513231469.0, + "step": 19833 + }, + { + "epoch": 2.1781243136393584, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1488537788391113, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7387229204177856, + "num_tokens": 513260543.0, + "step": 19834 + }, + { + "epoch": 2.178234131341972, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5562920570373535, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7536928653717041, + "num_tokens": 513283046.0, + "step": 19835 + }, + { + "epoch": 2.178343949044586, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.416917562484741, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.6943023204803467, + "num_tokens": 513309369.0, + "step": 19836 + }, + { + "epoch": 2.1784537667471997, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3698041439056396, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.6975215077400208, + "num_tokens": 513337836.0, + "step": 19837 + }, + { + "epoch": 2.1785635844498135, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.0636987686157227, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.690510630607605, + "num_tokens": 513370115.0, + "step": 19838 + }, + { + "epoch": 2.1786734021524268, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.326432466506958, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7210032343864441, + "num_tokens": 513398324.0, + "step": 19839 + }, + { + "epoch": 2.1787832198550405, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.276702880859375, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7308800220489502, + "num_tokens": 513427101.0, + "step": 19840 + }, + { + "epoch": 2.1788930375576543, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.407339572906494, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7504104375839233, + "num_tokens": 513450935.0, + "step": 19841 + }, + { + "epoch": 2.179002855260268, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4234681129455566, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7276965975761414, + "num_tokens": 513476918.0, + "step": 19842 + }, + { + "epoch": 2.179112672962882, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.177560567855835, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.729821503162384, + "num_tokens": 513508803.0, + "step": 19843 + }, + { + "epoch": 2.179222490665495, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4964029788970947, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7278809547424316, + "num_tokens": 513533374.0, + "step": 19844 + }, + { + "epoch": 2.179332308368109, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5584652423858643, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7437116503715515, + "num_tokens": 513557068.0, + "step": 19845 + }, + { + "epoch": 2.1794421260707226, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4713382720947266, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7406470775604248, + "num_tokens": 513579110.0, + "step": 19846 + }, + { + "epoch": 2.1795519437733364, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.329118490219116, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.727180004119873, + "num_tokens": 513604626.0, + "step": 19847 + }, + { + "epoch": 2.17966176147595, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3104043006896973, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7164740562438965, + "num_tokens": 513630768.0, + "step": 19848 + }, + { + "epoch": 2.1797715791785635, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2612314224243164, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7443632483482361, + "num_tokens": 513656694.0, + "step": 19849 + }, + { + "epoch": 2.179881396881177, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2825934886932373, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7178412675857544, + "num_tokens": 513685769.0, + "step": 19850 + }, + { + "epoch": 2.179991214583791, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.05731201171875, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7053239345550537, + "num_tokens": 513719014.0, + "step": 19851 + }, + { + "epoch": 2.1801010322864047, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4995224475860596, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7226318120956421, + "num_tokens": 513745490.0, + "step": 19852 + }, + { + "epoch": 2.180210849989018, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.258485794067383, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7204709649085999, + "num_tokens": 513774431.0, + "step": 19853 + }, + { + "epoch": 2.180320667691632, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.320417881011963, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7309873104095459, + "num_tokens": 513799223.0, + "step": 19854 + }, + { + "epoch": 2.1804304853942456, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2542216777801514, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7228176593780518, + "num_tokens": 513826121.0, + "step": 19855 + }, + { + "epoch": 2.1805403030968593, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.0725414752960205, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7198693752288818, + "num_tokens": 513857479.0, + "step": 19856 + }, + { + "epoch": 2.180650120799473, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.697298288345337, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7206835746765137, + "num_tokens": 513878498.0, + "step": 19857 + }, + { + "epoch": 2.1807599385020864, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.718301773071289, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7353771924972534, + "num_tokens": 513899898.0, + "step": 19858 + }, + { + "epoch": 2.1808697562047, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1788077354431152, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7202719449996948, + "num_tokens": 513929478.0, + "step": 19859 + }, + { + "epoch": 2.180979573907314, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2230868339538574, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7258090376853943, + "num_tokens": 513956905.0, + "step": 19860 + }, + { + "epoch": 2.1810893916099277, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1964383125305176, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7440487146377563, + "num_tokens": 513984325.0, + "step": 19861 + }, + { + "epoch": 2.181199209312541, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.184197187423706, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7082329392433167, + "num_tokens": 514013885.0, + "step": 19862 + }, + { + "epoch": 2.1813090270151547, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1127216815948486, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7223060727119446, + "num_tokens": 514045712.0, + "step": 19863 + }, + { + "epoch": 2.1814188447177685, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6690673828125, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7339698076248169, + "num_tokens": 514067777.0, + "step": 19864 + }, + { + "epoch": 2.1815286624203822, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 3.064314365386963, + "learning_rate": 1e-06, + "loss": 0.7472, + "mean_token_accuracy": 0.7572256326675415, + "num_tokens": 514084239.0, + "step": 19865 + }, + { + "epoch": 2.181638480122996, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3658199310302734, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7139883637428284, + "num_tokens": 514111071.0, + "step": 19866 + }, + { + "epoch": 2.1817482978256093, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2652909755706787, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7465750575065613, + "num_tokens": 514139515.0, + "step": 19867 + }, + { + "epoch": 2.181858115528223, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.346630334854126, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7166721820831299, + "num_tokens": 514166181.0, + "step": 19868 + }, + { + "epoch": 2.181967933230837, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.187568426132202, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.7428942918777466, + "num_tokens": 514194972.0, + "step": 19869 + }, + { + "epoch": 2.1820777509334506, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.231497287750244, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7195609211921692, + "num_tokens": 514224534.0, + "step": 19870 + }, + { + "epoch": 2.1821875686360643, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.089989185333252, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7417490482330322, + "num_tokens": 514256824.0, + "step": 19871 + }, + { + "epoch": 2.1822973863386776, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1990244388580322, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7386331558227539, + "num_tokens": 514284891.0, + "step": 19872 + }, + { + "epoch": 2.1824072040412914, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6521174907684326, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7467932105064392, + "num_tokens": 514306297.0, + "step": 19873 + }, + { + "epoch": 2.182517021743905, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3765478134155273, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7403531074523926, + "num_tokens": 514331575.0, + "step": 19874 + }, + { + "epoch": 2.182626839446519, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4973366260528564, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.729227602481842, + "num_tokens": 514355046.0, + "step": 19875 + }, + { + "epoch": 2.1827366571491327, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.324960470199585, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7315211296081543, + "num_tokens": 514381567.0, + "step": 19876 + }, + { + "epoch": 2.182846474851746, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4496500492095947, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7422934770584106, + "num_tokens": 514404830.0, + "step": 19877 + }, + { + "epoch": 2.1829562925543597, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.11246395111084, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7369434237480164, + "num_tokens": 514434722.0, + "step": 19878 + }, + { + "epoch": 2.1830661102569735, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.551820755004883, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7215596437454224, + "num_tokens": 514457611.0, + "step": 19879 + }, + { + "epoch": 2.1831759279595873, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.8463048934936523, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7457796335220337, + "num_tokens": 514477830.0, + "step": 19880 + }, + { + "epoch": 2.1832857456622006, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2821695804595947, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7418509721755981, + "num_tokens": 514503871.0, + "step": 19881 + }, + { + "epoch": 2.1833955633648143, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6532561779022217, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7312674522399902, + "num_tokens": 514525892.0, + "step": 19882 + }, + { + "epoch": 2.183505381067428, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.115699052810669, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7201505303382874, + "num_tokens": 514557113.0, + "step": 19883 + }, + { + "epoch": 2.183615198770042, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.118326425552368, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7353585958480835, + "num_tokens": 514588219.0, + "step": 19884 + }, + { + "epoch": 2.1837250164726556, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2209126949310303, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7406734824180603, + "num_tokens": 514615084.0, + "step": 19885 + }, + { + "epoch": 2.183834834175269, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3326354026794434, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7350024580955505, + "num_tokens": 514639856.0, + "step": 19886 + }, + { + "epoch": 2.1839446518778827, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.903433322906494, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7469197511672974, + "num_tokens": 514659269.0, + "step": 19887 + }, + { + "epoch": 2.1840544695804964, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5041344165802, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7307224273681641, + "num_tokens": 514682620.0, + "step": 19888 + }, + { + "epoch": 2.18416428728311, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.693315267562866, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7192403078079224, + "num_tokens": 514705976.0, + "step": 19889 + }, + { + "epoch": 2.1842741049857235, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.589067220687866, + "learning_rate": 1e-06, + "loss": 0.7595, + "mean_token_accuracy": 0.7555250525474548, + "num_tokens": 514727483.0, + "step": 19890 + }, + { + "epoch": 2.1843839226883373, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.547694444656372, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7276169657707214, + "num_tokens": 514750353.0, + "step": 19891 + }, + { + "epoch": 2.184493740390951, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.246447801589966, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7211732268333435, + "num_tokens": 514776854.0, + "step": 19892 + }, + { + "epoch": 2.1846035580935648, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.057450771331787, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.731923520565033, + "num_tokens": 514809145.0, + "step": 19893 + }, + { + "epoch": 2.1847133757961785, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.198047161102295, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7278186082839966, + "num_tokens": 514837614.0, + "step": 19894 + }, + { + "epoch": 2.184823193498792, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.408726692199707, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7021001577377319, + "num_tokens": 514863122.0, + "step": 19895 + }, + { + "epoch": 2.1849330112014056, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.557156562805176, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7464630603790283, + "num_tokens": 514884835.0, + "step": 19896 + }, + { + "epoch": 2.1850428289040194, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.378464937210083, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7376312017440796, + "num_tokens": 514910027.0, + "step": 19897 + }, + { + "epoch": 2.185152646606633, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.404660940170288, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7577839493751526, + "num_tokens": 514934175.0, + "step": 19898 + }, + { + "epoch": 2.185262464309247, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.595611572265625, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7271048426628113, + "num_tokens": 514956012.0, + "step": 19899 + }, + { + "epoch": 2.18537228201186, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.341130256652832, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.733330488204956, + "num_tokens": 514982460.0, + "step": 19900 + }, + { + "epoch": 2.185482099714474, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.573719024658203, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7482198476791382, + "num_tokens": 515005156.0, + "step": 19901 + }, + { + "epoch": 2.1855919174170877, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.439154863357544, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7248430252075195, + "num_tokens": 515030699.0, + "step": 19902 + }, + { + "epoch": 2.1857017351197014, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4094889163970947, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7293946743011475, + "num_tokens": 515055876.0, + "step": 19903 + }, + { + "epoch": 2.1858115528223148, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.208212375640869, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7143542766571045, + "num_tokens": 515086888.0, + "step": 19904 + }, + { + "epoch": 2.1859213705249285, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.253777503967285, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7374860048294067, + "num_tokens": 515115200.0, + "step": 19905 + }, + { + "epoch": 2.1860311882275423, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.291672945022583, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7309548854827881, + "num_tokens": 515141586.0, + "step": 19906 + }, + { + "epoch": 2.186141005930156, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3816165924072266, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.733887791633606, + "num_tokens": 515167298.0, + "step": 19907 + }, + { + "epoch": 2.18625082363277, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6226675510406494, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7354390025138855, + "num_tokens": 515190325.0, + "step": 19908 + }, + { + "epoch": 2.186360641335383, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.396815538406372, + "learning_rate": 1e-06, + "loss": 0.7921, + "mean_token_accuracy": 0.7452983856201172, + "num_tokens": 515215158.0, + "step": 19909 + }, + { + "epoch": 2.186470459037997, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.158726930618286, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7430218458175659, + "num_tokens": 515244882.0, + "step": 19910 + }, + { + "epoch": 2.1865802767406106, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.687824249267578, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7384392023086548, + "num_tokens": 515265951.0, + "step": 19911 + }, + { + "epoch": 2.1866900944432244, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6307666301727295, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7259918451309204, + "num_tokens": 515287808.0, + "step": 19912 + }, + { + "epoch": 2.1867999121458377, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.523033380508423, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7359154224395752, + "num_tokens": 515312532.0, + "step": 19913 + }, + { + "epoch": 2.1869097298484514, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.9959030151367188, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7472264766693115, + "num_tokens": 515330522.0, + "step": 19914 + }, + { + "epoch": 2.187019547551065, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.6252474784851074, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7086696624755859, + "num_tokens": 515353629.0, + "step": 19915 + }, + { + "epoch": 2.187129365253679, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.561856508255005, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.73896723985672, + "num_tokens": 515377622.0, + "step": 19916 + }, + { + "epoch": 2.1872391829562927, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4891159534454346, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7262212038040161, + "num_tokens": 515402491.0, + "step": 19917 + }, + { + "epoch": 2.187349000658906, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.519516944885254, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7537680268287659, + "num_tokens": 515427669.0, + "step": 19918 + }, + { + "epoch": 2.18745881836152, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3050642013549805, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.703892707824707, + "num_tokens": 515455641.0, + "step": 19919 + }, + { + "epoch": 2.1875686360641335, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.485821485519409, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7450107336044312, + "num_tokens": 515480433.0, + "step": 19920 + }, + { + "epoch": 2.1876784537667473, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.33101487159729, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7319351434707642, + "num_tokens": 515506270.0, + "step": 19921 + }, + { + "epoch": 2.187788271469361, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4602162837982178, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7324877977371216, + "num_tokens": 515532306.0, + "step": 19922 + }, + { + "epoch": 2.1878980891719744, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4044442176818848, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7178115844726562, + "num_tokens": 515556984.0, + "step": 19923 + }, + { + "epoch": 2.188007906874588, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4587738513946533, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7297267913818359, + "num_tokens": 515581961.0, + "step": 19924 + }, + { + "epoch": 2.188117724577202, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.906435012817383, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7392640113830566, + "num_tokens": 515599473.0, + "step": 19925 + }, + { + "epoch": 2.1882275422798156, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3727872371673584, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7145399451255798, + "num_tokens": 515623550.0, + "step": 19926 + }, + { + "epoch": 2.1883373599824294, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4395699501037598, + "learning_rate": 1e-06, + "loss": 0.7217, + "mean_token_accuracy": 0.7618780136108398, + "num_tokens": 515647512.0, + "step": 19927 + }, + { + "epoch": 2.1884471776850427, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.763408899307251, + "learning_rate": 1e-06, + "loss": 0.7642, + "mean_token_accuracy": 0.7597134113311768, + "num_tokens": 515666805.0, + "step": 19928 + }, + { + "epoch": 2.1885569953876565, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5314927101135254, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7425442934036255, + "num_tokens": 515691097.0, + "step": 19929 + }, + { + "epoch": 2.1886668130902702, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1611855030059814, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.7015289664268494, + "num_tokens": 515721818.0, + "step": 19930 + }, + { + "epoch": 2.188776630792884, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1935319900512695, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7110463976860046, + "num_tokens": 515754030.0, + "step": 19931 + }, + { + "epoch": 2.1888864484954973, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3542752265930176, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7324334383010864, + "num_tokens": 515780601.0, + "step": 19932 + }, + { + "epoch": 2.188996266198111, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.435668706893921, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7340580821037292, + "num_tokens": 515803712.0, + "step": 19933 + }, + { + "epoch": 2.189106083900725, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4633617401123047, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7518696188926697, + "num_tokens": 515826421.0, + "step": 19934 + }, + { + "epoch": 2.1892159016033386, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.421657085418701, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7129773497581482, + "num_tokens": 515852185.0, + "step": 19935 + }, + { + "epoch": 2.1893257193059523, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.312141180038452, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7203425765037537, + "num_tokens": 515880345.0, + "step": 19936 + }, + { + "epoch": 2.1894355370085656, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.185708999633789, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7273926138877869, + "num_tokens": 515909406.0, + "step": 19937 + }, + { + "epoch": 2.1895453547111794, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.1669020652770996, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7353508472442627, + "num_tokens": 515938440.0, + "step": 19938 + }, + { + "epoch": 2.189655172413793, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5832905769348145, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7301612496376038, + "num_tokens": 515960058.0, + "step": 19939 + }, + { + "epoch": 2.189764990116407, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.737450122833252, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7368769645690918, + "num_tokens": 515980139.0, + "step": 19940 + }, + { + "epoch": 2.18987480781902, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.5970699787139893, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7345504760742188, + "num_tokens": 516000716.0, + "step": 19941 + }, + { + "epoch": 2.189984625521634, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.980114459991455, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7411215305328369, + "num_tokens": 516019694.0, + "step": 19942 + }, + { + "epoch": 2.1900944432242477, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.210261106491089, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7147001028060913, + "num_tokens": 516049134.0, + "step": 19943 + }, + { + "epoch": 2.1902042609268615, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.2997584342956543, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7212327718734741, + "num_tokens": 516076668.0, + "step": 19944 + }, + { + "epoch": 2.1903140786294752, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4480340480804443, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7351406812667847, + "num_tokens": 516101344.0, + "step": 19945 + }, + { + "epoch": 2.1904238963320886, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4092459678649902, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7170534133911133, + "num_tokens": 516127699.0, + "step": 19946 + }, + { + "epoch": 2.1905337140347023, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.286590337753296, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7408862113952637, + "num_tokens": 516153496.0, + "step": 19947 + }, + { + "epoch": 2.190643531737316, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.489999294281006, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.710097074508667, + "num_tokens": 516179007.0, + "step": 19948 + }, + { + "epoch": 2.19075334943993, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.360508918762207, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7311339378356934, + "num_tokens": 516203928.0, + "step": 19949 + }, + { + "epoch": 2.1908631671425436, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.4676103591918945, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7402780055999756, + "num_tokens": 516227878.0, + "step": 19950 + }, + { + "epoch": 2.190972984845157, + "ewc_loss": 0.00017642974853515625, + "grad_norm": 2.3158187866210938, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7251173257827759, + "num_tokens": 516254967.0, + "step": 19951 + }, + { + "epoch": 2.1910828025477707, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5113589763641357, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7546852231025696, + "num_tokens": 516277432.0, + "step": 19952 + }, + { + "epoch": 2.1911926202503844, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6927456855773926, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7225906848907471, + "num_tokens": 516300329.0, + "step": 19953 + }, + { + "epoch": 2.191302437952998, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4016082286834717, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7129342555999756, + "num_tokens": 516328649.0, + "step": 19954 + }, + { + "epoch": 2.1914122556556115, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3998148441314697, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7239770293235779, + "num_tokens": 516353113.0, + "step": 19955 + }, + { + "epoch": 2.1915220733582252, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1188113689422607, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7400831580162048, + "num_tokens": 516383463.0, + "step": 19956 + }, + { + "epoch": 2.191631891060839, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3640928268432617, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7498358488082886, + "num_tokens": 516407754.0, + "step": 19957 + }, + { + "epoch": 2.1917417087634528, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6539340019226074, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.7500694990158081, + "num_tokens": 516428975.0, + "step": 19958 + }, + { + "epoch": 2.1918515264660665, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.863008975982666, + "learning_rate": 1e-06, + "loss": 0.7772, + "mean_token_accuracy": 0.7471160888671875, + "num_tokens": 516448000.0, + "step": 19959 + }, + { + "epoch": 2.19196134416868, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1928348541259766, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7128368616104126, + "num_tokens": 516475023.0, + "step": 19960 + }, + { + "epoch": 2.1920711618712936, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3298091888427734, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7383275032043457, + "num_tokens": 516501229.0, + "step": 19961 + }, + { + "epoch": 2.1921809795739073, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2471139430999756, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7323879599571228, + "num_tokens": 516532095.0, + "step": 19962 + }, + { + "epoch": 2.192290797276521, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4873173236846924, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7301645278930664, + "num_tokens": 516555590.0, + "step": 19963 + }, + { + "epoch": 2.1924006149791344, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1959667205810547, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7249531745910645, + "num_tokens": 516586018.0, + "step": 19964 + }, + { + "epoch": 2.192510432681748, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.499683380126953, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.719292402267456, + "num_tokens": 516610101.0, + "step": 19965 + }, + { + "epoch": 2.192620250384362, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.44497013092041, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7312828898429871, + "num_tokens": 516634371.0, + "step": 19966 + }, + { + "epoch": 2.1927300680869757, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3978185653686523, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7474520206451416, + "num_tokens": 516659537.0, + "step": 19967 + }, + { + "epoch": 2.1928398857895894, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4642646312713623, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7364928722381592, + "num_tokens": 516684554.0, + "step": 19968 + }, + { + "epoch": 2.1929497034922028, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.549344301223755, + "learning_rate": 1e-06, + "loss": 0.7719, + "mean_token_accuracy": 0.7530131340026855, + "num_tokens": 516706127.0, + "step": 19969 + }, + { + "epoch": 2.1930595211948165, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.416508197784424, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7363987565040588, + "num_tokens": 516729918.0, + "step": 19970 + }, + { + "epoch": 2.1931693388974303, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.336958646774292, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.745903491973877, + "num_tokens": 516754609.0, + "step": 19971 + }, + { + "epoch": 2.193279156600044, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5418686866760254, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7151267528533936, + "num_tokens": 516779618.0, + "step": 19972 + }, + { + "epoch": 2.193388974302658, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3529391288757324, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7390146255493164, + "num_tokens": 516803725.0, + "step": 19973 + }, + { + "epoch": 2.193498792005271, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.324568033218384, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7515251636505127, + "num_tokens": 516829951.0, + "step": 19974 + }, + { + "epoch": 2.193608609707885, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1469428539276123, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7040442228317261, + "num_tokens": 516861823.0, + "step": 19975 + }, + { + "epoch": 2.1937184274104986, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3241453170776367, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7224111557006836, + "num_tokens": 516888694.0, + "step": 19976 + }, + { + "epoch": 2.1938282451131124, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2945616245269775, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.751838207244873, + "num_tokens": 516913859.0, + "step": 19977 + }, + { + "epoch": 2.193938062815726, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.761594533920288, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7343048453330994, + "num_tokens": 516934106.0, + "step": 19978 + }, + { + "epoch": 2.1940478805183394, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2376933097839355, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7388885021209717, + "num_tokens": 516962220.0, + "step": 19979 + }, + { + "epoch": 2.194157698220953, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.225818634033203, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7265741229057312, + "num_tokens": 516992097.0, + "step": 19980 + }, + { + "epoch": 2.194267515923567, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.106905698776245, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7314913272857666, + "num_tokens": 517024985.0, + "step": 19981 + }, + { + "epoch": 2.1943773336261807, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.446274518966675, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7294233441352844, + "num_tokens": 517049686.0, + "step": 19982 + }, + { + "epoch": 2.194487151328794, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5927236080169678, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7313247323036194, + "num_tokens": 517073345.0, + "step": 19983 + }, + { + "epoch": 2.1945969690314078, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.321749448776245, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7340903878211975, + "num_tokens": 517099659.0, + "step": 19984 + }, + { + "epoch": 2.1947067867340215, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.179516077041626, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7224538326263428, + "num_tokens": 517127778.0, + "step": 19985 + }, + { + "epoch": 2.1948166044366353, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6983745098114014, + "learning_rate": 1e-06, + "loss": 0.7146, + "mean_token_accuracy": 0.7646573781967163, + "num_tokens": 517147665.0, + "step": 19986 + }, + { + "epoch": 2.194926422139249, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.261317253112793, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7174906730651855, + "num_tokens": 517173900.0, + "step": 19987 + }, + { + "epoch": 2.1950362398418624, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.317580223083496, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7401219606399536, + "num_tokens": 517201170.0, + "step": 19988 + }, + { + "epoch": 2.195146057544476, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.199655055999756, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6912586688995361, + "num_tokens": 517230427.0, + "step": 19989 + }, + { + "epoch": 2.19525587524709, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.095228910446167, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7130621671676636, + "num_tokens": 517265792.0, + "step": 19990 + }, + { + "epoch": 2.1953656929497036, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.623239278793335, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7376905679702759, + "num_tokens": 517287592.0, + "step": 19991 + }, + { + "epoch": 2.195475510652317, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.511274814605713, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7516069412231445, + "num_tokens": 517310886.0, + "step": 19992 + }, + { + "epoch": 2.1955853283549307, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.604010581970215, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.7421679496765137, + "num_tokens": 517332572.0, + "step": 19993 + }, + { + "epoch": 2.1956951460575445, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5828678607940674, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7343062162399292, + "num_tokens": 517355634.0, + "step": 19994 + }, + { + "epoch": 2.195804963760158, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1142637729644775, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7013004422187805, + "num_tokens": 517385468.0, + "step": 19995 + }, + { + "epoch": 2.195914781462772, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.0490128993988037, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7218959331512451, + "num_tokens": 517417204.0, + "step": 19996 + }, + { + "epoch": 2.1960245991653853, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4699249267578125, + "learning_rate": 1e-06, + "loss": 0.7578, + "mean_token_accuracy": 0.7567841410636902, + "num_tokens": 517441033.0, + "step": 19997 + }, + { + "epoch": 2.196134416867999, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3206300735473633, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.711438000202179, + "num_tokens": 517469258.0, + "step": 19998 + }, + { + "epoch": 2.196244234570613, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3385467529296875, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7096544504165649, + "num_tokens": 517495697.0, + "step": 19999 + }, + { + "epoch": 2.1963540522732266, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.128993511199951, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7309216856956482, + "num_tokens": 517526328.0, + "step": 20000 + }, + { + "epoch": 2.1964638699758403, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2814114093780518, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7051436901092529, + "num_tokens": 517556477.0, + "step": 20001 + }, + { + "epoch": 2.1965736876784536, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.310805082321167, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7282358407974243, + "num_tokens": 517583923.0, + "step": 20002 + }, + { + "epoch": 2.1966835053810674, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.0783214569091797, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7117480635643005, + "num_tokens": 517618071.0, + "step": 20003 + }, + { + "epoch": 2.196793323083681, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.551581382751465, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.7640267610549927, + "num_tokens": 517640507.0, + "step": 20004 + }, + { + "epoch": 2.196903140786295, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3668057918548584, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.737115740776062, + "num_tokens": 517665619.0, + "step": 20005 + }, + { + "epoch": 2.197012958488908, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1877365112304688, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7423880696296692, + "num_tokens": 517693362.0, + "step": 20006 + }, + { + "epoch": 2.197122776191522, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.47868013381958, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7299126386642456, + "num_tokens": 517716288.0, + "step": 20007 + }, + { + "epoch": 2.1972325938941357, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3638694286346436, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7121990919113159, + "num_tokens": 517741774.0, + "step": 20008 + }, + { + "epoch": 2.1973424115967495, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.185525894165039, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7161060571670532, + "num_tokens": 517773664.0, + "step": 20009 + }, + { + "epoch": 2.1974522292993632, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.574136972427368, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7420184016227722, + "num_tokens": 517793926.0, + "step": 20010 + }, + { + "epoch": 2.1975620470019765, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1581132411956787, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7197400331497192, + "num_tokens": 517823769.0, + "step": 20011 + }, + { + "epoch": 2.1976718647045903, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.507598876953125, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7347226142883301, + "num_tokens": 517849202.0, + "step": 20012 + }, + { + "epoch": 2.197781682407204, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.409144401550293, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7207520008087158, + "num_tokens": 517877265.0, + "step": 20013 + }, + { + "epoch": 2.197891500109818, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4464311599731445, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7365115880966187, + "num_tokens": 517900377.0, + "step": 20014 + }, + { + "epoch": 2.198001317812431, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1582021713256836, + "learning_rate": 1e-06, + "loss": 0.755, + "mean_token_accuracy": 0.7560670375823975, + "num_tokens": 517927383.0, + "step": 20015 + }, + { + "epoch": 2.198111135515045, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1461222171783447, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7094824314117432, + "num_tokens": 517957042.0, + "step": 20016 + }, + { + "epoch": 2.1982209532176586, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3906524181365967, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7194348573684692, + "num_tokens": 517982093.0, + "step": 20017 + }, + { + "epoch": 2.1983307709202724, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.304783582687378, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7510986328125, + "num_tokens": 518006726.0, + "step": 20018 + }, + { + "epoch": 2.198440588622886, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1723155975341797, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7263013124465942, + "num_tokens": 518038182.0, + "step": 20019 + }, + { + "epoch": 2.1985504063254995, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4307515621185303, + "learning_rate": 1e-06, + "loss": 0.7593, + "mean_token_accuracy": 0.7562441825866699, + "num_tokens": 518061637.0, + "step": 20020 + }, + { + "epoch": 2.1986602240281132, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.270704507827759, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7195360064506531, + "num_tokens": 518091519.0, + "step": 20021 + }, + { + "epoch": 2.198770041730727, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4102470874786377, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7104947566986084, + "num_tokens": 518116665.0, + "step": 20022 + }, + { + "epoch": 2.1988798594333407, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1132800579071045, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7176176309585571, + "num_tokens": 518151433.0, + "step": 20023 + }, + { + "epoch": 2.1989896771359545, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.756558418273926, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7129966020584106, + "num_tokens": 518172320.0, + "step": 20024 + }, + { + "epoch": 2.199099494838568, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6659433841705322, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7625190019607544, + "num_tokens": 518192686.0, + "step": 20025 + }, + { + "epoch": 2.1992093125411816, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.382467746734619, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.717119038105011, + "num_tokens": 518219581.0, + "step": 20026 + }, + { + "epoch": 2.1993191302437953, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2518272399902344, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7294488549232483, + "num_tokens": 518247489.0, + "step": 20027 + }, + { + "epoch": 2.199428947946409, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.601292371749878, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7377036213874817, + "num_tokens": 518269276.0, + "step": 20028 + }, + { + "epoch": 2.199538765649023, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2965190410614014, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7492226362228394, + "num_tokens": 518297160.0, + "step": 20029 + }, + { + "epoch": 2.199648583351636, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.407972812652588, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7118654847145081, + "num_tokens": 518324758.0, + "step": 20030 + }, + { + "epoch": 2.19975840105425, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3458127975463867, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7449281811714172, + "num_tokens": 518351349.0, + "step": 20031 + }, + { + "epoch": 2.1998682187568637, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2199859619140625, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7403958439826965, + "num_tokens": 518379043.0, + "step": 20032 + }, + { + "epoch": 2.1999780364594774, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2218854427337646, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7250624895095825, + "num_tokens": 518408717.0, + "step": 20033 + }, + { + "epoch": 2.2000878541620907, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.553345203399658, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7237508296966553, + "num_tokens": 518433853.0, + "step": 20034 + }, + { + "epoch": 2.2001976718647045, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.614956855773926, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7154616117477417, + "num_tokens": 518457724.0, + "step": 20035 + }, + { + "epoch": 2.2003074895673183, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2155380249023438, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7535970211029053, + "num_tokens": 518485575.0, + "step": 20036 + }, + { + "epoch": 2.200417307269932, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2646241188049316, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7117912769317627, + "num_tokens": 518517005.0, + "step": 20037 + }, + { + "epoch": 2.2005271249725458, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.477790117263794, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7134619951248169, + "num_tokens": 518541236.0, + "step": 20038 + }, + { + "epoch": 2.200636942675159, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.633404016494751, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7396228313446045, + "num_tokens": 518565424.0, + "step": 20039 + }, + { + "epoch": 2.200746760377773, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.360722541809082, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.746906042098999, + "num_tokens": 518589389.0, + "step": 20040 + }, + { + "epoch": 2.2008565780803866, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4252095222473145, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7284491062164307, + "num_tokens": 518614095.0, + "step": 20041 + }, + { + "epoch": 2.2009663957830004, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.479081153869629, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7252122163772583, + "num_tokens": 518638191.0, + "step": 20042 + }, + { + "epoch": 2.2010762134856137, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6684978008270264, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.734515368938446, + "num_tokens": 518660484.0, + "step": 20043 + }, + { + "epoch": 2.2011860311882274, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.266683578491211, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7148622274398804, + "num_tokens": 518687926.0, + "step": 20044 + }, + { + "epoch": 2.201295848890841, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.143366575241089, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7161365747451782, + "num_tokens": 518719279.0, + "step": 20045 + }, + { + "epoch": 2.201405666593455, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.0495612621307373, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7193484902381897, + "num_tokens": 518753491.0, + "step": 20046 + }, + { + "epoch": 2.2015154842960687, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.240139961242676, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7208622694015503, + "num_tokens": 518780926.0, + "step": 20047 + }, + { + "epoch": 2.201625301998682, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.507122278213501, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7067165374755859, + "num_tokens": 518805478.0, + "step": 20048 + }, + { + "epoch": 2.2017351197012958, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5020439624786377, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7314300537109375, + "num_tokens": 518830675.0, + "step": 20049 + }, + { + "epoch": 2.2018449374039095, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5295419692993164, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.749825656414032, + "num_tokens": 518854521.0, + "step": 20050 + }, + { + "epoch": 2.2019547551065233, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4810948371887207, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7219455242156982, + "num_tokens": 518882182.0, + "step": 20051 + }, + { + "epoch": 2.202064572809137, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2308104038238525, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.728821337223053, + "num_tokens": 518911153.0, + "step": 20052 + }, + { + "epoch": 2.2021743905117503, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2811672687530518, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7214715480804443, + "num_tokens": 518938259.0, + "step": 20053 + }, + { + "epoch": 2.202284208214364, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4198973178863525, + "learning_rate": 1e-06, + "loss": 0.7285, + "mean_token_accuracy": 0.7667207717895508, + "num_tokens": 518960702.0, + "step": 20054 + }, + { + "epoch": 2.202394025916978, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2407171726226807, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7139480113983154, + "num_tokens": 518989298.0, + "step": 20055 + }, + { + "epoch": 2.2025038436195916, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3982369899749756, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7106137275695801, + "num_tokens": 519014972.0, + "step": 20056 + }, + { + "epoch": 2.2026136613222054, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6077988147735596, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7399728298187256, + "num_tokens": 519038925.0, + "step": 20057 + }, + { + "epoch": 2.2027234790248187, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3253731727600098, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7288562655448914, + "num_tokens": 519064807.0, + "step": 20058 + }, + { + "epoch": 2.2028332967274324, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.537654399871826, + "learning_rate": 1e-06, + "loss": 0.7532, + "mean_token_accuracy": 0.7575178146362305, + "num_tokens": 519086927.0, + "step": 20059 + }, + { + "epoch": 2.202943114430046, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3456757068634033, + "learning_rate": 1e-06, + "loss": 0.7742, + "mean_token_accuracy": 0.7562369108200073, + "num_tokens": 519111766.0, + "step": 20060 + }, + { + "epoch": 2.20305293213266, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.200634717941284, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7136869430541992, + "num_tokens": 519140981.0, + "step": 20061 + }, + { + "epoch": 2.2031627498352733, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.449486255645752, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7182778120040894, + "num_tokens": 519165772.0, + "step": 20062 + }, + { + "epoch": 2.203272567537887, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.311272382736206, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7478017807006836, + "num_tokens": 519191829.0, + "step": 20063 + }, + { + "epoch": 2.203382385240501, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.499109983444214, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7134568691253662, + "num_tokens": 519216179.0, + "step": 20064 + }, + { + "epoch": 2.2034922029431145, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1596195697784424, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.749755322933197, + "num_tokens": 519245075.0, + "step": 20065 + }, + { + "epoch": 2.2036020206457283, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2821462154388428, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7342818975448608, + "num_tokens": 519272624.0, + "step": 20066 + }, + { + "epoch": 2.2037118383483416, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.411637544631958, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7265985012054443, + "num_tokens": 519297407.0, + "step": 20067 + }, + { + "epoch": 2.2038216560509554, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.010112762451172, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7064077854156494, + "num_tokens": 519335419.0, + "step": 20068 + }, + { + "epoch": 2.203931473753569, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6957225799560547, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7545703649520874, + "num_tokens": 519355629.0, + "step": 20069 + }, + { + "epoch": 2.204041291456183, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.363762378692627, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7330451607704163, + "num_tokens": 519381439.0, + "step": 20070 + }, + { + "epoch": 2.204151109158796, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4451637268066406, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7192527055740356, + "num_tokens": 519404641.0, + "step": 20071 + }, + { + "epoch": 2.20426092686141, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.204604387283325, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7232475876808167, + "num_tokens": 519432333.0, + "step": 20072 + }, + { + "epoch": 2.2043707445640237, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2641103267669678, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7197384834289551, + "num_tokens": 519462964.0, + "step": 20073 + }, + { + "epoch": 2.2044805622666375, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.368229389190674, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7356698513031006, + "num_tokens": 519487752.0, + "step": 20074 + }, + { + "epoch": 2.2045903799692512, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.208587169647217, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7245445251464844, + "num_tokens": 519516482.0, + "step": 20075 + }, + { + "epoch": 2.2047001976718645, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4612350463867188, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.750578761100769, + "num_tokens": 519540401.0, + "step": 20076 + }, + { + "epoch": 2.2048100153744783, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4960994720458984, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7177531123161316, + "num_tokens": 519564891.0, + "step": 20077 + }, + { + "epoch": 2.204919833077092, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1438779830932617, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7191659212112427, + "num_tokens": 519595185.0, + "step": 20078 + }, + { + "epoch": 2.205029650779706, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.397156000137329, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7259339094161987, + "num_tokens": 519620293.0, + "step": 20079 + }, + { + "epoch": 2.2051394684823196, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5715065002441406, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7402668595314026, + "num_tokens": 519643636.0, + "step": 20080 + }, + { + "epoch": 2.205249286184933, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.18760085105896, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7174225449562073, + "num_tokens": 519671611.0, + "step": 20081 + }, + { + "epoch": 2.2053591038875466, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.0003821849823, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7152458429336548, + "num_tokens": 519706511.0, + "step": 20082 + }, + { + "epoch": 2.2054689215901604, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.506932497024536, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7410961389541626, + "num_tokens": 519729225.0, + "step": 20083 + }, + { + "epoch": 2.205578739292774, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1901867389678955, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7286350727081299, + "num_tokens": 519757241.0, + "step": 20084 + }, + { + "epoch": 2.2056885569953875, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2495439052581787, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7359914183616638, + "num_tokens": 519783692.0, + "step": 20085 + }, + { + "epoch": 2.205798374698001, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.483328342437744, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7542442083358765, + "num_tokens": 519807339.0, + "step": 20086 + }, + { + "epoch": 2.205908192400615, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5163300037384033, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7436729669570923, + "num_tokens": 519830355.0, + "step": 20087 + }, + { + "epoch": 2.2060180101032287, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.197266101837158, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7305681705474854, + "num_tokens": 519859643.0, + "step": 20088 + }, + { + "epoch": 2.2061278278058425, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.534885883331299, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7421000003814697, + "num_tokens": 519883210.0, + "step": 20089 + }, + { + "epoch": 2.206237645508456, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6258950233459473, + "learning_rate": 1e-06, + "loss": 0.7708, + "mean_token_accuracy": 0.7551181316375732, + "num_tokens": 519903342.0, + "step": 20090 + }, + { + "epoch": 2.2063474632110696, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2071352005004883, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7370685935020447, + "num_tokens": 519930300.0, + "step": 20091 + }, + { + "epoch": 2.2064572809136833, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1772119998931885, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7219791412353516, + "num_tokens": 519960474.0, + "step": 20092 + }, + { + "epoch": 2.206567098616297, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2571542263031006, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7390801906585693, + "num_tokens": 519986070.0, + "step": 20093 + }, + { + "epoch": 2.2066769163189104, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4327399730682373, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7414373159408569, + "num_tokens": 520011594.0, + "step": 20094 + }, + { + "epoch": 2.206786734021524, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.479771375656128, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7151284217834473, + "num_tokens": 520037705.0, + "step": 20095 + }, + { + "epoch": 2.206896551724138, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.7920596599578857, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7410155534744263, + "num_tokens": 520056695.0, + "step": 20096 + }, + { + "epoch": 2.2070063694267517, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.515714406967163, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.7420897483825684, + "num_tokens": 520079533.0, + "step": 20097 + }, + { + "epoch": 2.2071161871293654, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.541724443435669, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7278091311454773, + "num_tokens": 520104959.0, + "step": 20098 + }, + { + "epoch": 2.2072260048319787, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6049773693084717, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7479410767555237, + "num_tokens": 520126584.0, + "step": 20099 + }, + { + "epoch": 2.2073358225345925, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4212467670440674, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7342775464057922, + "num_tokens": 520149976.0, + "step": 20100 + }, + { + "epoch": 2.2074456402372062, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4276540279388428, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7343894839286804, + "num_tokens": 520177916.0, + "step": 20101 + }, + { + "epoch": 2.20755545793982, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2471630573272705, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7253788709640503, + "num_tokens": 520205584.0, + "step": 20102 + }, + { + "epoch": 2.2076652756424338, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2938451766967773, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7234097719192505, + "num_tokens": 520234122.0, + "step": 20103 + }, + { + "epoch": 2.207775093345047, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.526123046875, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7077289819717407, + "num_tokens": 520258717.0, + "step": 20104 + }, + { + "epoch": 2.207884911047661, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4587225914001465, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7418624758720398, + "num_tokens": 520284535.0, + "step": 20105 + }, + { + "epoch": 2.2079947287502746, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3267154693603516, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7526018023490906, + "num_tokens": 520310647.0, + "step": 20106 + }, + { + "epoch": 2.2081045464528883, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.0260984897613525, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7105380892753601, + "num_tokens": 520344625.0, + "step": 20107 + }, + { + "epoch": 2.208214364155502, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.319063663482666, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.736973226070404, + "num_tokens": 520368818.0, + "step": 20108 + }, + { + "epoch": 2.2083241818581154, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.166517496109009, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7321091890335083, + "num_tokens": 520396062.0, + "step": 20109 + }, + { + "epoch": 2.208433999560729, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.263566493988037, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7285168766975403, + "num_tokens": 520423499.0, + "step": 20110 + }, + { + "epoch": 2.208543817263343, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 1.9255808591842651, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.716118574142456, + "num_tokens": 520460909.0, + "step": 20111 + }, + { + "epoch": 2.2086536349659567, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3434386253356934, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7371008396148682, + "num_tokens": 520487116.0, + "step": 20112 + }, + { + "epoch": 2.20876345266857, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1861467361450195, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7105072736740112, + "num_tokens": 520516152.0, + "step": 20113 + }, + { + "epoch": 2.2088732703711838, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4348936080932617, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7460211515426636, + "num_tokens": 520539944.0, + "step": 20114 + }, + { + "epoch": 2.2089830880737975, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2021591663360596, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7358975410461426, + "num_tokens": 520567663.0, + "step": 20115 + }, + { + "epoch": 2.2090929057764113, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6345632076263428, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7520185708999634, + "num_tokens": 520590056.0, + "step": 20116 + }, + { + "epoch": 2.209202723479025, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.545017957687378, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7422840595245361, + "num_tokens": 520612078.0, + "step": 20117 + }, + { + "epoch": 2.2093125411816383, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.8699445724487305, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7353230714797974, + "num_tokens": 520631927.0, + "step": 20118 + }, + { + "epoch": 2.209422358884252, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4060208797454834, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7262347936630249, + "num_tokens": 520658703.0, + "step": 20119 + }, + { + "epoch": 2.209532176586866, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.7357914447784424, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7402136921882629, + "num_tokens": 520677388.0, + "step": 20120 + }, + { + "epoch": 2.2096419942894796, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2831027507781982, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7143164277076721, + "num_tokens": 520706453.0, + "step": 20121 + }, + { + "epoch": 2.209751811992093, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2265710830688477, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7302349209785461, + "num_tokens": 520736876.0, + "step": 20122 + }, + { + "epoch": 2.2098616296947067, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5399041175842285, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6965296268463135, + "num_tokens": 520762241.0, + "step": 20123 + }, + { + "epoch": 2.2099714473973204, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5485870838165283, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7333305478096008, + "num_tokens": 520785181.0, + "step": 20124 + }, + { + "epoch": 2.210081265099934, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.276210069656372, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7262258529663086, + "num_tokens": 520812759.0, + "step": 20125 + }, + { + "epoch": 2.210191082802548, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3369698524475098, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7421064972877502, + "num_tokens": 520839324.0, + "step": 20126 + }, + { + "epoch": 2.2103009005051613, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3259403705596924, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.733972430229187, + "num_tokens": 520867844.0, + "step": 20127 + }, + { + "epoch": 2.210410718207775, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2669296264648438, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7115280628204346, + "num_tokens": 520897337.0, + "step": 20128 + }, + { + "epoch": 2.2105205359103888, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.381939172744751, + "learning_rate": 1e-06, + "loss": 0.7737, + "mean_token_accuracy": 0.754463791847229, + "num_tokens": 520922710.0, + "step": 20129 + }, + { + "epoch": 2.2106303536130025, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.417001247406006, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.719732403755188, + "num_tokens": 520948579.0, + "step": 20130 + }, + { + "epoch": 2.2107401713156163, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.490481376647949, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7205286026000977, + "num_tokens": 520976304.0, + "step": 20131 + }, + { + "epoch": 2.2108499890182296, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.219860553741455, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6939400434494019, + "num_tokens": 521005547.0, + "step": 20132 + }, + { + "epoch": 2.2109598067208434, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.364326000213623, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7302219867706299, + "num_tokens": 521032555.0, + "step": 20133 + }, + { + "epoch": 2.211069624423457, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.45356822013855, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7094390392303467, + "num_tokens": 521059832.0, + "step": 20134 + }, + { + "epoch": 2.211179442126071, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.253430128097534, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7182033061981201, + "num_tokens": 521088278.0, + "step": 20135 + }, + { + "epoch": 2.211289259828684, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5429277420043945, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7185513973236084, + "num_tokens": 521111519.0, + "step": 20136 + }, + { + "epoch": 2.211399077531298, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.24619722366333, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7360560894012451, + "num_tokens": 521139870.0, + "step": 20137 + }, + { + "epoch": 2.2115088952339117, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4057772159576416, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7330718040466309, + "num_tokens": 521167763.0, + "step": 20138 + }, + { + "epoch": 2.2116187129365255, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6626946926116943, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7304888963699341, + "num_tokens": 521193950.0, + "step": 20139 + }, + { + "epoch": 2.211728530639139, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.171649217605591, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7230293154716492, + "num_tokens": 521227519.0, + "step": 20140 + }, + { + "epoch": 2.2118383483417525, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.8467767238616943, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7369720339775085, + "num_tokens": 521247804.0, + "step": 20141 + }, + { + "epoch": 2.2119481660443663, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.648089647293091, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7216431498527527, + "num_tokens": 521269010.0, + "step": 20142 + }, + { + "epoch": 2.21205798374698, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4521708488464355, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7234205007553101, + "num_tokens": 521293769.0, + "step": 20143 + }, + { + "epoch": 2.212167801449594, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4444987773895264, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7386194467544556, + "num_tokens": 521318842.0, + "step": 20144 + }, + { + "epoch": 2.212277619152207, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.258540391921997, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7279845476150513, + "num_tokens": 521346698.0, + "step": 20145 + }, + { + "epoch": 2.212387436854821, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4521288871765137, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7429301738739014, + "num_tokens": 521371578.0, + "step": 20146 + }, + { + "epoch": 2.2124972545574346, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3376882076263428, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7439366579055786, + "num_tokens": 521397721.0, + "step": 20147 + }, + { + "epoch": 2.2126070722600484, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.60398006439209, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7293431758880615, + "num_tokens": 521418959.0, + "step": 20148 + }, + { + "epoch": 2.212716889962662, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.270951747894287, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7225770950317383, + "num_tokens": 521447973.0, + "step": 20149 + }, + { + "epoch": 2.2128267076652754, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.243408203125, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7288789749145508, + "num_tokens": 521475234.0, + "step": 20150 + }, + { + "epoch": 2.212936525367889, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4403040409088135, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7432197332382202, + "num_tokens": 521499190.0, + "step": 20151 + }, + { + "epoch": 2.213046343070503, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4096899032592773, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7307307124137878, + "num_tokens": 521525491.0, + "step": 20152 + }, + { + "epoch": 2.2131561607731167, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5367748737335205, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7246295213699341, + "num_tokens": 521548624.0, + "step": 20153 + }, + { + "epoch": 2.2132659784757305, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1230576038360596, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7122923731803894, + "num_tokens": 521579307.0, + "step": 20154 + }, + { + "epoch": 2.213375796178344, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4833426475524902, + "learning_rate": 1e-06, + "loss": 0.7657, + "mean_token_accuracy": 0.7516202926635742, + "num_tokens": 521602114.0, + "step": 20155 + }, + { + "epoch": 2.2134856138809575, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4295120239257812, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7510737180709839, + "num_tokens": 521625068.0, + "step": 20156 + }, + { + "epoch": 2.2135954315835713, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.342658758163452, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7375513911247253, + "num_tokens": 521649794.0, + "step": 20157 + }, + { + "epoch": 2.213705249286185, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.7749404907226562, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.729597806930542, + "num_tokens": 521669433.0, + "step": 20158 + }, + { + "epoch": 2.213815066988799, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.554392099380493, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7238413095474243, + "num_tokens": 521691755.0, + "step": 20159 + }, + { + "epoch": 2.213924884691412, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3360812664031982, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7324955463409424, + "num_tokens": 521717300.0, + "step": 20160 + }, + { + "epoch": 2.214034702394026, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2743096351623535, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.702012836933136, + "num_tokens": 521746826.0, + "step": 20161 + }, + { + "epoch": 2.2141445200966396, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.700045347213745, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7425585985183716, + "num_tokens": 521767377.0, + "step": 20162 + }, + { + "epoch": 2.2142543377992534, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3300423622131348, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7320554256439209, + "num_tokens": 521792448.0, + "step": 20163 + }, + { + "epoch": 2.2143641555018667, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3573105335235596, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7340059280395508, + "num_tokens": 521817314.0, + "step": 20164 + }, + { + "epoch": 2.2144739732044805, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4941482543945312, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7231735587120056, + "num_tokens": 521844677.0, + "step": 20165 + }, + { + "epoch": 2.2145837909070942, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6499807834625244, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7599708437919617, + "num_tokens": 521865563.0, + "step": 20166 + }, + { + "epoch": 2.214693608609708, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.721731662750244, + "learning_rate": 1e-06, + "loss": 0.828, + "mean_token_accuracy": 0.7393639087677002, + "num_tokens": 521885758.0, + "step": 20167 + }, + { + "epoch": 2.2148034263123217, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6146085262298584, + "learning_rate": 1e-06, + "loss": 0.7519, + "mean_token_accuracy": 0.7639275193214417, + "num_tokens": 521907417.0, + "step": 20168 + }, + { + "epoch": 2.214913244014935, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.308084011077881, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7278033494949341, + "num_tokens": 521934415.0, + "step": 20169 + }, + { + "epoch": 2.215023061717549, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.650236129760742, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7216989398002625, + "num_tokens": 521956542.0, + "step": 20170 + }, + { + "epoch": 2.2151328794201626, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4410393238067627, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.7492153644561768, + "num_tokens": 521981406.0, + "step": 20171 + }, + { + "epoch": 2.2152426971227763, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.347597122192383, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7247543334960938, + "num_tokens": 522006913.0, + "step": 20172 + }, + { + "epoch": 2.2153525148253896, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3830525875091553, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7193880081176758, + "num_tokens": 522032649.0, + "step": 20173 + }, + { + "epoch": 2.2154623325280034, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.651294708251953, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7371931076049805, + "num_tokens": 522054575.0, + "step": 20174 + }, + { + "epoch": 2.215572150230617, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 1.9929540157318115, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7031930685043335, + "num_tokens": 522088974.0, + "step": 20175 + }, + { + "epoch": 2.215681967933231, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.5358073711395264, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.728001594543457, + "num_tokens": 522112488.0, + "step": 20176 + }, + { + "epoch": 2.2157917856358447, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.597278356552124, + "learning_rate": 1e-06, + "loss": 0.7589, + "mean_token_accuracy": 0.7556805610656738, + "num_tokens": 522133063.0, + "step": 20177 + }, + { + "epoch": 2.215901603338458, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2583975791931152, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7184315919876099, + "num_tokens": 522161195.0, + "step": 20178 + }, + { + "epoch": 2.2160114210410717, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2741823196411133, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7330220937728882, + "num_tokens": 522188462.0, + "step": 20179 + }, + { + "epoch": 2.2161212387436855, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.20298171043396, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.714703381061554, + "num_tokens": 522218385.0, + "step": 20180 + }, + { + "epoch": 2.2162310564462993, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.093754291534424, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7386761903762817, + "num_tokens": 522250015.0, + "step": 20181 + }, + { + "epoch": 2.216340874148913, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3097002506256104, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7280921936035156, + "num_tokens": 522276403.0, + "step": 20182 + }, + { + "epoch": 2.2164506918515263, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.139281988143921, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7148297429084778, + "num_tokens": 522309096.0, + "step": 20183 + }, + { + "epoch": 2.21656050955414, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.270387649536133, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7097223997116089, + "num_tokens": 522337840.0, + "step": 20184 + }, + { + "epoch": 2.216670327256754, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.369767189025879, + "learning_rate": 1e-06, + "loss": 0.7307, + "mean_token_accuracy": 0.7640409469604492, + "num_tokens": 522361322.0, + "step": 20185 + }, + { + "epoch": 2.2167801449593676, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6126232147216797, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7508383989334106, + "num_tokens": 522382718.0, + "step": 20186 + }, + { + "epoch": 2.216889962661981, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3744149208068848, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7525544166564941, + "num_tokens": 522407048.0, + "step": 20187 + }, + { + "epoch": 2.2169997803645947, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2812747955322266, + "learning_rate": 1e-06, + "loss": 0.726, + "mean_token_accuracy": 0.7705293297767639, + "num_tokens": 522432975.0, + "step": 20188 + }, + { + "epoch": 2.2171095980672084, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3696300983428955, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7357001304626465, + "num_tokens": 522460947.0, + "step": 20189 + }, + { + "epoch": 2.217219415769822, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1150360107421875, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7362872362136841, + "num_tokens": 522491659.0, + "step": 20190 + }, + { + "epoch": 2.217329233472436, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6003599166870117, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7364001274108887, + "num_tokens": 522513786.0, + "step": 20191 + }, + { + "epoch": 2.2174390511750492, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.311500310897827, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7387212514877319, + "num_tokens": 522540420.0, + "step": 20192 + }, + { + "epoch": 2.217548868877663, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.7415761947631836, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7540921568870544, + "num_tokens": 522560474.0, + "step": 20193 + }, + { + "epoch": 2.2176586865802768, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4643752574920654, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7438195943832397, + "num_tokens": 522583465.0, + "step": 20194 + }, + { + "epoch": 2.2177685042828905, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.337510108947754, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.734472930431366, + "num_tokens": 522609422.0, + "step": 20195 + }, + { + "epoch": 2.217878321985504, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.327521800994873, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7265706658363342, + "num_tokens": 522634318.0, + "step": 20196 + }, + { + "epoch": 2.2179881396881176, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.430447578430176, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.746549129486084, + "num_tokens": 522658245.0, + "step": 20197 + }, + { + "epoch": 2.2180979573907313, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.7179312705993652, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7315335273742676, + "num_tokens": 522682109.0, + "step": 20198 + }, + { + "epoch": 2.218207775093345, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.1114323139190674, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7274372577667236, + "num_tokens": 522713748.0, + "step": 20199 + }, + { + "epoch": 2.218317592795959, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.3245668411254883, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7040722370147705, + "num_tokens": 522742450.0, + "step": 20200 + }, + { + "epoch": 2.218427410498572, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2946152687072754, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7224321365356445, + "num_tokens": 522771817.0, + "step": 20201 + }, + { + "epoch": 2.218537228201186, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.2161972522735596, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.713878870010376, + "num_tokens": 522799701.0, + "step": 20202 + }, + { + "epoch": 2.2186470459037997, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.4376819133758545, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7332308888435364, + "num_tokens": 522824493.0, + "step": 20203 + }, + { + "epoch": 2.2187568636064134, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.6018948554992676, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7230536937713623, + "num_tokens": 522846289.0, + "step": 20204 + }, + { + "epoch": 2.218866681309027, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.677370548248291, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7479684352874756, + "num_tokens": 522867454.0, + "step": 20205 + }, + { + "epoch": 2.2189764990116405, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.486229181289673, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7325035333633423, + "num_tokens": 522889945.0, + "step": 20206 + }, + { + "epoch": 2.2190863167142543, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 1.9634363651275635, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7343446016311646, + "num_tokens": 522923535.0, + "step": 20207 + }, + { + "epoch": 2.219196134416868, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.598942995071411, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7331071496009827, + "num_tokens": 522945158.0, + "step": 20208 + }, + { + "epoch": 2.219305952119482, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.106473207473755, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7299655675888062, + "num_tokens": 522975291.0, + "step": 20209 + }, + { + "epoch": 2.2194157698220955, + "ewc_loss": 0.0001773834228515625, + "grad_norm": 2.213820219039917, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7366373538970947, + "num_tokens": 523002590.0, + "step": 20210 + }, + { + "epoch": 2.219525587524709, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.727710723876953, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7199847102165222, + "num_tokens": 523024099.0, + "step": 20211 + }, + { + "epoch": 2.2196354052273226, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5272324085235596, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7297173738479614, + "num_tokens": 523046434.0, + "step": 20212 + }, + { + "epoch": 2.2197452229299364, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5622646808624268, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7393381595611572, + "num_tokens": 523069020.0, + "step": 20213 + }, + { + "epoch": 2.21985504063255, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5665180683135986, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7385690212249756, + "num_tokens": 523092677.0, + "step": 20214 + }, + { + "epoch": 2.2199648583351634, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1399052143096924, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.725954532623291, + "num_tokens": 523125967.0, + "step": 20215 + }, + { + "epoch": 2.220074676037777, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4000957012176514, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.734275221824646, + "num_tokens": 523152192.0, + "step": 20216 + }, + { + "epoch": 2.220184493740391, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4945404529571533, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7256193161010742, + "num_tokens": 523174814.0, + "step": 20217 + }, + { + "epoch": 2.2202943114430047, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4751503467559814, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7190282344818115, + "num_tokens": 523199753.0, + "step": 20218 + }, + { + "epoch": 2.2204041291456185, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2398428916931152, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7213906049728394, + "num_tokens": 523228109.0, + "step": 20219 + }, + { + "epoch": 2.220513946848232, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2337753772735596, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7231953740119934, + "num_tokens": 523257988.0, + "step": 20220 + }, + { + "epoch": 2.2206237645508455, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.7488372325897217, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7434475421905518, + "num_tokens": 523277505.0, + "step": 20221 + }, + { + "epoch": 2.2207335822534593, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2459523677825928, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7259394526481628, + "num_tokens": 523304940.0, + "step": 20222 + }, + { + "epoch": 2.220843399956073, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2986905574798584, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.717201292514801, + "num_tokens": 523332632.0, + "step": 20223 + }, + { + "epoch": 2.2209532176586864, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4032580852508545, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.704540491104126, + "num_tokens": 523361444.0, + "step": 20224 + }, + { + "epoch": 2.2210630353613, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6410140991210938, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7369973659515381, + "num_tokens": 523383465.0, + "step": 20225 + }, + { + "epoch": 2.221172853063914, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.483365535736084, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7501168251037598, + "num_tokens": 523406351.0, + "step": 20226 + }, + { + "epoch": 2.2212826707665276, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4224164485931396, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.7478143572807312, + "num_tokens": 523432018.0, + "step": 20227 + }, + { + "epoch": 2.2213924884691414, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.229548454284668, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7126073837280273, + "num_tokens": 523463555.0, + "step": 20228 + }, + { + "epoch": 2.2215023061717547, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6993162631988525, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7363184094429016, + "num_tokens": 523485185.0, + "step": 20229 + }, + { + "epoch": 2.2216121238743685, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4746265411376953, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7118111252784729, + "num_tokens": 523509089.0, + "step": 20230 + }, + { + "epoch": 2.221721941576982, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.425234794616699, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7323237657546997, + "num_tokens": 523534084.0, + "step": 20231 + }, + { + "epoch": 2.221831759279596, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1674952507019043, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7176219820976257, + "num_tokens": 523564309.0, + "step": 20232 + }, + { + "epoch": 2.2219415769822097, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.347036361694336, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7425963282585144, + "num_tokens": 523588865.0, + "step": 20233 + }, + { + "epoch": 2.222051394684823, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1754326820373535, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7077507376670837, + "num_tokens": 523620490.0, + "step": 20234 + }, + { + "epoch": 2.222161212387437, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.442718267440796, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7378932237625122, + "num_tokens": 523645323.0, + "step": 20235 + }, + { + "epoch": 2.2222710300900506, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5045950412750244, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7209689021110535, + "num_tokens": 523668337.0, + "step": 20236 + }, + { + "epoch": 2.2223808477926643, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.354236602783203, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7355611324310303, + "num_tokens": 523692874.0, + "step": 20237 + }, + { + "epoch": 2.222490665495278, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.611279010772705, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.754142165184021, + "num_tokens": 523714459.0, + "step": 20238 + }, + { + "epoch": 2.2226004831978914, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.611598253250122, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7366271018981934, + "num_tokens": 523736494.0, + "step": 20239 + }, + { + "epoch": 2.222710300900505, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2104406356811523, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7185372114181519, + "num_tokens": 523766414.0, + "step": 20240 + }, + { + "epoch": 2.222820118603119, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.238234281539917, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7234170436859131, + "num_tokens": 523794278.0, + "step": 20241 + }, + { + "epoch": 2.2229299363057327, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 1.9601404666900635, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7452592849731445, + "num_tokens": 523829092.0, + "step": 20242 + }, + { + "epoch": 2.223039754008346, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.878636360168457, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7429608106613159, + "num_tokens": 523847786.0, + "step": 20243 + }, + { + "epoch": 2.2231495717109597, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6268792152404785, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7148624658584595, + "num_tokens": 523873224.0, + "step": 20244 + }, + { + "epoch": 2.2232593894135735, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3779373168945312, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7422688603401184, + "num_tokens": 523899560.0, + "step": 20245 + }, + { + "epoch": 2.2233692071161872, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2451062202453613, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7236372232437134, + "num_tokens": 523926135.0, + "step": 20246 + }, + { + "epoch": 2.223479024818801, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6949431896209717, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7380006313323975, + "num_tokens": 523947372.0, + "step": 20247 + }, + { + "epoch": 2.2235888425214143, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4102673530578613, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7254354953765869, + "num_tokens": 523972720.0, + "step": 20248 + }, + { + "epoch": 2.223698660224028, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.9307899475097656, + "learning_rate": 1e-06, + "loss": 0.7358, + "mean_token_accuracy": 0.7625982761383057, + "num_tokens": 523990408.0, + "step": 20249 + }, + { + "epoch": 2.223808477926642, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.438810348510742, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.734676718711853, + "num_tokens": 524018298.0, + "step": 20250 + }, + { + "epoch": 2.2239182956292556, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.216118574142456, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7232961058616638, + "num_tokens": 524047897.0, + "step": 20251 + }, + { + "epoch": 2.224028113331869, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4800801277160645, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7117277979850769, + "num_tokens": 524072670.0, + "step": 20252 + }, + { + "epoch": 2.2241379310344827, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.633780002593994, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7527974247932434, + "num_tokens": 524094237.0, + "step": 20253 + }, + { + "epoch": 2.2242477487370964, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.526134729385376, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7141788601875305, + "num_tokens": 524119778.0, + "step": 20254 + }, + { + "epoch": 2.22435756643971, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.249342679977417, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7307077646255493, + "num_tokens": 524147937.0, + "step": 20255 + }, + { + "epoch": 2.224467384142324, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 1.9694743156433105, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7312321662902832, + "num_tokens": 524182165.0, + "step": 20256 + }, + { + "epoch": 2.2245772018449372, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.738560676574707, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7556904554367065, + "num_tokens": 524201340.0, + "step": 20257 + }, + { + "epoch": 2.224687019547551, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4349074363708496, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7276411056518555, + "num_tokens": 524226267.0, + "step": 20258 + }, + { + "epoch": 2.2247968372501647, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.420990467071533, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7404907941818237, + "num_tokens": 524249983.0, + "step": 20259 + }, + { + "epoch": 2.2249066549527785, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2395429611206055, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7081440687179565, + "num_tokens": 524277426.0, + "step": 20260 + }, + { + "epoch": 2.2250164726553923, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.453948974609375, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7177064418792725, + "num_tokens": 524302820.0, + "step": 20261 + }, + { + "epoch": 2.2251262903580056, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2763659954071045, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7291271686553955, + "num_tokens": 524331358.0, + "step": 20262 + }, + { + "epoch": 2.2252361080606193, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3017592430114746, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7267947196960449, + "num_tokens": 524358418.0, + "step": 20263 + }, + { + "epoch": 2.225345925763233, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.570063591003418, + "learning_rate": 1e-06, + "loss": 0.7624, + "mean_token_accuracy": 0.7638026475906372, + "num_tokens": 524379393.0, + "step": 20264 + }, + { + "epoch": 2.225455743465847, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.24715518951416, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7428750991821289, + "num_tokens": 524406388.0, + "step": 20265 + }, + { + "epoch": 2.22556556116846, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1834630966186523, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7145845293998718, + "num_tokens": 524434241.0, + "step": 20266 + }, + { + "epoch": 2.225675378871074, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.251556158065796, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7170292735099792, + "num_tokens": 524461854.0, + "step": 20267 + }, + { + "epoch": 2.2257851965736877, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.080832004547119, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6932908296585083, + "num_tokens": 524496859.0, + "step": 20268 + }, + { + "epoch": 2.2258950142763014, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.498142719268799, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.718564510345459, + "num_tokens": 524523319.0, + "step": 20269 + }, + { + "epoch": 2.226004831978915, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.738189220428467, + "learning_rate": 1e-06, + "loss": 0.7861, + "mean_token_accuracy": 0.7546988725662231, + "num_tokens": 524543772.0, + "step": 20270 + }, + { + "epoch": 2.2261146496815285, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.339648723602295, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7159596681594849, + "num_tokens": 524570526.0, + "step": 20271 + }, + { + "epoch": 2.2262244673841423, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.8322625160217285, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7397924065589905, + "num_tokens": 524591893.0, + "step": 20272 + }, + { + "epoch": 2.226334285086756, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5007050037384033, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7308775186538696, + "num_tokens": 524617959.0, + "step": 20273 + }, + { + "epoch": 2.2264441027893698, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1122825145721436, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7399730682373047, + "num_tokens": 524650018.0, + "step": 20274 + }, + { + "epoch": 2.226553920491983, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3359014987945557, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7282355427742004, + "num_tokens": 524677642.0, + "step": 20275 + }, + { + "epoch": 2.226663738194597, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.575395345687866, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7281123399734497, + "num_tokens": 524700131.0, + "step": 20276 + }, + { + "epoch": 2.2267735558972106, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.487532138824463, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7472081184387207, + "num_tokens": 524723625.0, + "step": 20277 + }, + { + "epoch": 2.2268833735998244, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3349804878234863, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7226486802101135, + "num_tokens": 524749654.0, + "step": 20278 + }, + { + "epoch": 2.226993191302438, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6512489318847656, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7440570592880249, + "num_tokens": 524770999.0, + "step": 20279 + }, + { + "epoch": 2.2271030090050514, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.235992431640625, + "learning_rate": 1e-06, + "loss": 0.791, + "mean_token_accuracy": 0.7478876113891602, + "num_tokens": 524798476.0, + "step": 20280 + }, + { + "epoch": 2.227212826707665, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1934783458709717, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7192307710647583, + "num_tokens": 524830257.0, + "step": 20281 + }, + { + "epoch": 2.227322644410279, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.196460008621216, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7375960350036621, + "num_tokens": 524860125.0, + "step": 20282 + }, + { + "epoch": 2.2274324621128927, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.210529327392578, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7318694591522217, + "num_tokens": 524886906.0, + "step": 20283 + }, + { + "epoch": 2.2275422798155065, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3830673694610596, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7282574772834778, + "num_tokens": 524913202.0, + "step": 20284 + }, + { + "epoch": 2.2276520975181198, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.806811809539795, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.753760576248169, + "num_tokens": 524934388.0, + "step": 20285 + }, + { + "epoch": 2.2277619152207335, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.7447829246520996, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7454516887664795, + "num_tokens": 524954701.0, + "step": 20286 + }, + { + "epoch": 2.2278717329233473, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4627556800842285, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7417806386947632, + "num_tokens": 524982373.0, + "step": 20287 + }, + { + "epoch": 2.227981550625961, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4536020755767822, + "learning_rate": 1e-06, + "loss": 0.8274, + "mean_token_accuracy": 0.7429988980293274, + "num_tokens": 525006604.0, + "step": 20288 + }, + { + "epoch": 2.228091368328575, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5610482692718506, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.747459352016449, + "num_tokens": 525030729.0, + "step": 20289 + }, + { + "epoch": 2.228201186031188, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3795769214630127, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7191472053527832, + "num_tokens": 525057137.0, + "step": 20290 + }, + { + "epoch": 2.228311003733802, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5341837406158447, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7273669242858887, + "num_tokens": 525083368.0, + "step": 20291 + }, + { + "epoch": 2.2284208214364156, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.481703281402588, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7245060801506042, + "num_tokens": 525106564.0, + "step": 20292 + }, + { + "epoch": 2.2285306391390294, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.398996591567993, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7349903583526611, + "num_tokens": 525131346.0, + "step": 20293 + }, + { + "epoch": 2.2286404568416427, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4507176876068115, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7259925007820129, + "num_tokens": 525153578.0, + "step": 20294 + }, + { + "epoch": 2.2287502745442564, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.26708984375, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7361817359924316, + "num_tokens": 525180802.0, + "step": 20295 + }, + { + "epoch": 2.22886009224687, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.323699951171875, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7128485441207886, + "num_tokens": 525207531.0, + "step": 20296 + }, + { + "epoch": 2.228969909949484, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1933982372283936, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7412405014038086, + "num_tokens": 525235532.0, + "step": 20297 + }, + { + "epoch": 2.2290797276520977, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.501777172088623, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7148802876472473, + "num_tokens": 525260748.0, + "step": 20298 + }, + { + "epoch": 2.229189545354711, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2421321868896484, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7263835668563843, + "num_tokens": 525290637.0, + "step": 20299 + }, + { + "epoch": 2.229299363057325, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.686673879623413, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7299559712409973, + "num_tokens": 525311476.0, + "step": 20300 + }, + { + "epoch": 2.2294091807599385, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3656773567199707, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7406482100486755, + "num_tokens": 525336307.0, + "step": 20301 + }, + { + "epoch": 2.2295189984625523, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4343268871307373, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7297065258026123, + "num_tokens": 525360801.0, + "step": 20302 + }, + { + "epoch": 2.2296288161651656, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.8125486373901367, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7281215190887451, + "num_tokens": 525381272.0, + "step": 20303 + }, + { + "epoch": 2.2297386338677794, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.454655408859253, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7125614285469055, + "num_tokens": 525405143.0, + "step": 20304 + }, + { + "epoch": 2.229848451570393, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3140993118286133, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7230774164199829, + "num_tokens": 525432417.0, + "step": 20305 + }, + { + "epoch": 2.229958269273007, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2876882553100586, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7246695756912231, + "num_tokens": 525458740.0, + "step": 20306 + }, + { + "epoch": 2.2300680869756206, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4586434364318848, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7274935841560364, + "num_tokens": 525483970.0, + "step": 20307 + }, + { + "epoch": 2.230177904678234, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.307234048843384, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.730225682258606, + "num_tokens": 525511566.0, + "step": 20308 + }, + { + "epoch": 2.2302877223808477, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5127756595611572, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7496379613876343, + "num_tokens": 525534404.0, + "step": 20309 + }, + { + "epoch": 2.2303975400834615, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.407533884048462, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7240849733352661, + "num_tokens": 525559670.0, + "step": 20310 + }, + { + "epoch": 2.2305073577860752, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5428199768066406, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.74616539478302, + "num_tokens": 525582828.0, + "step": 20311 + }, + { + "epoch": 2.230617175488689, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.559556484222412, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7334597110748291, + "num_tokens": 525605437.0, + "step": 20312 + }, + { + "epoch": 2.2307269931913023, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5035452842712402, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.746681809425354, + "num_tokens": 525628750.0, + "step": 20313 + }, + { + "epoch": 2.230836810893916, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4334871768951416, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7322807908058167, + "num_tokens": 525652681.0, + "step": 20314 + }, + { + "epoch": 2.23094662859653, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.517216682434082, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7436590790748596, + "num_tokens": 525675705.0, + "step": 20315 + }, + { + "epoch": 2.2310564462991436, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2931878566741943, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7351003885269165, + "num_tokens": 525702444.0, + "step": 20316 + }, + { + "epoch": 2.231166264001757, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.5783660411834717, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7473548054695129, + "num_tokens": 525725438.0, + "step": 20317 + }, + { + "epoch": 2.2312760817043706, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.1531312465667725, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7317923903465271, + "num_tokens": 525755652.0, + "step": 20318 + }, + { + "epoch": 2.2313858994069844, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3046960830688477, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7375588417053223, + "num_tokens": 525781826.0, + "step": 20319 + }, + { + "epoch": 2.231495717109598, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 1.9780657291412354, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7073454260826111, + "num_tokens": 525820208.0, + "step": 20320 + }, + { + "epoch": 2.231605534812212, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4504125118255615, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7375545501708984, + "num_tokens": 525841822.0, + "step": 20321 + }, + { + "epoch": 2.2317153525148252, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.132606029510498, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7052749395370483, + "num_tokens": 525873699.0, + "step": 20322 + }, + { + "epoch": 2.231825170217439, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6613364219665527, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7186222672462463, + "num_tokens": 525896504.0, + "step": 20323 + }, + { + "epoch": 2.2319349879200527, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.2889063358306885, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7473902106285095, + "num_tokens": 525923052.0, + "step": 20324 + }, + { + "epoch": 2.2320448056226665, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.3782107830047607, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7217239141464233, + "num_tokens": 525950202.0, + "step": 20325 + }, + { + "epoch": 2.23215462332528, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.021317481994629, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.7008674144744873, + "num_tokens": 525983447.0, + "step": 20326 + }, + { + "epoch": 2.2322644410278936, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.345456838607788, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7027202844619751, + "num_tokens": 526010645.0, + "step": 20327 + }, + { + "epoch": 2.2323742587305073, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.62050199508667, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7204340696334839, + "num_tokens": 526033189.0, + "step": 20328 + }, + { + "epoch": 2.232484076433121, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.482248067855835, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7245990633964539, + "num_tokens": 526057076.0, + "step": 20329 + }, + { + "epoch": 2.232593894135735, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.18489670753479, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7120267152786255, + "num_tokens": 526088194.0, + "step": 20330 + }, + { + "epoch": 2.232703711838348, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4602365493774414, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7092487215995789, + "num_tokens": 526112889.0, + "step": 20331 + }, + { + "epoch": 2.232813529540962, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.22672963142395, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7315822243690491, + "num_tokens": 526140702.0, + "step": 20332 + }, + { + "epoch": 2.2329233472435757, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.0378050804138184, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.6976770162582397, + "num_tokens": 526172674.0, + "step": 20333 + }, + { + "epoch": 2.2330331649461894, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.6210033893585205, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7496776580810547, + "num_tokens": 526194778.0, + "step": 20334 + }, + { + "epoch": 2.233142982648803, + "ewc_loss": 0.00017833709716796875, + "grad_norm": 2.4068198204040527, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7260926961898804, + "num_tokens": 526221856.0, + "step": 20335 + }, + { + "epoch": 2.2332528003514165, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5267493724823, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7406449317932129, + "num_tokens": 526245796.0, + "step": 20336 + }, + { + "epoch": 2.2333626180540302, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3797807693481445, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.7417815923690796, + "num_tokens": 526269914.0, + "step": 20337 + }, + { + "epoch": 2.233472435756644, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4730594158172607, + "learning_rate": 1e-06, + "loss": 0.7678, + "mean_token_accuracy": 0.762641429901123, + "num_tokens": 526292531.0, + "step": 20338 + }, + { + "epoch": 2.2335822534592578, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.7547600269317627, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7426281571388245, + "num_tokens": 526311733.0, + "step": 20339 + }, + { + "epoch": 2.2336920711618715, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.190985918045044, + "learning_rate": 1e-06, + "loss": 0.769, + "mean_token_accuracy": 0.755774199962616, + "num_tokens": 526338515.0, + "step": 20340 + }, + { + "epoch": 2.233801888864485, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1006357669830322, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.7459602952003479, + "num_tokens": 526371474.0, + "step": 20341 + }, + { + "epoch": 2.2339117065670986, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.904447078704834, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7415221929550171, + "num_tokens": 526390129.0, + "step": 20342 + }, + { + "epoch": 2.2340215242697123, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3081963062286377, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7361049652099609, + "num_tokens": 526417402.0, + "step": 20343 + }, + { + "epoch": 2.234131341972326, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4127838611602783, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.757437527179718, + "num_tokens": 526440645.0, + "step": 20344 + }, + { + "epoch": 2.2342411596749394, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2977426052093506, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7570605278015137, + "num_tokens": 526466874.0, + "step": 20345 + }, + { + "epoch": 2.234350977377553, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4869225025177, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7347248792648315, + "num_tokens": 526490122.0, + "step": 20346 + }, + { + "epoch": 2.234460795080167, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1950929164886475, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7096041440963745, + "num_tokens": 526517745.0, + "step": 20347 + }, + { + "epoch": 2.2345706127827807, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.560701370239258, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7456132173538208, + "num_tokens": 526542473.0, + "step": 20348 + }, + { + "epoch": 2.2346804304853944, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.399768352508545, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7065199017524719, + "num_tokens": 526567896.0, + "step": 20349 + }, + { + "epoch": 2.2347902481880078, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.7404115200042725, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7479231357574463, + "num_tokens": 526588527.0, + "step": 20350 + }, + { + "epoch": 2.2349000658906215, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5769622325897217, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.749591588973999, + "num_tokens": 526610904.0, + "step": 20351 + }, + { + "epoch": 2.2350098835932353, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.266827344894409, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7343826293945312, + "num_tokens": 526638307.0, + "step": 20352 + }, + { + "epoch": 2.235119701295849, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6619575023651123, + "learning_rate": 1e-06, + "loss": 0.775, + "mean_token_accuracy": 0.7503587603569031, + "num_tokens": 526657833.0, + "step": 20353 + }, + { + "epoch": 2.2352295189984623, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.183703660964966, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.71383136510849, + "num_tokens": 526688820.0, + "step": 20354 + }, + { + "epoch": 2.235339336701076, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.69340181350708, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7313981056213379, + "num_tokens": 526712067.0, + "step": 20355 + }, + { + "epoch": 2.23544915440369, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5735819339752197, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.74130779504776, + "num_tokens": 526734605.0, + "step": 20356 + }, + { + "epoch": 2.2355589721063036, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2684690952301025, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7163827419281006, + "num_tokens": 526763488.0, + "step": 20357 + }, + { + "epoch": 2.2356687898089174, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.370659351348877, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.73139888048172, + "num_tokens": 526788452.0, + "step": 20358 + }, + { + "epoch": 2.2357786075115307, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4487338066101074, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7299809455871582, + "num_tokens": 526812108.0, + "step": 20359 + }, + { + "epoch": 2.2358884252141444, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.483035087585449, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7384526133537292, + "num_tokens": 526836705.0, + "step": 20360 + }, + { + "epoch": 2.235998242916758, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.381969451904297, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7503929734230042, + "num_tokens": 526860268.0, + "step": 20361 + }, + { + "epoch": 2.236108060619372, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.215682029724121, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7285813093185425, + "num_tokens": 526888612.0, + "step": 20362 + }, + { + "epoch": 2.2362178783219857, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1373472213745117, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7587953805923462, + "num_tokens": 526918780.0, + "step": 20363 + }, + { + "epoch": 2.236327696024599, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.177107334136963, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.741122841835022, + "num_tokens": 526947039.0, + "step": 20364 + }, + { + "epoch": 2.236437513727213, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5941147804260254, + "learning_rate": 1e-06, + "loss": 0.7347, + "mean_token_accuracy": 0.764332115650177, + "num_tokens": 526967088.0, + "step": 20365 + }, + { + "epoch": 2.2365473314298265, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.661886215209961, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7304826974868774, + "num_tokens": 526988400.0, + "step": 20366 + }, + { + "epoch": 2.2366571491324403, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3779942989349365, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7174133658409119, + "num_tokens": 527015572.0, + "step": 20367 + }, + { + "epoch": 2.2367669668350536, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.270183801651001, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7445728778839111, + "num_tokens": 527041790.0, + "step": 20368 + }, + { + "epoch": 2.2368767845376674, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.38611102104187, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7349177598953247, + "num_tokens": 527066764.0, + "step": 20369 + }, + { + "epoch": 2.236986602240281, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.412646532058716, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7327562570571899, + "num_tokens": 527092131.0, + "step": 20370 + }, + { + "epoch": 2.237096419942895, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5053296089172363, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7172716856002808, + "num_tokens": 527115857.0, + "step": 20371 + }, + { + "epoch": 2.2372062376455086, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.229363203048706, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7185392379760742, + "num_tokens": 527146876.0, + "step": 20372 + }, + { + "epoch": 2.237316055348122, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.64471173286438, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7470014095306396, + "num_tokens": 527167614.0, + "step": 20373 + }, + { + "epoch": 2.2374258730507357, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4886467456817627, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7321782112121582, + "num_tokens": 527191952.0, + "step": 20374 + }, + { + "epoch": 2.2375356907533495, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6408603191375732, + "learning_rate": 1e-06, + "loss": 0.7647, + "mean_token_accuracy": 0.757475733757019, + "num_tokens": 527211500.0, + "step": 20375 + }, + { + "epoch": 2.237645508455963, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.435826063156128, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7400235533714294, + "num_tokens": 527236686.0, + "step": 20376 + }, + { + "epoch": 2.2377553261585765, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4945919513702393, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.740399956703186, + "num_tokens": 527260395.0, + "step": 20377 + }, + { + "epoch": 2.2378651438611903, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.612208604812622, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7296794056892395, + "num_tokens": 527282151.0, + "step": 20378 + }, + { + "epoch": 2.237974961563804, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6102237701416016, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7297296524047852, + "num_tokens": 527306096.0, + "step": 20379 + }, + { + "epoch": 2.238084779266418, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.9240477085113525, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.731468677520752, + "num_tokens": 527325235.0, + "step": 20380 + }, + { + "epoch": 2.2381945969690316, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5984299182891846, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7525995969772339, + "num_tokens": 527347389.0, + "step": 20381 + }, + { + "epoch": 2.238304414671645, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4377646446228027, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7147097587585449, + "num_tokens": 527373511.0, + "step": 20382 + }, + { + "epoch": 2.2384142323742586, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1182193756103516, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7302018404006958, + "num_tokens": 527402166.0, + "step": 20383 + }, + { + "epoch": 2.2385240500768724, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.762233257293701, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7397748231887817, + "num_tokens": 527422959.0, + "step": 20384 + }, + { + "epoch": 2.238633867779486, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.334064483642578, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7230448722839355, + "num_tokens": 527450587.0, + "step": 20385 + }, + { + "epoch": 2.2387436854821, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3396382331848145, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7453324794769287, + "num_tokens": 527476126.0, + "step": 20386 + }, + { + "epoch": 2.238853503184713, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6823689937591553, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7332857847213745, + "num_tokens": 527496509.0, + "step": 20387 + }, + { + "epoch": 2.238963320887327, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5742075443267822, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7426905632019043, + "num_tokens": 527518868.0, + "step": 20388 + }, + { + "epoch": 2.2390731385899407, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.355804920196533, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7152178883552551, + "num_tokens": 527547509.0, + "step": 20389 + }, + { + "epoch": 2.2391829562925545, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2346363067626953, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7252951264381409, + "num_tokens": 527579173.0, + "step": 20390 + }, + { + "epoch": 2.2392927739951682, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.780911684036255, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.751672625541687, + "num_tokens": 527599196.0, + "step": 20391 + }, + { + "epoch": 2.2394025916977816, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.080609083175659, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.736351490020752, + "num_tokens": 527629669.0, + "step": 20392 + }, + { + "epoch": 2.2395124094003953, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1442630290985107, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7309937477111816, + "num_tokens": 527660922.0, + "step": 20393 + }, + { + "epoch": 2.239622227103009, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.768221378326416, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7477123737335205, + "num_tokens": 527680675.0, + "step": 20394 + }, + { + "epoch": 2.239732044805623, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.35125994682312, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7457473874092102, + "num_tokens": 527703669.0, + "step": 20395 + }, + { + "epoch": 2.239841862508236, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3019917011260986, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.6996010541915894, + "num_tokens": 527735082.0, + "step": 20396 + }, + { + "epoch": 2.23995168021085, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.394195795059204, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7410473227500916, + "num_tokens": 527759437.0, + "step": 20397 + }, + { + "epoch": 2.2400614979134637, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1888153553009033, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7015249729156494, + "num_tokens": 527789906.0, + "step": 20398 + }, + { + "epoch": 2.2401713156160774, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.53117036819458, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.743401050567627, + "num_tokens": 527812118.0, + "step": 20399 + }, + { + "epoch": 2.240281133318691, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5044164657592773, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7213554978370667, + "num_tokens": 527837353.0, + "step": 20400 + }, + { + "epoch": 2.2403909510213045, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.0766067504882812, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7115521430969238, + "num_tokens": 527868683.0, + "step": 20401 + }, + { + "epoch": 2.2405007687239182, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4011635780334473, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7413370609283447, + "num_tokens": 527892039.0, + "step": 20402 + }, + { + "epoch": 2.240610586426532, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4322428703308105, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7233967781066895, + "num_tokens": 527918095.0, + "step": 20403 + }, + { + "epoch": 2.2407204041291457, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.693246841430664, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7205139994621277, + "num_tokens": 527938217.0, + "step": 20404 + }, + { + "epoch": 2.240830221831759, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.221750497817993, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7220240235328674, + "num_tokens": 527965025.0, + "step": 20405 + }, + { + "epoch": 2.240940039534373, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.444816827774048, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7122942209243774, + "num_tokens": 527989612.0, + "step": 20406 + }, + { + "epoch": 2.2410498572369866, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6806118488311768, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7538679838180542, + "num_tokens": 528009851.0, + "step": 20407 + }, + { + "epoch": 2.2411596749396003, + "ewc_loss": 0.000179290771484375, + "grad_norm": 1.9813803434371948, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7133695483207703, + "num_tokens": 528043087.0, + "step": 20408 + }, + { + "epoch": 2.241269492642214, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.212444305419922, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7082650065422058, + "num_tokens": 528072194.0, + "step": 20409 + }, + { + "epoch": 2.2413793103448274, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3269450664520264, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7142599821090698, + "num_tokens": 528098834.0, + "step": 20410 + }, + { + "epoch": 2.241489128047441, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5154778957366943, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7384876608848572, + "num_tokens": 528122187.0, + "step": 20411 + }, + { + "epoch": 2.241598945750055, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.0851638317108154, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7073574662208557, + "num_tokens": 528154044.0, + "step": 20412 + }, + { + "epoch": 2.2417087634526687, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5086758136749268, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7272650599479675, + "num_tokens": 528175938.0, + "step": 20413 + }, + { + "epoch": 2.2418185811552824, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2300164699554443, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7065314054489136, + "num_tokens": 528204258.0, + "step": 20414 + }, + { + "epoch": 2.2419283988578957, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.10028338432312, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.718791127204895, + "num_tokens": 528237244.0, + "step": 20415 + }, + { + "epoch": 2.2420382165605095, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3429617881774902, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7171803712844849, + "num_tokens": 528264249.0, + "step": 20416 + }, + { + "epoch": 2.2421480342631233, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.7213902473449707, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7363470792770386, + "num_tokens": 528284839.0, + "step": 20417 + }, + { + "epoch": 2.242257851965737, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4178478717803955, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7194088697433472, + "num_tokens": 528310191.0, + "step": 20418 + }, + { + "epoch": 2.2423676696683508, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3310811519622803, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7277373671531677, + "num_tokens": 528338473.0, + "step": 20419 + }, + { + "epoch": 2.242477487370964, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.110198736190796, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7233439087867737, + "num_tokens": 528368942.0, + "step": 20420 + }, + { + "epoch": 2.242587305073578, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4660861492156982, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.72041255235672, + "num_tokens": 528394289.0, + "step": 20421 + }, + { + "epoch": 2.2426971227761916, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2023465633392334, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6944557428359985, + "num_tokens": 528424723.0, + "step": 20422 + }, + { + "epoch": 2.2428069404788054, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.329421043395996, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7063537240028381, + "num_tokens": 528451787.0, + "step": 20423 + }, + { + "epoch": 2.2429167581814187, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1315789222717285, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7124876379966736, + "num_tokens": 528484557.0, + "step": 20424 + }, + { + "epoch": 2.2430265758840324, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3908233642578125, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7171969413757324, + "num_tokens": 528511132.0, + "step": 20425 + }, + { + "epoch": 2.243136393586646, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2585484981536865, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7279341816902161, + "num_tokens": 528539163.0, + "step": 20426 + }, + { + "epoch": 2.24324621128926, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2662665843963623, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7375599145889282, + "num_tokens": 528567797.0, + "step": 20427 + }, + { + "epoch": 2.2433560289918737, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2408981323242188, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7136354446411133, + "num_tokens": 528595182.0, + "step": 20428 + }, + { + "epoch": 2.243465846694487, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3509867191314697, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7403526306152344, + "num_tokens": 528620815.0, + "step": 20429 + }, + { + "epoch": 2.2435756643971008, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.09169340133667, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7174030542373657, + "num_tokens": 528653451.0, + "step": 20430 + }, + { + "epoch": 2.2436854820997145, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.1441402435302734, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7203354835510254, + "num_tokens": 528682954.0, + "step": 20431 + }, + { + "epoch": 2.2437952998023283, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2985236644744873, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7104251384735107, + "num_tokens": 528713588.0, + "step": 20432 + }, + { + "epoch": 2.2439051175049416, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.320113182067871, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7441792488098145, + "num_tokens": 528738256.0, + "step": 20433 + }, + { + "epoch": 2.2440149352075554, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3263497352600098, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7328158020973206, + "num_tokens": 528764360.0, + "step": 20434 + }, + { + "epoch": 2.244124752910169, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3287124633789062, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7253909111022949, + "num_tokens": 528791968.0, + "step": 20435 + }, + { + "epoch": 2.244234570612783, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2993481159210205, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7166889905929565, + "num_tokens": 528818306.0, + "step": 20436 + }, + { + "epoch": 2.2443443883153966, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.441161632537842, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7417250275611877, + "num_tokens": 528844044.0, + "step": 20437 + }, + { + "epoch": 2.24445420601801, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.385749340057373, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.749019980430603, + "num_tokens": 528869218.0, + "step": 20438 + }, + { + "epoch": 2.2445640237206237, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.285794258117676, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7403610348701477, + "num_tokens": 528895694.0, + "step": 20439 + }, + { + "epoch": 2.2446738414232374, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.429729700088501, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7142236232757568, + "num_tokens": 528924203.0, + "step": 20440 + }, + { + "epoch": 2.244783659125851, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3840432167053223, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7305353879928589, + "num_tokens": 528948436.0, + "step": 20441 + }, + { + "epoch": 2.244893476828465, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.636406421661377, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7452436685562134, + "num_tokens": 528969233.0, + "step": 20442 + }, + { + "epoch": 2.2450032945310783, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3742737770080566, + "learning_rate": 1e-06, + "loss": 0.7565, + "mean_token_accuracy": 0.7588914632797241, + "num_tokens": 528993458.0, + "step": 20443 + }, + { + "epoch": 2.245113112233692, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.4436562061309814, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7341402769088745, + "num_tokens": 529019352.0, + "step": 20444 + }, + { + "epoch": 2.245222929936306, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.5033249855041504, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7226258516311646, + "num_tokens": 529043241.0, + "step": 20445 + }, + { + "epoch": 2.2453327476389195, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.313422918319702, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7212404012680054, + "num_tokens": 529071459.0, + "step": 20446 + }, + { + "epoch": 2.245442565341533, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.2833452224731445, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7264678478240967, + "num_tokens": 529097966.0, + "step": 20447 + }, + { + "epoch": 2.2455523830441466, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.419200897216797, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7179845571517944, + "num_tokens": 529123413.0, + "step": 20448 + }, + { + "epoch": 2.2456622007467604, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.3420443534851074, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.709775447845459, + "num_tokens": 529151037.0, + "step": 20449 + }, + { + "epoch": 2.245772018449374, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5621261596679688, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7342665195465088, + "num_tokens": 529173773.0, + "step": 20450 + }, + { + "epoch": 2.245881836151988, + "ewc_loss": 0.000179290771484375, + "grad_norm": 2.6063246726989746, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.7511401772499084, + "num_tokens": 529195203.0, + "step": 20451 + }, + { + "epoch": 2.245991653854601, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.433197259902954, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.7465057969093323, + "num_tokens": 529218317.0, + "step": 20452 + }, + { + "epoch": 2.246101471557215, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4021575450897217, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7430843114852905, + "num_tokens": 529242615.0, + "step": 20453 + }, + { + "epoch": 2.2462112892598287, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.355886936187744, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7276221513748169, + "num_tokens": 529268984.0, + "step": 20454 + }, + { + "epoch": 2.2463211069624425, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3651676177978516, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7360073328018188, + "num_tokens": 529293638.0, + "step": 20455 + }, + { + "epoch": 2.246430924665056, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.507960319519043, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7342706918716431, + "num_tokens": 529317290.0, + "step": 20456 + }, + { + "epoch": 2.2465407423676695, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.785672903060913, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.745761513710022, + "num_tokens": 529336431.0, + "step": 20457 + }, + { + "epoch": 2.2466505600702833, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.179145097732544, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7042087316513062, + "num_tokens": 529366266.0, + "step": 20458 + }, + { + "epoch": 2.246760377772897, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.8513615131378174, + "learning_rate": 1e-06, + "loss": 0.7814, + "mean_token_accuracy": 0.7508845329284668, + "num_tokens": 529383851.0, + "step": 20459 + }, + { + "epoch": 2.246870195475511, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3879520893096924, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7334741353988647, + "num_tokens": 529409177.0, + "step": 20460 + }, + { + "epoch": 2.246980013178124, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1136913299560547, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7264205813407898, + "num_tokens": 529439888.0, + "step": 20461 + }, + { + "epoch": 2.247089830880738, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5234105587005615, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7495702505111694, + "num_tokens": 529462820.0, + "step": 20462 + }, + { + "epoch": 2.2471996485833516, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.476710796356201, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7346680760383606, + "num_tokens": 529486913.0, + "step": 20463 + }, + { + "epoch": 2.2473094662859654, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.239960193634033, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7375228404998779, + "num_tokens": 529513992.0, + "step": 20464 + }, + { + "epoch": 2.247419283988579, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3608758449554443, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7377854585647583, + "num_tokens": 529540114.0, + "step": 20465 + }, + { + "epoch": 2.2475291016911925, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2554969787597656, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7321521043777466, + "num_tokens": 529566298.0, + "step": 20466 + }, + { + "epoch": 2.2476389193938062, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.399698495864868, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7277370691299438, + "num_tokens": 529591700.0, + "step": 20467 + }, + { + "epoch": 2.24774873709642, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.741670846939087, + "learning_rate": 1e-06, + "loss": 0.7202, + "mean_token_accuracy": 0.7675628662109375, + "num_tokens": 529612780.0, + "step": 20468 + }, + { + "epoch": 2.2478585547990337, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1874845027923584, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7158584594726562, + "num_tokens": 529641782.0, + "step": 20469 + }, + { + "epoch": 2.2479683725016475, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.267970085144043, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7284702658653259, + "num_tokens": 529668437.0, + "step": 20470 + }, + { + "epoch": 2.248078190204261, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5147948265075684, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7155343890190125, + "num_tokens": 529692229.0, + "step": 20471 + }, + { + "epoch": 2.2481880079068746, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.078075885772705, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7361118793487549, + "num_tokens": 529722955.0, + "step": 20472 + }, + { + "epoch": 2.2482978256094883, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.27970290184021, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7381648421287537, + "num_tokens": 529750229.0, + "step": 20473 + }, + { + "epoch": 2.248407643312102, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3023858070373535, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7220516204833984, + "num_tokens": 529777962.0, + "step": 20474 + }, + { + "epoch": 2.2485174610147154, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2730283737182617, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7295442819595337, + "num_tokens": 529805968.0, + "step": 20475 + }, + { + "epoch": 2.248627278717329, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.517355442047119, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7518898248672485, + "num_tokens": 529828306.0, + "step": 20476 + }, + { + "epoch": 2.248737096419943, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3538269996643066, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7212457656860352, + "num_tokens": 529855477.0, + "step": 20477 + }, + { + "epoch": 2.2488469141225567, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.55061936378479, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7476820349693298, + "num_tokens": 529879914.0, + "step": 20478 + }, + { + "epoch": 2.2489567318251704, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5082740783691406, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7392807006835938, + "num_tokens": 529902911.0, + "step": 20479 + }, + { + "epoch": 2.2490665495277837, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.0941035747528076, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7071664929389954, + "num_tokens": 529936411.0, + "step": 20480 + }, + { + "epoch": 2.2491763672303975, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.602692127227783, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7362158298492432, + "num_tokens": 529957626.0, + "step": 20481 + }, + { + "epoch": 2.2492861849330112, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5896074771881104, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7278910279273987, + "num_tokens": 529981233.0, + "step": 20482 + }, + { + "epoch": 2.249396002635625, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3497300148010254, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7340863943099976, + "num_tokens": 530008794.0, + "step": 20483 + }, + { + "epoch": 2.2495058203382383, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.248737335205078, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7227082252502441, + "num_tokens": 530036435.0, + "step": 20484 + }, + { + "epoch": 2.249615638040852, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3226871490478516, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7353460192680359, + "num_tokens": 530060470.0, + "step": 20485 + }, + { + "epoch": 2.249725455743466, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2381319999694824, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7179995775222778, + "num_tokens": 530090529.0, + "step": 20486 + }, + { + "epoch": 2.2498352734460796, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.338081121444702, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7369912266731262, + "num_tokens": 530117324.0, + "step": 20487 + }, + { + "epoch": 2.2499450911486933, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4472737312316895, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7367206811904907, + "num_tokens": 530139881.0, + "step": 20488 + }, + { + "epoch": 2.2500549088513067, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7522666454315186, + "learning_rate": 1e-06, + "loss": 0.7217, + "mean_token_accuracy": 0.7586532831192017, + "num_tokens": 530159955.0, + "step": 20489 + }, + { + "epoch": 2.2501647265539204, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.444347858428955, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7263184189796448, + "num_tokens": 530183815.0, + "step": 20490 + }, + { + "epoch": 2.250274544256534, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3922030925750732, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7324785590171814, + "num_tokens": 530209864.0, + "step": 20491 + }, + { + "epoch": 2.250384361959148, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4227230548858643, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.738764762878418, + "num_tokens": 530236326.0, + "step": 20492 + }, + { + "epoch": 2.2504941796617617, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.640960693359375, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7348963022232056, + "num_tokens": 530256498.0, + "step": 20493 + }, + { + "epoch": 2.250603997364375, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.41831374168396, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7306300401687622, + "num_tokens": 530280928.0, + "step": 20494 + }, + { + "epoch": 2.2507138150669888, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.60392165184021, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.731741189956665, + "num_tokens": 530301881.0, + "step": 20495 + }, + { + "epoch": 2.2508236327696025, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.214160919189453, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7038670182228088, + "num_tokens": 530333726.0, + "step": 20496 + }, + { + "epoch": 2.2509334504722163, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3716886043548584, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7445696592330933, + "num_tokens": 530360414.0, + "step": 20497 + }, + { + "epoch": 2.25104326817483, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2951650619506836, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7124197483062744, + "num_tokens": 530387789.0, + "step": 20498 + }, + { + "epoch": 2.2511530858774433, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5779497623443604, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7241219282150269, + "num_tokens": 530410661.0, + "step": 20499 + }, + { + "epoch": 2.251262903580057, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.232830286026001, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7248647809028625, + "num_tokens": 530438675.0, + "step": 20500 + }, + { + "epoch": 2.251372721282671, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.262554407119751, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7281051874160767, + "num_tokens": 530466771.0, + "step": 20501 + }, + { + "epoch": 2.2514825389852846, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3998517990112305, + "learning_rate": 1e-06, + "loss": 0.7321, + "mean_token_accuracy": 0.7637693881988525, + "num_tokens": 530489353.0, + "step": 20502 + }, + { + "epoch": 2.251592356687898, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4883697032928467, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7264851331710815, + "num_tokens": 530514382.0, + "step": 20503 + }, + { + "epoch": 2.2517021743905117, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.390300750732422, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7196052074432373, + "num_tokens": 530539874.0, + "step": 20504 + }, + { + "epoch": 2.2518119920931254, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4749226570129395, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7373839020729065, + "num_tokens": 530564468.0, + "step": 20505 + }, + { + "epoch": 2.251921809795739, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.162503719329834, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.742970883846283, + "num_tokens": 530592630.0, + "step": 20506 + }, + { + "epoch": 2.2520316274983525, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2789483070373535, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7191569209098816, + "num_tokens": 530621981.0, + "step": 20507 + }, + { + "epoch": 2.2521414452009663, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5430407524108887, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.737193763256073, + "num_tokens": 530646013.0, + "step": 20508 + }, + { + "epoch": 2.25225126290358, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.496222972869873, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.741712749004364, + "num_tokens": 530669801.0, + "step": 20509 + }, + { + "epoch": 2.2523610806061938, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2955706119537354, + "learning_rate": 1e-06, + "loss": 0.7613, + "mean_token_accuracy": 0.764772891998291, + "num_tokens": 530695794.0, + "step": 20510 + }, + { + "epoch": 2.2524708983088075, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2388246059417725, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7263146042823792, + "num_tokens": 530725142.0, + "step": 20511 + }, + { + "epoch": 2.252580716011421, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1930155754089355, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7387125492095947, + "num_tokens": 530754745.0, + "step": 20512 + }, + { + "epoch": 2.2526905337140346, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5292437076568604, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7292752265930176, + "num_tokens": 530778615.0, + "step": 20513 + }, + { + "epoch": 2.2528003514166484, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3120155334472656, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.740839421749115, + "num_tokens": 530806303.0, + "step": 20514 + }, + { + "epoch": 2.252910169119262, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.142674207687378, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7305600643157959, + "num_tokens": 530836596.0, + "step": 20515 + }, + { + "epoch": 2.253019986821876, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3445584774017334, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7258421778678894, + "num_tokens": 530862211.0, + "step": 20516 + }, + { + "epoch": 2.253129804524489, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2085988521575928, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7075831294059753, + "num_tokens": 530892342.0, + "step": 20517 + }, + { + "epoch": 2.253239622227103, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1584970951080322, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7000889778137207, + "num_tokens": 530924070.0, + "step": 20518 + }, + { + "epoch": 2.2533494399297167, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2480416297912598, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7043546438217163, + "num_tokens": 530953116.0, + "step": 20519 + }, + { + "epoch": 2.2534592576323305, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2022111415863037, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7308570146560669, + "num_tokens": 530981887.0, + "step": 20520 + }, + { + "epoch": 2.253569075334944, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2197351455688477, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7325728535652161, + "num_tokens": 531009953.0, + "step": 20521 + }, + { + "epoch": 2.2536788930375575, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.0635242462158203, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7276567220687866, + "num_tokens": 531042200.0, + "step": 20522 + }, + { + "epoch": 2.2537887107401713, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2101969718933105, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.749006986618042, + "num_tokens": 531071165.0, + "step": 20523 + }, + { + "epoch": 2.253898528442785, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4608709812164307, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.745850682258606, + "num_tokens": 531095433.0, + "step": 20524 + }, + { + "epoch": 2.254008346145399, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3526060581207275, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7270557880401611, + "num_tokens": 531120481.0, + "step": 20525 + }, + { + "epoch": 2.254118163848012, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.6560075283050537, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7441279888153076, + "num_tokens": 531143677.0, + "step": 20526 + }, + { + "epoch": 2.254227981550626, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2105319499969482, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7141175866127014, + "num_tokens": 531173405.0, + "step": 20527 + }, + { + "epoch": 2.2543377992532396, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3977396488189697, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7087452411651611, + "num_tokens": 531200292.0, + "step": 20528 + }, + { + "epoch": 2.2544476169558534, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5096046924591064, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7334502339363098, + "num_tokens": 531223427.0, + "step": 20529 + }, + { + "epoch": 2.2545574346584667, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.40541934967041, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7161873579025269, + "num_tokens": 531248597.0, + "step": 20530 + }, + { + "epoch": 2.2546672523610805, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5562338829040527, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7439722418785095, + "num_tokens": 531269512.0, + "step": 20531 + }, + { + "epoch": 2.254777070063694, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2243332862854004, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.720881462097168, + "num_tokens": 531298178.0, + "step": 20532 + }, + { + "epoch": 2.254886887766308, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.259847402572632, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7380648851394653, + "num_tokens": 531325600.0, + "step": 20533 + }, + { + "epoch": 2.2549967054689217, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.377332925796509, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.727181613445282, + "num_tokens": 531351729.0, + "step": 20534 + }, + { + "epoch": 2.255106523171535, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5022387504577637, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7347121238708496, + "num_tokens": 531376139.0, + "step": 20535 + }, + { + "epoch": 2.255216340874149, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.482150077819824, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7195461392402649, + "num_tokens": 531402112.0, + "step": 20536 + }, + { + "epoch": 2.2553261585767626, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1730692386627197, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7192858457565308, + "num_tokens": 531431035.0, + "step": 20537 + }, + { + "epoch": 2.2554359762793763, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2254998683929443, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7220281958580017, + "num_tokens": 531459968.0, + "step": 20538 + }, + { + "epoch": 2.25554579398199, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5304110050201416, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7180132865905762, + "num_tokens": 531482442.0, + "step": 20539 + }, + { + "epoch": 2.2556556116846034, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.386298179626465, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7445981502532959, + "num_tokens": 531506980.0, + "step": 20540 + }, + { + "epoch": 2.255765429387217, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.201969861984253, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7345337271690369, + "num_tokens": 531536863.0, + "step": 20541 + }, + { + "epoch": 2.255875247089831, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4614572525024414, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7404721975326538, + "num_tokens": 531561662.0, + "step": 20542 + }, + { + "epoch": 2.2559850647924446, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.6665892601013184, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7501100301742554, + "num_tokens": 531583732.0, + "step": 20543 + }, + { + "epoch": 2.2560948824950584, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.211834669113159, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7542803287506104, + "num_tokens": 531611244.0, + "step": 20544 + }, + { + "epoch": 2.2562047001976717, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.9537549018859863, + "learning_rate": 1e-06, + "loss": 0.7335, + "mean_token_accuracy": 0.7639438509941101, + "num_tokens": 531629118.0, + "step": 20545 + }, + { + "epoch": 2.2563145179002855, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.101388692855835, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6933225393295288, + "num_tokens": 531660232.0, + "step": 20546 + }, + { + "epoch": 2.2564243356028992, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.192873239517212, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.738433837890625, + "num_tokens": 531691246.0, + "step": 20547 + }, + { + "epoch": 2.256534153305513, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 1.9420342445373535, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7092984914779663, + "num_tokens": 531727836.0, + "step": 20548 + }, + { + "epoch": 2.2566439710081267, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7254953384399414, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.743472695350647, + "num_tokens": 531748360.0, + "step": 20549 + }, + { + "epoch": 2.25675378871074, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1202502250671387, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.6974731683731079, + "num_tokens": 531782369.0, + "step": 20550 + }, + { + "epoch": 2.256863606413354, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7128477096557617, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7177190780639648, + "num_tokens": 531803350.0, + "step": 20551 + }, + { + "epoch": 2.2569734241159676, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.360499858856201, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7265434265136719, + "num_tokens": 531830245.0, + "step": 20552 + }, + { + "epoch": 2.2570832418185813, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2565555572509766, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7367583513259888, + "num_tokens": 531857075.0, + "step": 20553 + }, + { + "epoch": 2.2571930595211946, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1271238327026367, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7205948233604431, + "num_tokens": 531887055.0, + "step": 20554 + }, + { + "epoch": 2.2573028772238084, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.338151454925537, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7214602828025818, + "num_tokens": 531914390.0, + "step": 20555 + }, + { + "epoch": 2.257412694926422, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1628105640411377, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7291683554649353, + "num_tokens": 531944376.0, + "step": 20556 + }, + { + "epoch": 2.257522512629036, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.392367362976074, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7305583953857422, + "num_tokens": 531970531.0, + "step": 20557 + }, + { + "epoch": 2.2576323303316492, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3727807998657227, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7386919260025024, + "num_tokens": 531997450.0, + "step": 20558 + }, + { + "epoch": 2.257742148034263, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2692794799804688, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7386143207550049, + "num_tokens": 532023100.0, + "step": 20559 + }, + { + "epoch": 2.2578519657368767, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2690796852111816, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7287995219230652, + "num_tokens": 532049239.0, + "step": 20560 + }, + { + "epoch": 2.2579617834394905, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.031999349594116, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7254674434661865, + "num_tokens": 532083272.0, + "step": 20561 + }, + { + "epoch": 2.2580716011421043, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.155292510986328, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7247656583786011, + "num_tokens": 532112550.0, + "step": 20562 + }, + { + "epoch": 2.2581814188447176, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1314828395843506, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7251623868942261, + "num_tokens": 532141991.0, + "step": 20563 + }, + { + "epoch": 2.2582912365473313, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1808502674102783, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7379197478294373, + "num_tokens": 532169669.0, + "step": 20564 + }, + { + "epoch": 2.258401054249945, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.562061071395874, + "learning_rate": 1e-06, + "loss": 0.6698, + "mean_token_accuracy": 0.7833056449890137, + "num_tokens": 532189423.0, + "step": 20565 + }, + { + "epoch": 2.258510871952559, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5390732288360596, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7308058142662048, + "num_tokens": 532213736.0, + "step": 20566 + }, + { + "epoch": 2.2586206896551726, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.179868221282959, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7066546082496643, + "num_tokens": 532242845.0, + "step": 20567 + }, + { + "epoch": 2.258730507357786, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2791671752929688, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7436912655830383, + "num_tokens": 532269830.0, + "step": 20568 + }, + { + "epoch": 2.2588403250603997, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4099583625793457, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7441058158874512, + "num_tokens": 532294985.0, + "step": 20569 + }, + { + "epoch": 2.2589501427630134, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2461233139038086, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7475007772445679, + "num_tokens": 532322489.0, + "step": 20570 + }, + { + "epoch": 2.259059960465627, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.451714515686035, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7287395000457764, + "num_tokens": 532346418.0, + "step": 20571 + }, + { + "epoch": 2.259169778168241, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5482678413391113, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7470122575759888, + "num_tokens": 532368992.0, + "step": 20572 + }, + { + "epoch": 2.2592795958708543, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5171971321105957, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7180700898170471, + "num_tokens": 532394273.0, + "step": 20573 + }, + { + "epoch": 2.259389413573468, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7294788360595703, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7277837991714478, + "num_tokens": 532416996.0, + "step": 20574 + }, + { + "epoch": 2.2594992312760818, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4998834133148193, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7314383387565613, + "num_tokens": 532442497.0, + "step": 20575 + }, + { + "epoch": 2.2596090489786955, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4229936599731445, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7217103242874146, + "num_tokens": 532469051.0, + "step": 20576 + }, + { + "epoch": 2.259718866681309, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4126477241516113, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7311416864395142, + "num_tokens": 532495620.0, + "step": 20577 + }, + { + "epoch": 2.2598286843839226, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5340304374694824, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7401319742202759, + "num_tokens": 532518777.0, + "step": 20578 + }, + { + "epoch": 2.2599385020865363, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3998727798461914, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7471691370010376, + "num_tokens": 532543082.0, + "step": 20579 + }, + { + "epoch": 2.26004831978915, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.192949056625366, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7227271795272827, + "num_tokens": 532571941.0, + "step": 20580 + }, + { + "epoch": 2.260158137491764, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.382603168487549, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7357706427574158, + "num_tokens": 532596864.0, + "step": 20581 + }, + { + "epoch": 2.260267955194377, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.268458366394043, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7156836986541748, + "num_tokens": 532624412.0, + "step": 20582 + }, + { + "epoch": 2.260377772896991, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.619561195373535, + "learning_rate": 1e-06, + "loss": 0.7556, + "mean_token_accuracy": 0.7534990310668945, + "num_tokens": 532645362.0, + "step": 20583 + }, + { + "epoch": 2.2604875905996047, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.8432819843292236, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7346962690353394, + "num_tokens": 532665259.0, + "step": 20584 + }, + { + "epoch": 2.2605974083022184, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.479313373565674, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7165257930755615, + "num_tokens": 532688492.0, + "step": 20585 + }, + { + "epoch": 2.2607072260048318, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.547266721725464, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7412105798721313, + "num_tokens": 532712808.0, + "step": 20586 + }, + { + "epoch": 2.2608170437074455, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.198718547821045, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7401160001754761, + "num_tokens": 532741960.0, + "step": 20587 + }, + { + "epoch": 2.2609268614100593, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.0445501804351807, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7262464761734009, + "num_tokens": 532775010.0, + "step": 20588 + }, + { + "epoch": 2.261036679112673, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.397052049636841, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7298696637153625, + "num_tokens": 532799610.0, + "step": 20589 + }, + { + "epoch": 2.261146496815287, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5372314453125, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7394382953643799, + "num_tokens": 532822318.0, + "step": 20590 + }, + { + "epoch": 2.2612563145179, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.356813907623291, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7429900765419006, + "num_tokens": 532848644.0, + "step": 20591 + }, + { + "epoch": 2.261366132220514, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.345590591430664, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7185028195381165, + "num_tokens": 532874857.0, + "step": 20592 + }, + { + "epoch": 2.2614759499231276, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.348900556564331, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7350143790245056, + "num_tokens": 532901140.0, + "step": 20593 + }, + { + "epoch": 2.2615857676257414, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1175358295440674, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6982553005218506, + "num_tokens": 532933633.0, + "step": 20594 + }, + { + "epoch": 2.261695585328355, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.387789249420166, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7192389369010925, + "num_tokens": 532960356.0, + "step": 20595 + }, + { + "epoch": 2.2618054030309684, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.6642158031463623, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7231506705284119, + "num_tokens": 532982102.0, + "step": 20596 + }, + { + "epoch": 2.261915220733582, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.353996753692627, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7385942339897156, + "num_tokens": 533008657.0, + "step": 20597 + }, + { + "epoch": 2.262025038436196, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.012420177459717, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7153480648994446, + "num_tokens": 533040050.0, + "step": 20598 + }, + { + "epoch": 2.2621348561388097, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3951687812805176, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7172234654426575, + "num_tokens": 533065654.0, + "step": 20599 + }, + { + "epoch": 2.2622446738414235, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.452878475189209, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7467012405395508, + "num_tokens": 533089382.0, + "step": 20600 + }, + { + "epoch": 2.262354491544037, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1882872581481934, + "learning_rate": 1e-06, + "loss": 1.0738, + "mean_token_accuracy": 0.6979909539222717, + "num_tokens": 533119669.0, + "step": 20601 + }, + { + "epoch": 2.2624643092466505, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7446188926696777, + "learning_rate": 1e-06, + "loss": 0.7127, + "mean_token_accuracy": 0.7744511365890503, + "num_tokens": 533138406.0, + "step": 20602 + }, + { + "epoch": 2.2625741269492643, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.243640184402466, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.738145112991333, + "num_tokens": 533165108.0, + "step": 20603 + }, + { + "epoch": 2.262683944651878, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1499576568603516, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7292313575744629, + "num_tokens": 533196382.0, + "step": 20604 + }, + { + "epoch": 2.2627937623544914, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.375518560409546, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7293000221252441, + "num_tokens": 533221981.0, + "step": 20605 + }, + { + "epoch": 2.262903580057105, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.462946653366089, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7157891392707825, + "num_tokens": 533248532.0, + "step": 20606 + }, + { + "epoch": 2.263013397759719, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1699905395507812, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7210913300514221, + "num_tokens": 533276848.0, + "step": 20607 + }, + { + "epoch": 2.2631232154623326, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.121951103210449, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7044832706451416, + "num_tokens": 533309991.0, + "step": 20608 + }, + { + "epoch": 2.263233033164946, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3690168857574463, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7412815690040588, + "num_tokens": 533337717.0, + "step": 20609 + }, + { + "epoch": 2.2633428508675597, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4272868633270264, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7245814204216003, + "num_tokens": 533361471.0, + "step": 20610 + }, + { + "epoch": 2.2634526685701735, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.298661231994629, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7334562540054321, + "num_tokens": 533390149.0, + "step": 20611 + }, + { + "epoch": 2.263562486272787, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4549760818481445, + "learning_rate": 1e-06, + "loss": 0.7445, + "mean_token_accuracy": 0.7610552310943604, + "num_tokens": 533412367.0, + "step": 20612 + }, + { + "epoch": 2.263672303975401, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.381150007247925, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7325187921524048, + "num_tokens": 533439118.0, + "step": 20613 + }, + { + "epoch": 2.2637821216780143, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.49062180519104, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7157208323478699, + "num_tokens": 533465411.0, + "step": 20614 + }, + { + "epoch": 2.263891939380628, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4870049953460693, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7268486618995667, + "num_tokens": 533491447.0, + "step": 20615 + }, + { + "epoch": 2.264001757083242, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.6255435943603516, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7423025369644165, + "num_tokens": 533513943.0, + "step": 20616 + }, + { + "epoch": 2.2641115747858556, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4590795040130615, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7104662656784058, + "num_tokens": 533541238.0, + "step": 20617 + }, + { + "epoch": 2.2642213924884693, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3256325721740723, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7422662377357483, + "num_tokens": 533565011.0, + "step": 20618 + }, + { + "epoch": 2.2643312101910826, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.0769574642181396, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7386002540588379, + "num_tokens": 533598006.0, + "step": 20619 + }, + { + "epoch": 2.2644410278936964, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.456207752227783, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.707679808139801, + "num_tokens": 533623268.0, + "step": 20620 + }, + { + "epoch": 2.26455084559631, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1816368103027344, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7304506301879883, + "num_tokens": 533652105.0, + "step": 20621 + }, + { + "epoch": 2.264660663298924, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.8272151947021484, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7350342273712158, + "num_tokens": 533670980.0, + "step": 20622 + }, + { + "epoch": 2.2647704810015377, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.442337989807129, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.724729061126709, + "num_tokens": 533697461.0, + "step": 20623 + }, + { + "epoch": 2.264880298704151, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.272775650024414, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7284265756607056, + "num_tokens": 533723770.0, + "step": 20624 + }, + { + "epoch": 2.2649901164067647, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.1054937839508057, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7401552200317383, + "num_tokens": 533752403.0, + "step": 20625 + }, + { + "epoch": 2.2650999341093785, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3932008743286133, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7348071336746216, + "num_tokens": 533777072.0, + "step": 20626 + }, + { + "epoch": 2.2652097518119922, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.645594358444214, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7428251504898071, + "num_tokens": 533800909.0, + "step": 20627 + }, + { + "epoch": 2.265319569514606, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.4305877685546875, + "learning_rate": 1e-06, + "loss": 0.7686, + "mean_token_accuracy": 0.7560081481933594, + "num_tokens": 533825942.0, + "step": 20628 + }, + { + "epoch": 2.2654293872172193, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3031129837036133, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7455294132232666, + "num_tokens": 533851246.0, + "step": 20629 + }, + { + "epoch": 2.265539204919833, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7386984825134277, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7472771406173706, + "num_tokens": 533871288.0, + "step": 20630 + }, + { + "epoch": 2.265649022622447, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.3498523235321045, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7157009840011597, + "num_tokens": 533897718.0, + "step": 20631 + }, + { + "epoch": 2.2657588403250606, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4794113636016846, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.74259352684021, + "num_tokens": 533919924.0, + "step": 20632 + }, + { + "epoch": 2.265868658027674, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.354539155960083, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7238442301750183, + "num_tokens": 533945249.0, + "step": 20633 + }, + { + "epoch": 2.2659784757302877, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.282384157180786, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7087175250053406, + "num_tokens": 533973516.0, + "step": 20634 + }, + { + "epoch": 2.2660882934329014, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4825499057769775, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7484170794487, + "num_tokens": 533997573.0, + "step": 20635 + }, + { + "epoch": 2.266198111135515, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.829350709915161, + "learning_rate": 1e-06, + "loss": 0.7432, + "mean_token_accuracy": 0.7583214044570923, + "num_tokens": 534016343.0, + "step": 20636 + }, + { + "epoch": 2.2663079288381285, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.166841983795166, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7229630947113037, + "num_tokens": 534046554.0, + "step": 20637 + }, + { + "epoch": 2.2664177465407422, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5313665866851807, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7391351461410522, + "num_tokens": 534070506.0, + "step": 20638 + }, + { + "epoch": 2.266527564243356, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.2492568492889404, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7153708934783936, + "num_tokens": 534098458.0, + "step": 20639 + }, + { + "epoch": 2.2666373819459698, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.7287635803222656, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7464351654052734, + "num_tokens": 534119676.0, + "step": 20640 + }, + { + "epoch": 2.2667471996485835, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3056952953338623, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7558226585388184, + "num_tokens": 534147125.0, + "step": 20641 + }, + { + "epoch": 2.266857017351197, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.280460834503174, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6944767832756042, + "num_tokens": 534175831.0, + "step": 20642 + }, + { + "epoch": 2.2669668350538106, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.327321767807007, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7459667921066284, + "num_tokens": 534202981.0, + "step": 20643 + }, + { + "epoch": 2.2670766527564243, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.3641722202301025, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7133923172950745, + "num_tokens": 534232202.0, + "step": 20644 + }, + { + "epoch": 2.267186470459038, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.5238327980041504, + "learning_rate": 1e-06, + "loss": 0.7846, + "mean_token_accuracy": 0.754022479057312, + "num_tokens": 534256887.0, + "step": 20645 + }, + { + "epoch": 2.267296288161652, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.282759666442871, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.7450812458992004, + "num_tokens": 534283370.0, + "step": 20646 + }, + { + "epoch": 2.267406105864265, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.57027530670166, + "learning_rate": 1e-06, + "loss": 0.789, + "mean_token_accuracy": 0.7497380971908569, + "num_tokens": 534305855.0, + "step": 20647 + }, + { + "epoch": 2.267515923566879, + "ewc_loss": 0.00018024444580078125, + "grad_norm": 2.929819107055664, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7310190200805664, + "num_tokens": 534324343.0, + "step": 20648 + }, + { + "epoch": 2.2676257412694927, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4365856647491455, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7437151670455933, + "num_tokens": 534349733.0, + "step": 20649 + }, + { + "epoch": 2.2677355589721064, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4360930919647217, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.71898353099823, + "num_tokens": 534376612.0, + "step": 20650 + }, + { + "epoch": 2.26784537667472, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.2234034538269043, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7361769676208496, + "num_tokens": 534405067.0, + "step": 20651 + }, + { + "epoch": 2.2679551943773335, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1800036430358887, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7018948793411255, + "num_tokens": 534434465.0, + "step": 20652 + }, + { + "epoch": 2.2680650120799473, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.6401913166046143, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7257219552993774, + "num_tokens": 534457313.0, + "step": 20653 + }, + { + "epoch": 2.268174829782561, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.331294536590576, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7385975122451782, + "num_tokens": 534483692.0, + "step": 20654 + }, + { + "epoch": 2.2682846474851748, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.48481822013855, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7268715500831604, + "num_tokens": 534506119.0, + "step": 20655 + }, + { + "epoch": 2.268394465187788, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.156883478164673, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7068946361541748, + "num_tokens": 534539490.0, + "step": 20656 + }, + { + "epoch": 2.268504282890402, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.285808563232422, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7593787908554077, + "num_tokens": 534566135.0, + "step": 20657 + }, + { + "epoch": 2.2686141005930156, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4532101154327393, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7435140013694763, + "num_tokens": 534591151.0, + "step": 20658 + }, + { + "epoch": 2.2687239182956294, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1829910278320312, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7277055978775024, + "num_tokens": 534620416.0, + "step": 20659 + }, + { + "epoch": 2.2688337359982427, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.514810562133789, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7432538866996765, + "num_tokens": 534643272.0, + "step": 20660 + }, + { + "epoch": 2.2689435537008564, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.2051711082458496, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7161329388618469, + "num_tokens": 534674244.0, + "step": 20661 + }, + { + "epoch": 2.26905337140347, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1646482944488525, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7455191612243652, + "num_tokens": 534702621.0, + "step": 20662 + }, + { + "epoch": 2.269163189106084, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.624387264251709, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7403995990753174, + "num_tokens": 534724554.0, + "step": 20663 + }, + { + "epoch": 2.2692730068086977, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.2986629009246826, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7131060361862183, + "num_tokens": 534754776.0, + "step": 20664 + }, + { + "epoch": 2.269382824511311, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4147837162017822, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7317422032356262, + "num_tokens": 534781638.0, + "step": 20665 + }, + { + "epoch": 2.2694926422139248, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.737448215484619, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7477121353149414, + "num_tokens": 534800711.0, + "step": 20666 + }, + { + "epoch": 2.2696024599165385, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1966488361358643, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7045345902442932, + "num_tokens": 534833402.0, + "step": 20667 + }, + { + "epoch": 2.2697122776191523, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.3724567890167236, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7451101541519165, + "num_tokens": 534858057.0, + "step": 20668 + }, + { + "epoch": 2.269822095321766, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.2604386806488037, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7087811231613159, + "num_tokens": 534884852.0, + "step": 20669 + }, + { + "epoch": 2.2699319130243794, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.6323347091674805, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7379711866378784, + "num_tokens": 534906785.0, + "step": 20670 + }, + { + "epoch": 2.270041730726993, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1585581302642822, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7268843650817871, + "num_tokens": 534939871.0, + "step": 20671 + }, + { + "epoch": 2.270151548429607, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.892220973968506, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.757828414440155, + "num_tokens": 534957889.0, + "step": 20672 + }, + { + "epoch": 2.2702613661322206, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.2683751583099365, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7131874561309814, + "num_tokens": 534988918.0, + "step": 20673 + }, + { + "epoch": 2.2703711838348344, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.406236410140991, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7194526195526123, + "num_tokens": 535016145.0, + "step": 20674 + }, + { + "epoch": 2.2704810015374477, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4940390586853027, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7455978393554688, + "num_tokens": 535040159.0, + "step": 20675 + }, + { + "epoch": 2.2705908192400615, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.1460366249084473, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7478302121162415, + "num_tokens": 535068420.0, + "step": 20676 + }, + { + "epoch": 2.270700636942675, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.834096670150757, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7316133975982666, + "num_tokens": 535086914.0, + "step": 20677 + }, + { + "epoch": 2.270810454645289, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.6209678649902344, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7462591528892517, + "num_tokens": 535107860.0, + "step": 20678 + }, + { + "epoch": 2.2709202723479027, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.3758015632629395, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.737814724445343, + "num_tokens": 535134240.0, + "step": 20679 + }, + { + "epoch": 2.271030090050516, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.6309611797332764, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7158610820770264, + "num_tokens": 535159061.0, + "step": 20680 + }, + { + "epoch": 2.27113990775313, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.387242317199707, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7388550639152527, + "num_tokens": 535183824.0, + "step": 20681 + }, + { + "epoch": 2.2712497254557436, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.516766309738159, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7261959314346313, + "num_tokens": 535209253.0, + "step": 20682 + }, + { + "epoch": 2.2713595431583573, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.5887856483459473, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7222917079925537, + "num_tokens": 535232310.0, + "step": 20683 + }, + { + "epoch": 2.2714693608609706, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.510806083679199, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.738921582698822, + "num_tokens": 535254604.0, + "step": 20684 + }, + { + "epoch": 2.2715791785635844, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.0954179763793945, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7172916531562805, + "num_tokens": 535288578.0, + "step": 20685 + }, + { + "epoch": 2.271688996266198, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.4874818325042725, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.722741425037384, + "num_tokens": 535313776.0, + "step": 20686 + }, + { + "epoch": 2.271798813968812, + "ewc_loss": 0.0001811981201171875, + "grad_norm": 2.132375955581665, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7278485894203186, + "num_tokens": 535343650.0, + "step": 20687 + }, + { + "epoch": 2.271908631671425, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.759242296218872, + "learning_rate": 1e-06, + "loss": 0.7865, + "mean_token_accuracy": 0.7464696168899536, + "num_tokens": 535363387.0, + "step": 20688 + }, + { + "epoch": 2.272018449374039, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6036171913146973, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7559983730316162, + "num_tokens": 535385358.0, + "step": 20689 + }, + { + "epoch": 2.2721282670766527, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.300330638885498, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.697281539440155, + "num_tokens": 535416079.0, + "step": 20690 + }, + { + "epoch": 2.2722380847792665, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4199862480163574, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7398309707641602, + "num_tokens": 535440675.0, + "step": 20691 + }, + { + "epoch": 2.2723479024818802, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.216226577758789, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7205543518066406, + "num_tokens": 535469567.0, + "step": 20692 + }, + { + "epoch": 2.2724577201844935, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.084688663482666, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.703847348690033, + "num_tokens": 535503782.0, + "step": 20693 + }, + { + "epoch": 2.2725675378871073, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3095178604125977, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7470342516899109, + "num_tokens": 535529052.0, + "step": 20694 + }, + { + "epoch": 2.272677355589721, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.864536762237549, + "learning_rate": 1e-06, + "loss": 0.7777, + "mean_token_accuracy": 0.7502859830856323, + "num_tokens": 535550344.0, + "step": 20695 + }, + { + "epoch": 2.272787173292335, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5040323734283447, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7277589440345764, + "num_tokens": 535574011.0, + "step": 20696 + }, + { + "epoch": 2.2728969909949486, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.198673725128174, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7215477228164673, + "num_tokens": 535604376.0, + "step": 20697 + }, + { + "epoch": 2.273006808697562, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2577462196350098, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7259500622749329, + "num_tokens": 535633623.0, + "step": 20698 + }, + { + "epoch": 2.2731166264001756, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4806034564971924, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7218254804611206, + "num_tokens": 535658667.0, + "step": 20699 + }, + { + "epoch": 2.2732264441027894, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.450813055038452, + "learning_rate": 1e-06, + "loss": 0.7651, + "mean_token_accuracy": 0.7611462473869324, + "num_tokens": 535682680.0, + "step": 20700 + }, + { + "epoch": 2.273336261805403, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5638158321380615, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7515594363212585, + "num_tokens": 535704776.0, + "step": 20701 + }, + { + "epoch": 2.273446079508017, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.9277875423431396, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7272855043411255, + "num_tokens": 535731321.0, + "step": 20702 + }, + { + "epoch": 2.2735558972106302, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.849759578704834, + "learning_rate": 1e-06, + "loss": 0.7483, + "mean_token_accuracy": 0.7525787353515625, + "num_tokens": 535749357.0, + "step": 20703 + }, + { + "epoch": 2.273665714913244, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.221895694732666, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7451709508895874, + "num_tokens": 535776414.0, + "step": 20704 + }, + { + "epoch": 2.2737755326158577, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.550889015197754, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7089866995811462, + "num_tokens": 535800583.0, + "step": 20705 + }, + { + "epoch": 2.2738853503184715, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.280308246612549, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7265985012054443, + "num_tokens": 535827948.0, + "step": 20706 + }, + { + "epoch": 2.273995168021085, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.281916618347168, + "learning_rate": 1e-06, + "loss": 0.7792, + "mean_token_accuracy": 0.748503565788269, + "num_tokens": 535855953.0, + "step": 20707 + }, + { + "epoch": 2.2741049857236986, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5751633644104004, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7519967555999756, + "num_tokens": 535878268.0, + "step": 20708 + }, + { + "epoch": 2.2742148034263123, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.267439842224121, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7302602529525757, + "num_tokens": 535906332.0, + "step": 20709 + }, + { + "epoch": 2.274324621128926, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.618212938308716, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7288897037506104, + "num_tokens": 535930170.0, + "step": 20710 + }, + { + "epoch": 2.2744344388315394, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.437598705291748, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7325024604797363, + "num_tokens": 535956811.0, + "step": 20711 + }, + { + "epoch": 2.274544256534153, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.211838722229004, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.731389045715332, + "num_tokens": 535987062.0, + "step": 20712 + }, + { + "epoch": 2.274654074236767, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.223660945892334, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7457050085067749, + "num_tokens": 536016727.0, + "step": 20713 + }, + { + "epoch": 2.2747638919393807, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4795947074890137, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7263227701187134, + "num_tokens": 536040280.0, + "step": 20714 + }, + { + "epoch": 2.2748737096419944, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4011569023132324, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7197376489639282, + "num_tokens": 536067056.0, + "step": 20715 + }, + { + "epoch": 2.2749835273446077, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6029844284057617, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7130504846572876, + "num_tokens": 536090216.0, + "step": 20716 + }, + { + "epoch": 2.2750933450472215, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5228917598724365, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7417147159576416, + "num_tokens": 536113676.0, + "step": 20717 + }, + { + "epoch": 2.2752031627498353, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.38103985786438, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7426552772521973, + "num_tokens": 536140983.0, + "step": 20718 + }, + { + "epoch": 2.275312980452449, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.512472629547119, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7101466655731201, + "num_tokens": 536164566.0, + "step": 20719 + }, + { + "epoch": 2.2754227981550628, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3809921741485596, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7425991296768188, + "num_tokens": 536189222.0, + "step": 20720 + }, + { + "epoch": 2.275532615857676, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7844903469085693, + "learning_rate": 1e-06, + "loss": 0.7776, + "mean_token_accuracy": 0.7443863749504089, + "num_tokens": 536207842.0, + "step": 20721 + }, + { + "epoch": 2.27564243356029, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.264399766921997, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7339191436767578, + "num_tokens": 536238776.0, + "step": 20722 + }, + { + "epoch": 2.2757522512629036, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7042057514190674, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.73528653383255, + "num_tokens": 536258384.0, + "step": 20723 + }, + { + "epoch": 2.2758620689655173, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1916189193725586, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7220867872238159, + "num_tokens": 536284965.0, + "step": 20724 + }, + { + "epoch": 2.275971886668131, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2801718711853027, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7336591482162476, + "num_tokens": 536313428.0, + "step": 20725 + }, + { + "epoch": 2.2760817043707444, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5003457069396973, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7185778617858887, + "num_tokens": 536339517.0, + "step": 20726 + }, + { + "epoch": 2.276191522073358, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.260040044784546, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7319943904876709, + "num_tokens": 536365847.0, + "step": 20727 + }, + { + "epoch": 2.276301339775972, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.289198398590088, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7083145380020142, + "num_tokens": 536393986.0, + "step": 20728 + }, + { + "epoch": 2.2764111574785857, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2655389308929443, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7071620225906372, + "num_tokens": 536422355.0, + "step": 20729 + }, + { + "epoch": 2.2765209751811994, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5197231769561768, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7266448736190796, + "num_tokens": 536447075.0, + "step": 20730 + }, + { + "epoch": 2.2766307928838128, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1873059272766113, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6969467997550964, + "num_tokens": 536478409.0, + "step": 20731 + }, + { + "epoch": 2.2767406105864265, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.376082181930542, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7175231575965881, + "num_tokens": 536505080.0, + "step": 20732 + }, + { + "epoch": 2.2768504282890403, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.371553659439087, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7509684562683105, + "num_tokens": 536529626.0, + "step": 20733 + }, + { + "epoch": 2.276960245991654, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.238520622253418, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7445783615112305, + "num_tokens": 536556253.0, + "step": 20734 + }, + { + "epoch": 2.2770700636942673, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6508493423461914, + "learning_rate": 1e-06, + "loss": 0.7444, + "mean_token_accuracy": 0.7578704357147217, + "num_tokens": 536578185.0, + "step": 20735 + }, + { + "epoch": 2.277179881396881, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6237940788269043, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7262499332427979, + "num_tokens": 536601047.0, + "step": 20736 + }, + { + "epoch": 2.277289699099495, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1967368125915527, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7395858764648438, + "num_tokens": 536631078.0, + "step": 20737 + }, + { + "epoch": 2.2773995168021086, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.859657049179077, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7291967272758484, + "num_tokens": 536649282.0, + "step": 20738 + }, + { + "epoch": 2.277509334504722, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3301327228546143, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7177069783210754, + "num_tokens": 536676114.0, + "step": 20739 + }, + { + "epoch": 2.2776191522073357, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.408799648284912, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7417392730712891, + "num_tokens": 536701464.0, + "step": 20740 + }, + { + "epoch": 2.2777289699099494, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.036592483520508, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7183197140693665, + "num_tokens": 536733933.0, + "step": 20741 + }, + { + "epoch": 2.277838787612563, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.088042974472046, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7405352592468262, + "num_tokens": 536764702.0, + "step": 20742 + }, + { + "epoch": 2.277948605315177, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3813211917877197, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7157812714576721, + "num_tokens": 536789400.0, + "step": 20743 + }, + { + "epoch": 2.2780584230177903, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.632671356201172, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.75682532787323, + "num_tokens": 536810565.0, + "step": 20744 + }, + { + "epoch": 2.278168240720404, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4341747760772705, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7305907011032104, + "num_tokens": 536835174.0, + "step": 20745 + }, + { + "epoch": 2.278278058423018, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4185574054718018, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7269442677497864, + "num_tokens": 536860085.0, + "step": 20746 + }, + { + "epoch": 2.2783878761256315, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5632576942443848, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7329563498497009, + "num_tokens": 536883060.0, + "step": 20747 + }, + { + "epoch": 2.2784976938282453, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.166775941848755, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7237491607666016, + "num_tokens": 536912050.0, + "step": 20748 + }, + { + "epoch": 2.2786075115308586, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5313851833343506, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7379900813102722, + "num_tokens": 536934709.0, + "step": 20749 + }, + { + "epoch": 2.2787173292334724, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1745119094848633, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7394611835479736, + "num_tokens": 536963123.0, + "step": 20750 + }, + { + "epoch": 2.278827146936086, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1272695064544678, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.718519389629364, + "num_tokens": 536994318.0, + "step": 20751 + }, + { + "epoch": 2.2789369646387, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.8947536945343018, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7395931482315063, + "num_tokens": 537013036.0, + "step": 20752 + }, + { + "epoch": 2.2790467823413136, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2804160118103027, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7064473628997803, + "num_tokens": 537042769.0, + "step": 20753 + }, + { + "epoch": 2.279156600043927, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3081631660461426, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.737173318862915, + "num_tokens": 537070224.0, + "step": 20754 + }, + { + "epoch": 2.2792664177465407, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2355692386627197, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.703016459941864, + "num_tokens": 537100398.0, + "step": 20755 + }, + { + "epoch": 2.2793762354491545, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.308206796646118, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7398487329483032, + "num_tokens": 537124562.0, + "step": 20756 + }, + { + "epoch": 2.279486053151768, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.572754383087158, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7204352021217346, + "num_tokens": 537148606.0, + "step": 20757 + }, + { + "epoch": 2.2795958708543815, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2450404167175293, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7204653024673462, + "num_tokens": 537177375.0, + "step": 20758 + }, + { + "epoch": 2.2797056885569953, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5117053985595703, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7515830993652344, + "num_tokens": 537198936.0, + "step": 20759 + }, + { + "epoch": 2.279815506259609, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.510831832885742, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.750757098197937, + "num_tokens": 537222002.0, + "step": 20760 + }, + { + "epoch": 2.279925323962223, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1960971355438232, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7234793901443481, + "num_tokens": 537252011.0, + "step": 20761 + }, + { + "epoch": 2.2800351416648366, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4133126735687256, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7425550222396851, + "num_tokens": 537275960.0, + "step": 20762 + }, + { + "epoch": 2.28014495936745, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.098797082901001, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7327605485916138, + "num_tokens": 537306553.0, + "step": 20763 + }, + { + "epoch": 2.2802547770700636, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5457346439361572, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7435036301612854, + "num_tokens": 537329446.0, + "step": 20764 + }, + { + "epoch": 2.2803645947726774, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7315852642059326, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7373002767562866, + "num_tokens": 537352940.0, + "step": 20765 + }, + { + "epoch": 2.280474412475291, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2936880588531494, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7408394813537598, + "num_tokens": 537381267.0, + "step": 20766 + }, + { + "epoch": 2.2805842301779045, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3632237911224365, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7333747744560242, + "num_tokens": 537407946.0, + "step": 20767 + }, + { + "epoch": 2.280694047880518, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.761711597442627, + "learning_rate": 1e-06, + "loss": 0.7937, + "mean_token_accuracy": 0.7489527463912964, + "num_tokens": 537426873.0, + "step": 20768 + }, + { + "epoch": 2.280803865583132, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3809030055999756, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7290600538253784, + "num_tokens": 537453441.0, + "step": 20769 + }, + { + "epoch": 2.2809136832857457, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4579904079437256, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7294283509254456, + "num_tokens": 537477762.0, + "step": 20770 + }, + { + "epoch": 2.2810235009883595, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3494699001312256, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7225104570388794, + "num_tokens": 537504972.0, + "step": 20771 + }, + { + "epoch": 2.281133318690973, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.06705641746521, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7038441896438599, + "num_tokens": 537538606.0, + "step": 20772 + }, + { + "epoch": 2.2812431363935866, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.12194561958313, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7183699607849121, + "num_tokens": 537570654.0, + "step": 20773 + }, + { + "epoch": 2.2813529540962003, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3473899364471436, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7253835201263428, + "num_tokens": 537597100.0, + "step": 20774 + }, + { + "epoch": 2.281462771798814, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4017012119293213, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7375520467758179, + "num_tokens": 537621795.0, + "step": 20775 + }, + { + "epoch": 2.281572589501428, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.472212553024292, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7348926067352295, + "num_tokens": 537645966.0, + "step": 20776 + }, + { + "epoch": 2.281682407204041, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2759768962860107, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7140721678733826, + "num_tokens": 537675897.0, + "step": 20777 + }, + { + "epoch": 2.281792224906655, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3757832050323486, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7325351238250732, + "num_tokens": 537703630.0, + "step": 20778 + }, + { + "epoch": 2.2819020426092687, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.296863079071045, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7011010050773621, + "num_tokens": 537731596.0, + "step": 20779 + }, + { + "epoch": 2.2820118603118824, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.257373571395874, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7270795106887817, + "num_tokens": 537759309.0, + "step": 20780 + }, + { + "epoch": 2.282121678014496, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.613992929458618, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7354574799537659, + "num_tokens": 537779453.0, + "step": 20781 + }, + { + "epoch": 2.2822314957171095, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4654195308685303, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7446873188018799, + "num_tokens": 537804432.0, + "step": 20782 + }, + { + "epoch": 2.2823413134197232, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.143821954727173, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7351610660552979, + "num_tokens": 537833522.0, + "step": 20783 + }, + { + "epoch": 2.282451131122337, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2533788681030273, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7284059524536133, + "num_tokens": 537861995.0, + "step": 20784 + }, + { + "epoch": 2.2825609488249508, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6885619163513184, + "learning_rate": 1e-06, + "loss": 0.7821, + "mean_token_accuracy": 0.7485044002532959, + "num_tokens": 537883858.0, + "step": 20785 + }, + { + "epoch": 2.282670766527564, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.315221071243286, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7123290300369263, + "num_tokens": 537910691.0, + "step": 20786 + }, + { + "epoch": 2.282780584230178, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7471487522125244, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.728317141532898, + "num_tokens": 537932131.0, + "step": 20787 + }, + { + "epoch": 2.2828904019327916, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4265456199645996, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7114651203155518, + "num_tokens": 537959397.0, + "step": 20788 + }, + { + "epoch": 2.2830002196354053, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7725942134857178, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.72099369764328, + "num_tokens": 537980465.0, + "step": 20789 + }, + { + "epoch": 2.2831100373380186, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1175050735473633, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.71718430519104, + "num_tokens": 538012996.0, + "step": 20790 + }, + { + "epoch": 2.2832198550406324, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.0992534160614014, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7097846865653992, + "num_tokens": 538042730.0, + "step": 20791 + }, + { + "epoch": 2.283329672743246, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.752788543701172, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7354868054389954, + "num_tokens": 538064188.0, + "step": 20792 + }, + { + "epoch": 2.28343949044586, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.484027624130249, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7466402053833008, + "num_tokens": 538089270.0, + "step": 20793 + }, + { + "epoch": 2.2835493081484737, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2569963932037354, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7258880138397217, + "num_tokens": 538117966.0, + "step": 20794 + }, + { + "epoch": 2.283659125851087, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2135205268859863, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7436497211456299, + "num_tokens": 538146195.0, + "step": 20795 + }, + { + "epoch": 2.2837689435537007, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.334385395050049, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7324259877204895, + "num_tokens": 538171123.0, + "step": 20796 + }, + { + "epoch": 2.2838787612563145, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.484466075897217, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.713010847568512, + "num_tokens": 538195696.0, + "step": 20797 + }, + { + "epoch": 2.2839885789589283, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.195850372314453, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7275371551513672, + "num_tokens": 538224843.0, + "step": 20798 + }, + { + "epoch": 2.284098396661542, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.839503049850464, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7423298954963684, + "num_tokens": 538243669.0, + "step": 20799 + }, + { + "epoch": 2.2842082143641553, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2221243381500244, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7364705801010132, + "num_tokens": 538271782.0, + "step": 20800 + }, + { + "epoch": 2.284318032066769, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.421849489212036, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7263054847717285, + "num_tokens": 538295796.0, + "step": 20801 + }, + { + "epoch": 2.284427849769383, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5573527812957764, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7196125984191895, + "num_tokens": 538319690.0, + "step": 20802 + }, + { + "epoch": 2.2845376674719966, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4091546535491943, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7252444624900818, + "num_tokens": 538345376.0, + "step": 20803 + }, + { + "epoch": 2.2846474851746104, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.39170241355896, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.738978922367096, + "num_tokens": 538368600.0, + "step": 20804 + }, + { + "epoch": 2.2847573028772237, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2909164428710938, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7325918674468994, + "num_tokens": 538396941.0, + "step": 20805 + }, + { + "epoch": 2.2848671205798374, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.509571075439453, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7055350542068481, + "num_tokens": 538423386.0, + "step": 20806 + }, + { + "epoch": 2.284976938282451, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4197893142700195, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7223258018493652, + "num_tokens": 538449073.0, + "step": 20807 + }, + { + "epoch": 2.285086755985065, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.135702610015869, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7275605201721191, + "num_tokens": 538481914.0, + "step": 20808 + }, + { + "epoch": 2.2851965736876787, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6321029663085938, + "learning_rate": 1e-06, + "loss": 0.7754, + "mean_token_accuracy": 0.7518539428710938, + "num_tokens": 538502236.0, + "step": 20809 + }, + { + "epoch": 2.285306391390292, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7170777320861816, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7305632829666138, + "num_tokens": 538522773.0, + "step": 20810 + }, + { + "epoch": 2.2854162090929058, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.497070789337158, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.738933801651001, + "num_tokens": 538547524.0, + "step": 20811 + }, + { + "epoch": 2.2855260267955195, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5908381938934326, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7469487190246582, + "num_tokens": 538569996.0, + "step": 20812 + }, + { + "epoch": 2.2856358444981333, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2268223762512207, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7255387902259827, + "num_tokens": 538597151.0, + "step": 20813 + }, + { + "epoch": 2.2857456622007466, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.162849187850952, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7094871997833252, + "num_tokens": 538627382.0, + "step": 20814 + }, + { + "epoch": 2.2858554799033604, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4401044845581055, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7311326265335083, + "num_tokens": 538652426.0, + "step": 20815 + }, + { + "epoch": 2.285965297605974, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.274977207183838, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6892767548561096, + "num_tokens": 538681280.0, + "step": 20816 + }, + { + "epoch": 2.286075115308588, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2399232387542725, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.720251739025116, + "num_tokens": 538709258.0, + "step": 20817 + }, + { + "epoch": 2.286184933011201, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.493884563446045, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7316166162490845, + "num_tokens": 538734645.0, + "step": 20818 + }, + { + "epoch": 2.286294750713815, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4304823875427246, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7417725324630737, + "num_tokens": 538759285.0, + "step": 20819 + }, + { + "epoch": 2.2864045684164287, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1447980403900146, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.7020593881607056, + "num_tokens": 538789988.0, + "step": 20820 + }, + { + "epoch": 2.2865143861190425, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4750618934631348, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7381392121315002, + "num_tokens": 538813693.0, + "step": 20821 + }, + { + "epoch": 2.286624203821656, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.289649248123169, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7440048456192017, + "num_tokens": 538839601.0, + "step": 20822 + }, + { + "epoch": 2.2867340215242695, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.0758442878723145, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7277814149856567, + "num_tokens": 538869111.0, + "step": 20823 + }, + { + "epoch": 2.2868438392268833, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4189980030059814, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7281414270401001, + "num_tokens": 538895327.0, + "step": 20824 + }, + { + "epoch": 2.286953656929497, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6053974628448486, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7126768231391907, + "num_tokens": 538917551.0, + "step": 20825 + }, + { + "epoch": 2.287063474632111, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.171138286590576, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7162593603134155, + "num_tokens": 538948810.0, + "step": 20826 + }, + { + "epoch": 2.2871732923347246, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.335179328918457, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7051982879638672, + "num_tokens": 538977071.0, + "step": 20827 + }, + { + "epoch": 2.287283110037338, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2770943641662598, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7425535917282104, + "num_tokens": 539005221.0, + "step": 20828 + }, + { + "epoch": 2.2873929277399516, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1161491870880127, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7201483845710754, + "num_tokens": 539037391.0, + "step": 20829 + }, + { + "epoch": 2.2875027454425654, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 3.1085927486419678, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7401238679885864, + "num_tokens": 539054853.0, + "step": 20830 + }, + { + "epoch": 2.287612563145179, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5557801723480225, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7501574754714966, + "num_tokens": 539076794.0, + "step": 20831 + }, + { + "epoch": 2.287722380847793, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.444138526916504, + "learning_rate": 1e-06, + "loss": 0.7673, + "mean_token_accuracy": 0.762424647808075, + "num_tokens": 539099218.0, + "step": 20832 + }, + { + "epoch": 2.287832198550406, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.319032669067383, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7282402515411377, + "num_tokens": 539124379.0, + "step": 20833 + }, + { + "epoch": 2.28794201625302, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.342691659927368, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7267914414405823, + "num_tokens": 539150985.0, + "step": 20834 + }, + { + "epoch": 2.2880518339556337, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.669560194015503, + "learning_rate": 1e-06, + "loss": 0.799, + "mean_token_accuracy": 0.7443757057189941, + "num_tokens": 539171635.0, + "step": 20835 + }, + { + "epoch": 2.2881616516582475, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4561779499053955, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7413312196731567, + "num_tokens": 539195741.0, + "step": 20836 + }, + { + "epoch": 2.288271469360861, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.217615842819214, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7360529899597168, + "num_tokens": 539224786.0, + "step": 20837 + }, + { + "epoch": 2.2883812870634745, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.514099597930908, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7130483388900757, + "num_tokens": 539249631.0, + "step": 20838 + }, + { + "epoch": 2.2884911047660883, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.246103525161743, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7273998856544495, + "num_tokens": 539277615.0, + "step": 20839 + }, + { + "epoch": 2.288600922468702, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5082197189331055, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7314884066581726, + "num_tokens": 539300415.0, + "step": 20840 + }, + { + "epoch": 2.2887107401713154, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3850290775299072, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7425700426101685, + "num_tokens": 539324690.0, + "step": 20841 + }, + { + "epoch": 2.288820557873929, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.21435546875, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7408877611160278, + "num_tokens": 539353522.0, + "step": 20842 + }, + { + "epoch": 2.288930375576543, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.360873222351074, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7330735921859741, + "num_tokens": 539380717.0, + "step": 20843 + }, + { + "epoch": 2.2890401932791566, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.485630989074707, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.705314576625824, + "num_tokens": 539406095.0, + "step": 20844 + }, + { + "epoch": 2.2891500109817704, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.63289213180542, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.740833044052124, + "num_tokens": 539428055.0, + "step": 20845 + }, + { + "epoch": 2.2892598286843837, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4450466632843018, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7438586354255676, + "num_tokens": 539453392.0, + "step": 20846 + }, + { + "epoch": 2.2893696463869975, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.340895414352417, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7346876263618469, + "num_tokens": 539478657.0, + "step": 20847 + }, + { + "epoch": 2.2894794640896112, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3188345432281494, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7364315986633301, + "num_tokens": 539503239.0, + "step": 20848 + }, + { + "epoch": 2.289589281792225, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.268217086791992, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7509032487869263, + "num_tokens": 539529013.0, + "step": 20849 + }, + { + "epoch": 2.2896990994948387, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.396064281463623, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7465122938156128, + "num_tokens": 539552767.0, + "step": 20850 + }, + { + "epoch": 2.289808917197452, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.273920774459839, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7537794709205627, + "num_tokens": 539579488.0, + "step": 20851 + }, + { + "epoch": 2.289918734900066, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3055367469787598, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7209774255752563, + "num_tokens": 539606583.0, + "step": 20852 + }, + { + "epoch": 2.2900285526026796, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.472344160079956, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7170699834823608, + "num_tokens": 539631989.0, + "step": 20853 + }, + { + "epoch": 2.2901383703052933, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5572404861450195, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.727423906326294, + "num_tokens": 539656172.0, + "step": 20854 + }, + { + "epoch": 2.290248188007907, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3261237144470215, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7244582176208496, + "num_tokens": 539684394.0, + "step": 20855 + }, + { + "epoch": 2.2903580057105204, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2834267616271973, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7063709497451782, + "num_tokens": 539712695.0, + "step": 20856 + }, + { + "epoch": 2.290467823413134, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.24010968208313, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7136424779891968, + "num_tokens": 539741458.0, + "step": 20857 + }, + { + "epoch": 2.290577641115748, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3104310035705566, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7261213660240173, + "num_tokens": 539770286.0, + "step": 20858 + }, + { + "epoch": 2.2906874588183617, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.422200918197632, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.716425895690918, + "num_tokens": 539797632.0, + "step": 20859 + }, + { + "epoch": 2.2907972765209754, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.013610601425171, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7163073420524597, + "num_tokens": 539831806.0, + "step": 20860 + }, + { + "epoch": 2.2909070942235887, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.1652286052703857, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7255566120147705, + "num_tokens": 539859437.0, + "step": 20861 + }, + { + "epoch": 2.2910169119262025, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6648471355438232, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7209272980690002, + "num_tokens": 539880540.0, + "step": 20862 + }, + { + "epoch": 2.2911267296288162, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.8105154037475586, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7499289512634277, + "num_tokens": 539899973.0, + "step": 20863 + }, + { + "epoch": 2.29123654733143, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 1.9757161140441895, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7203556299209595, + "num_tokens": 539934351.0, + "step": 20864 + }, + { + "epoch": 2.2913463650340433, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.600219964981079, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7429797649383545, + "num_tokens": 539956011.0, + "step": 20865 + }, + { + "epoch": 2.291456182736657, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.603492498397827, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7426096796989441, + "num_tokens": 539976916.0, + "step": 20866 + }, + { + "epoch": 2.291566000439271, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3205020427703857, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7177530527114868, + "num_tokens": 540002841.0, + "step": 20867 + }, + { + "epoch": 2.2916758181418846, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3231916427612305, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7270203232765198, + "num_tokens": 540028328.0, + "step": 20868 + }, + { + "epoch": 2.291785635844498, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2739224433898926, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7258505821228027, + "num_tokens": 540054231.0, + "step": 20869 + }, + { + "epoch": 2.2918954535471117, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.752346992492676, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7361158132553101, + "num_tokens": 540072985.0, + "step": 20870 + }, + { + "epoch": 2.2920052712497254, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6010937690734863, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7436856627464294, + "num_tokens": 540093628.0, + "step": 20871 + }, + { + "epoch": 2.292115088952339, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.631838321685791, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7323760390281677, + "num_tokens": 540115163.0, + "step": 20872 + }, + { + "epoch": 2.292224906654953, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5703580379486084, + "learning_rate": 1e-06, + "loss": 0.794, + "mean_token_accuracy": 0.7479891777038574, + "num_tokens": 540137017.0, + "step": 20873 + }, + { + "epoch": 2.2923347243575662, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6673669815063477, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7318447828292847, + "num_tokens": 540158223.0, + "step": 20874 + }, + { + "epoch": 2.29244454206018, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.347442865371704, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7131342887878418, + "num_tokens": 540184447.0, + "step": 20875 + }, + { + "epoch": 2.2925543597627938, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.2118635177612305, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7191369533538818, + "num_tokens": 540214098.0, + "step": 20876 + }, + { + "epoch": 2.2926641774654075, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4199774265289307, + "learning_rate": 1e-06, + "loss": 0.7283, + "mean_token_accuracy": 0.7660465240478516, + "num_tokens": 540236665.0, + "step": 20877 + }, + { + "epoch": 2.2927739951680213, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3937745094299316, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.745419979095459, + "num_tokens": 540262127.0, + "step": 20878 + }, + { + "epoch": 2.2928838128706346, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3316152095794678, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.734620988368988, + "num_tokens": 540286459.0, + "step": 20879 + }, + { + "epoch": 2.2929936305732483, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.6452126502990723, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7236309051513672, + "num_tokens": 540308637.0, + "step": 20880 + }, + { + "epoch": 2.293103448275862, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4722790718078613, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.7525595426559448, + "num_tokens": 540330801.0, + "step": 20881 + }, + { + "epoch": 2.293213265978476, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.7464816570281982, + "learning_rate": 1e-06, + "loss": 0.7424, + "mean_token_accuracy": 0.7596509456634521, + "num_tokens": 540350290.0, + "step": 20882 + }, + { + "epoch": 2.2933230836810896, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.3521783351898193, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7388596534729004, + "num_tokens": 540375822.0, + "step": 20883 + }, + { + "epoch": 2.293432901383703, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.5885679721832275, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7258504033088684, + "num_tokens": 540400247.0, + "step": 20884 + }, + { + "epoch": 2.2935427190863167, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.8109893798828125, + "learning_rate": 1e-06, + "loss": 0.7156, + "mean_token_accuracy": 0.7627235651016235, + "num_tokens": 540419036.0, + "step": 20885 + }, + { + "epoch": 2.2936525367889304, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 1.9825137853622437, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6993613839149475, + "num_tokens": 540454269.0, + "step": 20886 + }, + { + "epoch": 2.293762354491544, + "ewc_loss": 0.00018215179443359375, + "grad_norm": 2.4100146293640137, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7408372759819031, + "num_tokens": 540478961.0, + "step": 20887 + }, + { + "epoch": 2.2938721721941575, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2970051765441895, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7453997135162354, + "num_tokens": 540505132.0, + "step": 20888 + }, + { + "epoch": 2.2939819898967713, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.8869407176971436, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.736111044883728, + "num_tokens": 540524365.0, + "step": 20889 + }, + { + "epoch": 2.294091807599385, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.403456926345825, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7227352857589722, + "num_tokens": 540551399.0, + "step": 20890 + }, + { + "epoch": 2.294201625301999, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4379920959472656, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7450064420700073, + "num_tokens": 540575806.0, + "step": 20891 + }, + { + "epoch": 2.294311443004612, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.397472858428955, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7299944758415222, + "num_tokens": 540599400.0, + "step": 20892 + }, + { + "epoch": 2.294421260707226, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.165964365005493, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7320137619972229, + "num_tokens": 540629135.0, + "step": 20893 + }, + { + "epoch": 2.2945310784098396, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.19697904586792, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.700247049331665, + "num_tokens": 540659440.0, + "step": 20894 + }, + { + "epoch": 2.2946408961124534, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.7208940982818604, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7475110292434692, + "num_tokens": 540680501.0, + "step": 20895 + }, + { + "epoch": 2.294750713815067, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4458775520324707, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7351232171058655, + "num_tokens": 540703039.0, + "step": 20896 + }, + { + "epoch": 2.2948605315176804, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2850759029388428, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7192471027374268, + "num_tokens": 540732665.0, + "step": 20897 + }, + { + "epoch": 2.294970349220294, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.362687110900879, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7287348508834839, + "num_tokens": 540759246.0, + "step": 20898 + }, + { + "epoch": 2.295080166922908, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1905782222747803, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7131851315498352, + "num_tokens": 540788127.0, + "step": 20899 + }, + { + "epoch": 2.2951899846255217, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.9061100482940674, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7477787137031555, + "num_tokens": 540806931.0, + "step": 20900 + }, + { + "epoch": 2.2952998023281355, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.8680155277252197, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7356952428817749, + "num_tokens": 540826294.0, + "step": 20901 + }, + { + "epoch": 2.2954096200307488, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.398280143737793, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7253999710083008, + "num_tokens": 540851938.0, + "step": 20902 + }, + { + "epoch": 2.2955194377333625, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.261641263961792, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7125105857849121, + "num_tokens": 540879803.0, + "step": 20903 + }, + { + "epoch": 2.2956292554359763, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3930530548095703, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7102895975112915, + "num_tokens": 540908534.0, + "step": 20904 + }, + { + "epoch": 2.29573907313859, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.622641086578369, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7237032055854797, + "num_tokens": 540930373.0, + "step": 20905 + }, + { + "epoch": 2.295848890841204, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.405585289001465, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7405626177787781, + "num_tokens": 540954331.0, + "step": 20906 + }, + { + "epoch": 2.295958708543817, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5916929244995117, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7458488941192627, + "num_tokens": 540975078.0, + "step": 20907 + }, + { + "epoch": 2.296068526246431, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.432678461074829, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7190181612968445, + "num_tokens": 541000601.0, + "step": 20908 + }, + { + "epoch": 2.2961783439490446, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3213071823120117, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7178082466125488, + "num_tokens": 541028698.0, + "step": 20909 + }, + { + "epoch": 2.2962881616516584, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.146292209625244, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7067584991455078, + "num_tokens": 541060964.0, + "step": 20910 + }, + { + "epoch": 2.296397979354272, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.498664617538452, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7270123958587646, + "num_tokens": 541083273.0, + "step": 20911 + }, + { + "epoch": 2.2965077970568855, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.51275634765625, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7054644227027893, + "num_tokens": 541107607.0, + "step": 20912 + }, + { + "epoch": 2.296617614759499, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.406266689300537, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7315490245819092, + "num_tokens": 541131056.0, + "step": 20913 + }, + { + "epoch": 2.296727432462113, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2854156494140625, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.719596266746521, + "num_tokens": 541158773.0, + "step": 20914 + }, + { + "epoch": 2.2968372501647267, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.466240406036377, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7381595373153687, + "num_tokens": 541181050.0, + "step": 20915 + }, + { + "epoch": 2.29694706786734, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3234663009643555, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7281723022460938, + "num_tokens": 541210007.0, + "step": 20916 + }, + { + "epoch": 2.297056885569954, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3057193756103516, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7205098867416382, + "num_tokens": 541235748.0, + "step": 20917 + }, + { + "epoch": 2.2971667032725676, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1777780055999756, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7236232757568359, + "num_tokens": 541267118.0, + "step": 20918 + }, + { + "epoch": 2.2972765209751813, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.529884099960327, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7350082397460938, + "num_tokens": 541290251.0, + "step": 20919 + }, + { + "epoch": 2.2973863386777946, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.095670700073242, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.709723949432373, + "num_tokens": 541321159.0, + "step": 20920 + }, + { + "epoch": 2.2974961563804084, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2324681282043457, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7213441133499146, + "num_tokens": 541348450.0, + "step": 20921 + }, + { + "epoch": 2.297605974083022, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4496939182281494, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7431915998458862, + "num_tokens": 541372135.0, + "step": 20922 + }, + { + "epoch": 2.297715791785636, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2125391960144043, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7222213745117188, + "num_tokens": 541401150.0, + "step": 20923 + }, + { + "epoch": 2.2978256094882497, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.0595974922180176, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7151108980178833, + "num_tokens": 541435158.0, + "step": 20924 + }, + { + "epoch": 2.297935427190863, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.9524054527282715, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7355015873908997, + "num_tokens": 541453374.0, + "step": 20925 + }, + { + "epoch": 2.2980452448934767, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4397974014282227, + "learning_rate": 1e-06, + "loss": 0.6867, + "mean_token_accuracy": 0.7824385166168213, + "num_tokens": 541475660.0, + "step": 20926 + }, + { + "epoch": 2.2981550625960905, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2462172508239746, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7512360215187073, + "num_tokens": 541503522.0, + "step": 20927 + }, + { + "epoch": 2.2982648802987042, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4163601398468018, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.736504852771759, + "num_tokens": 541528360.0, + "step": 20928 + }, + { + "epoch": 2.298374698001318, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.499150276184082, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7183132171630859, + "num_tokens": 541554598.0, + "step": 20929 + }, + { + "epoch": 2.2984845157039313, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4524986743927, + "learning_rate": 1e-06, + "loss": 0.7651, + "mean_token_accuracy": 0.7549587488174438, + "num_tokens": 541577840.0, + "step": 20930 + }, + { + "epoch": 2.298594333406545, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.248675584793091, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7085909843444824, + "num_tokens": 541608864.0, + "step": 20931 + }, + { + "epoch": 2.298704151109159, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2872064113616943, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7022489309310913, + "num_tokens": 541636803.0, + "step": 20932 + }, + { + "epoch": 2.2988139688117726, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2785608768463135, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7289023995399475, + "num_tokens": 541664820.0, + "step": 20933 + }, + { + "epoch": 2.2989237865143863, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.42999267578125, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7237964868545532, + "num_tokens": 541690577.0, + "step": 20934 + }, + { + "epoch": 2.2990336042169996, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2969918251037598, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7293453812599182, + "num_tokens": 541716375.0, + "step": 20935 + }, + { + "epoch": 2.2991434219196134, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.379204511642456, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7445013523101807, + "num_tokens": 541742331.0, + "step": 20936 + }, + { + "epoch": 2.299253239622227, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3479275703430176, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7362135052680969, + "num_tokens": 541767319.0, + "step": 20937 + }, + { + "epoch": 2.299363057324841, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2792906761169434, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7061296105384827, + "num_tokens": 541795125.0, + "step": 20938 + }, + { + "epoch": 2.2994728750274542, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.0316154956817627, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7133303880691528, + "num_tokens": 541830697.0, + "step": 20939 + }, + { + "epoch": 2.299582692730068, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.390225410461426, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.7394288778305054, + "num_tokens": 541856540.0, + "step": 20940 + }, + { + "epoch": 2.2996925104326817, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.7507972717285156, + "learning_rate": 1e-06, + "loss": 0.7691, + "mean_token_accuracy": 0.7528162002563477, + "num_tokens": 541876615.0, + "step": 20941 + }, + { + "epoch": 2.2998023281352955, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.175483226776123, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7231259942054749, + "num_tokens": 541903735.0, + "step": 20942 + }, + { + "epoch": 2.2999121458379093, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.112224578857422, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7109426259994507, + "num_tokens": 541935807.0, + "step": 20943 + }, + { + "epoch": 2.3000219635405226, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3589913845062256, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7390075922012329, + "num_tokens": 541962688.0, + "step": 20944 + }, + { + "epoch": 2.3001317812431363, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.315462350845337, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7198972105979919, + "num_tokens": 541988938.0, + "step": 20945 + }, + { + "epoch": 2.30024159894575, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6363096237182617, + "learning_rate": 1e-06, + "loss": 0.7467, + "mean_token_accuracy": 0.7606955170631409, + "num_tokens": 542010466.0, + "step": 20946 + }, + { + "epoch": 2.300351416648364, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4010019302368164, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7358651161193848, + "num_tokens": 542034293.0, + "step": 20947 + }, + { + "epoch": 2.300461234350977, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.465916156768799, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7123566269874573, + "num_tokens": 542059149.0, + "step": 20948 + }, + { + "epoch": 2.300571052053591, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.0902819633483887, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7284494042396545, + "num_tokens": 542089035.0, + "step": 20949 + }, + { + "epoch": 2.3006808697562047, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.39396333694458, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.6990424990653992, + "num_tokens": 542116054.0, + "step": 20950 + }, + { + "epoch": 2.3007906874588184, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.250370502471924, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7196260094642639, + "num_tokens": 542145231.0, + "step": 20951 + }, + { + "epoch": 2.300900505161432, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.231764793395996, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7244839072227478, + "num_tokens": 542173544.0, + "step": 20952 + }, + { + "epoch": 2.3010103228640455, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1800930500030518, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7110745906829834, + "num_tokens": 542202205.0, + "step": 20953 + }, + { + "epoch": 2.3011201405666593, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3801746368408203, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7239069938659668, + "num_tokens": 542227842.0, + "step": 20954 + }, + { + "epoch": 2.301229958269273, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.129714012145996, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.724635124206543, + "num_tokens": 542255069.0, + "step": 20955 + }, + { + "epoch": 2.3013397759718868, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3004329204559326, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7297714948654175, + "num_tokens": 542282401.0, + "step": 20956 + }, + { + "epoch": 2.3014495936745005, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3085052967071533, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.7541662454605103, + "num_tokens": 542308418.0, + "step": 20957 + }, + { + "epoch": 2.301559411377114, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.505845308303833, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7264238595962524, + "num_tokens": 542333647.0, + "step": 20958 + }, + { + "epoch": 2.3016692290797276, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2809295654296875, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7312109470367432, + "num_tokens": 542359218.0, + "step": 20959 + }, + { + "epoch": 2.3017790467823414, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.229025363922119, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7398020029067993, + "num_tokens": 542388516.0, + "step": 20960 + }, + { + "epoch": 2.301888864484955, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.265108346939087, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7161956429481506, + "num_tokens": 542417076.0, + "step": 20961 + }, + { + "epoch": 2.301998682187569, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5609893798828125, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.730485200881958, + "num_tokens": 542439792.0, + "step": 20962 + }, + { + "epoch": 2.302108499890182, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1949102878570557, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.723317563533783, + "num_tokens": 542469589.0, + "step": 20963 + }, + { + "epoch": 2.302218317592796, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4080686569213867, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7248827219009399, + "num_tokens": 542494322.0, + "step": 20964 + }, + { + "epoch": 2.3023281352954097, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3668301105499268, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7487807273864746, + "num_tokens": 542519882.0, + "step": 20965 + }, + { + "epoch": 2.3024379529980235, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2895233631134033, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7221925854682922, + "num_tokens": 542548456.0, + "step": 20966 + }, + { + "epoch": 2.3025477707006368, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.516571283340454, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7297021746635437, + "num_tokens": 542573067.0, + "step": 20967 + }, + { + "epoch": 2.3026575884032505, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2379441261291504, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7205791473388672, + "num_tokens": 542602039.0, + "step": 20968 + }, + { + "epoch": 2.3027674061058643, + "ewc_loss": 0.00018310546875, + "grad_norm": 3.910735845565796, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7498654127120972, + "num_tokens": 542622463.0, + "step": 20969 + }, + { + "epoch": 2.302877223808478, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.30059552192688, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7087465524673462, + "num_tokens": 542649997.0, + "step": 20970 + }, + { + "epoch": 2.3029870415110913, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2003297805786133, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7391365766525269, + "num_tokens": 542679946.0, + "step": 20971 + }, + { + "epoch": 2.303096859213705, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.48321795463562, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.736055314540863, + "num_tokens": 542704940.0, + "step": 20972 + }, + { + "epoch": 2.303206676916319, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1527137756347656, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7237986326217651, + "num_tokens": 542735386.0, + "step": 20973 + }, + { + "epoch": 2.3033164946189326, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3923556804656982, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.738857090473175, + "num_tokens": 542760679.0, + "step": 20974 + }, + { + "epoch": 2.3034263123215464, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.107757806777954, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7409892082214355, + "num_tokens": 542789594.0, + "step": 20975 + }, + { + "epoch": 2.3035361300241597, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2098162174224854, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7199731469154358, + "num_tokens": 542817666.0, + "step": 20976 + }, + { + "epoch": 2.3036459477267734, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3948404788970947, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7487314343452454, + "num_tokens": 542841983.0, + "step": 20977 + }, + { + "epoch": 2.303755765429387, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2613048553466797, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7333272695541382, + "num_tokens": 542869857.0, + "step": 20978 + }, + { + "epoch": 2.303865583132001, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.158748149871826, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.724210262298584, + "num_tokens": 542900501.0, + "step": 20979 + }, + { + "epoch": 2.3039754008346147, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.713141441345215, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7348936796188354, + "num_tokens": 542922771.0, + "step": 20980 + }, + { + "epoch": 2.304085218537228, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.463890790939331, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7519103288650513, + "num_tokens": 542945980.0, + "step": 20981 + }, + { + "epoch": 2.304195036239842, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6007909774780273, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7192158102989197, + "num_tokens": 542969733.0, + "step": 20982 + }, + { + "epoch": 2.3043048539424555, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3252875804901123, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7436602115631104, + "num_tokens": 542997008.0, + "step": 20983 + }, + { + "epoch": 2.3044146716450693, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4421093463897705, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7364230155944824, + "num_tokens": 543021143.0, + "step": 20984 + }, + { + "epoch": 2.304524489347683, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.392881393432617, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7424517869949341, + "num_tokens": 543045313.0, + "step": 20985 + }, + { + "epoch": 2.3046343070502964, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5086588859558105, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7421504855155945, + "num_tokens": 543068448.0, + "step": 20986 + }, + { + "epoch": 2.30474412475291, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1274898052215576, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7216604351997375, + "num_tokens": 543099282.0, + "step": 20987 + }, + { + "epoch": 2.304853942455524, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6891939640045166, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7323015332221985, + "num_tokens": 543120078.0, + "step": 20988 + }, + { + "epoch": 2.3049637601581376, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4702908992767334, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7336719036102295, + "num_tokens": 543143251.0, + "step": 20989 + }, + { + "epoch": 2.3050735778607514, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3838019371032715, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7198053002357483, + "num_tokens": 543167052.0, + "step": 20990 + }, + { + "epoch": 2.3051833955633647, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6514086723327637, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7287203073501587, + "num_tokens": 543188857.0, + "step": 20991 + }, + { + "epoch": 2.3052932132659785, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4059691429138184, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7330718040466309, + "num_tokens": 543215243.0, + "step": 20992 + }, + { + "epoch": 2.3054030309685922, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.545538902282715, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7247121334075928, + "num_tokens": 543239720.0, + "step": 20993 + }, + { + "epoch": 2.305512848671206, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6891367435455322, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7382630705833435, + "num_tokens": 543261966.0, + "step": 20994 + }, + { + "epoch": 2.3056226663738193, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6995255947113037, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7467976808547974, + "num_tokens": 543281693.0, + "step": 20995 + }, + { + "epoch": 2.305732484076433, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5549418926239014, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.742796003818512, + "num_tokens": 543303138.0, + "step": 20996 + }, + { + "epoch": 2.305842301779047, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4591362476348877, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7058305740356445, + "num_tokens": 543330442.0, + "step": 20997 + }, + { + "epoch": 2.3059521194816606, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1693027019500732, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7156056761741638, + "num_tokens": 543362534.0, + "step": 20998 + }, + { + "epoch": 2.306061937184274, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6723148822784424, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7386826276779175, + "num_tokens": 543382404.0, + "step": 20999 + }, + { + "epoch": 2.3061717548868876, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4640820026397705, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7198343873023987, + "num_tokens": 543407543.0, + "step": 21000 + }, + { + "epoch": 2.3062815725895014, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.515822410583496, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.743868350982666, + "num_tokens": 543430168.0, + "step": 21001 + }, + { + "epoch": 2.306391390292115, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.261563301086426, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7227760553359985, + "num_tokens": 543458488.0, + "step": 21002 + }, + { + "epoch": 2.306501207994729, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6140642166137695, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.741696298122406, + "num_tokens": 543481233.0, + "step": 21003 + }, + { + "epoch": 2.306611025697342, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.511004686355591, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7279653549194336, + "num_tokens": 543504692.0, + "step": 21004 + }, + { + "epoch": 2.306720843399956, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6049280166625977, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7396462559700012, + "num_tokens": 543526621.0, + "step": 21005 + }, + { + "epoch": 2.3068306611025697, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1734352111816406, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7270038723945618, + "num_tokens": 543554414.0, + "step": 21006 + }, + { + "epoch": 2.3069404788051835, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.111663579940796, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7065698504447937, + "num_tokens": 543587486.0, + "step": 21007 + }, + { + "epoch": 2.3070502965077972, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3171660900115967, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7173902988433838, + "num_tokens": 543615678.0, + "step": 21008 + }, + { + "epoch": 2.3071601142104106, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4912872314453125, + "learning_rate": 1e-06, + "loss": 0.7699, + "mean_token_accuracy": 0.7534456253051758, + "num_tokens": 543639927.0, + "step": 21009 + }, + { + "epoch": 2.3072699319130243, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3094682693481445, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7131088972091675, + "num_tokens": 543668013.0, + "step": 21010 + }, + { + "epoch": 2.307379749615638, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5190951824188232, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7486138939857483, + "num_tokens": 543692129.0, + "step": 21011 + }, + { + "epoch": 2.307489567318252, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5323872566223145, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7424721121788025, + "num_tokens": 543714848.0, + "step": 21012 + }, + { + "epoch": 2.3075993850208656, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4549801349639893, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7275080680847168, + "num_tokens": 543739756.0, + "step": 21013 + }, + { + "epoch": 2.307709202723479, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5474679470062256, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7279229164123535, + "num_tokens": 543761497.0, + "step": 21014 + }, + { + "epoch": 2.3078190204260927, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.7090039253234863, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7324851751327515, + "num_tokens": 543783793.0, + "step": 21015 + }, + { + "epoch": 2.3079288381287064, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.420332431793213, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7211954593658447, + "num_tokens": 543809473.0, + "step": 21016 + }, + { + "epoch": 2.30803865583132, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.51652455329895, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7363541126251221, + "num_tokens": 543832119.0, + "step": 21017 + }, + { + "epoch": 2.3081484735339335, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.401524066925049, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7286491394042969, + "num_tokens": 543858050.0, + "step": 21018 + }, + { + "epoch": 2.3082582912365472, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4139957427978516, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7270019054412842, + "num_tokens": 543883102.0, + "step": 21019 + }, + { + "epoch": 2.308368108939161, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.619068145751953, + "learning_rate": 1e-06, + "loss": 0.7833, + "mean_token_accuracy": 0.7525099515914917, + "num_tokens": 543904191.0, + "step": 21020 + }, + { + "epoch": 2.3084779266417748, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.622403144836426, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7368236780166626, + "num_tokens": 543925282.0, + "step": 21021 + }, + { + "epoch": 2.308587744344388, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5914371013641357, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.727358341217041, + "num_tokens": 543948199.0, + "step": 21022 + }, + { + "epoch": 2.308697562047002, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.0422608852386475, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7498183250427246, + "num_tokens": 543978822.0, + "step": 21023 + }, + { + "epoch": 2.3088073797496156, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1945621967315674, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7088958621025085, + "num_tokens": 544009634.0, + "step": 21024 + }, + { + "epoch": 2.3089171974522293, + "ewc_loss": 0.00018310546875, + "grad_norm": 1.9995973110198975, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7154163718223572, + "num_tokens": 544045825.0, + "step": 21025 + }, + { + "epoch": 2.309027015154843, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6019504070281982, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7373796105384827, + "num_tokens": 544066655.0, + "step": 21026 + }, + { + "epoch": 2.3091368328574564, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5642755031585693, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7218499183654785, + "num_tokens": 544088602.0, + "step": 21027 + }, + { + "epoch": 2.30924665056007, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4441676139831543, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7435129880905151, + "num_tokens": 544111969.0, + "step": 21028 + }, + { + "epoch": 2.309356468262684, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.3773727416992188, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7505424618721008, + "num_tokens": 544136419.0, + "step": 21029 + }, + { + "epoch": 2.3094662859652977, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.1228585243225098, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7237780690193176, + "num_tokens": 544167652.0, + "step": 21030 + }, + { + "epoch": 2.3095761036679114, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.156156063079834, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.724857747554779, + "num_tokens": 544196676.0, + "step": 21031 + }, + { + "epoch": 2.3096859213705248, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4334590435028076, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7385385036468506, + "num_tokens": 544222160.0, + "step": 21032 + }, + { + "epoch": 2.3097957390731385, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.33038067817688, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7161647081375122, + "num_tokens": 544250547.0, + "step": 21033 + }, + { + "epoch": 2.3099055567757523, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.5181000232696533, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7160660028457642, + "num_tokens": 544275216.0, + "step": 21034 + }, + { + "epoch": 2.310015374478366, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.2690320014953613, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7278406620025635, + "num_tokens": 544302957.0, + "step": 21035 + }, + { + "epoch": 2.31012519218098, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4313530921936035, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7220249176025391, + "num_tokens": 544330432.0, + "step": 21036 + }, + { + "epoch": 2.310235009883593, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4555702209472656, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7294941544532776, + "num_tokens": 544358067.0, + "step": 21037 + }, + { + "epoch": 2.310344827586207, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.6730077266693115, + "learning_rate": 1e-06, + "loss": 0.656, + "mean_token_accuracy": 0.7876600027084351, + "num_tokens": 544378540.0, + "step": 21038 + }, + { + "epoch": 2.3104546452888206, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4400477409362793, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.7595133781433105, + "num_tokens": 544402832.0, + "step": 21039 + }, + { + "epoch": 2.3105644629914344, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.4992563724517822, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7564226388931274, + "num_tokens": 544424797.0, + "step": 21040 + }, + { + "epoch": 2.310674280694048, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.158768892288208, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7227257490158081, + "num_tokens": 544457150.0, + "step": 21041 + }, + { + "epoch": 2.3107840983966614, + "ewc_loss": 0.00018310546875, + "grad_norm": 2.111668825149536, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7294273376464844, + "num_tokens": 544488923.0, + "step": 21042 + }, + { + "epoch": 2.310893916099275, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2590270042419434, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7358166575431824, + "num_tokens": 544518724.0, + "step": 21043 + }, + { + "epoch": 2.311003733801889, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 1.9926053285598755, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7350186109542847, + "num_tokens": 544551721.0, + "step": 21044 + }, + { + "epoch": 2.3111135515045027, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.559640645980835, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7171199321746826, + "num_tokens": 544575492.0, + "step": 21045 + }, + { + "epoch": 2.311223369207116, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2526798248291016, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7234734892845154, + "num_tokens": 544602952.0, + "step": 21046 + }, + { + "epoch": 2.3113331869097298, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.395597457885742, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.743727445602417, + "num_tokens": 544628679.0, + "step": 21047 + }, + { + "epoch": 2.3114430046123435, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3951027393341064, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7330507040023804, + "num_tokens": 544655406.0, + "step": 21048 + }, + { + "epoch": 2.3115528223149573, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.563582420349121, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7202776074409485, + "num_tokens": 544676038.0, + "step": 21049 + }, + { + "epoch": 2.3116626400175706, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.6320712566375732, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7512911558151245, + "num_tokens": 544696723.0, + "step": 21050 + }, + { + "epoch": 2.3117724577201844, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4143528938293457, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.74012291431427, + "num_tokens": 544720954.0, + "step": 21051 + }, + { + "epoch": 2.311882275422798, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2901175022125244, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7270511388778687, + "num_tokens": 544748396.0, + "step": 21052 + }, + { + "epoch": 2.311992093125412, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.417775869369507, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7380720376968384, + "num_tokens": 544771867.0, + "step": 21053 + }, + { + "epoch": 2.3121019108280256, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.424281358718872, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.748816967010498, + "num_tokens": 544794970.0, + "step": 21054 + }, + { + "epoch": 2.312211728530639, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2881181240081787, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7445751428604126, + "num_tokens": 544821841.0, + "step": 21055 + }, + { + "epoch": 2.3123215462332527, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.349290132522583, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7260990142822266, + "num_tokens": 544849110.0, + "step": 21056 + }, + { + "epoch": 2.3124313639358665, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4085092544555664, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7237699627876282, + "num_tokens": 544876829.0, + "step": 21057 + }, + { + "epoch": 2.31254118163848, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2204415798187256, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7470256090164185, + "num_tokens": 544904983.0, + "step": 21058 + }, + { + "epoch": 2.312650999341094, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.739725112915039, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7168159484863281, + "num_tokens": 544926799.0, + "step": 21059 + }, + { + "epoch": 2.3127608170437073, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.332503318786621, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7034131288528442, + "num_tokens": 544956576.0, + "step": 21060 + }, + { + "epoch": 2.312870634746321, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.6370689868927, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7486163377761841, + "num_tokens": 544977466.0, + "step": 21061 + }, + { + "epoch": 2.312980452448935, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2591066360473633, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7140135169029236, + "num_tokens": 545006051.0, + "step": 21062 + }, + { + "epoch": 2.3130902701515486, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3522543907165527, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7181389331817627, + "num_tokens": 545032919.0, + "step": 21063 + }, + { + "epoch": 2.3132000878541623, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.314663887023926, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7285109162330627, + "num_tokens": 545058743.0, + "step": 21064 + }, + { + "epoch": 2.3133099055567756, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.5571470260620117, + "learning_rate": 1e-06, + "loss": 0.7889, + "mean_token_accuracy": 0.7487459182739258, + "num_tokens": 545080941.0, + "step": 21065 + }, + { + "epoch": 2.3134197232593894, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2301721572875977, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7178101539611816, + "num_tokens": 545111743.0, + "step": 21066 + }, + { + "epoch": 2.313529540962003, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4418389797210693, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7181863188743591, + "num_tokens": 545137372.0, + "step": 21067 + }, + { + "epoch": 2.313639358664617, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.505964517593384, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7248873114585876, + "num_tokens": 545161184.0, + "step": 21068 + }, + { + "epoch": 2.31374917636723, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.438962936401367, + "learning_rate": 1e-06, + "loss": 0.7405, + "mean_token_accuracy": 0.7636932730674744, + "num_tokens": 545184304.0, + "step": 21069 + }, + { + "epoch": 2.313858994069844, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3139536380767822, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.722720742225647, + "num_tokens": 545210656.0, + "step": 21070 + }, + { + "epoch": 2.3139688117724577, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4785897731781006, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7569260597229004, + "num_tokens": 545235013.0, + "step": 21071 + }, + { + "epoch": 2.3140786294750715, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.292189598083496, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.727994441986084, + "num_tokens": 545261614.0, + "step": 21072 + }, + { + "epoch": 2.314188447177685, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2948546409606934, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7327815294265747, + "num_tokens": 545288530.0, + "step": 21073 + }, + { + "epoch": 2.3142982648802986, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2949025630950928, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7167487144470215, + "num_tokens": 545316653.0, + "step": 21074 + }, + { + "epoch": 2.3144080825829123, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3450026512145996, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7330517768859863, + "num_tokens": 545344921.0, + "step": 21075 + }, + { + "epoch": 2.314517900285526, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.1309165954589844, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7077341675758362, + "num_tokens": 545375523.0, + "step": 21076 + }, + { + "epoch": 2.31462771798814, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.0373294353485107, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7296924591064453, + "num_tokens": 545407587.0, + "step": 21077 + }, + { + "epoch": 2.314737535690753, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.440424919128418, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7444148063659668, + "num_tokens": 545432350.0, + "step": 21078 + }, + { + "epoch": 2.314847353393367, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2913661003112793, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7465088963508606, + "num_tokens": 545460827.0, + "step": 21079 + }, + { + "epoch": 2.3149571710959806, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4688632488250732, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7374141216278076, + "num_tokens": 545484607.0, + "step": 21080 + }, + { + "epoch": 2.3150669887985944, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.385314702987671, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7249491810798645, + "num_tokens": 545508854.0, + "step": 21081 + }, + { + "epoch": 2.315176806501208, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3404529094696045, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7024128437042236, + "num_tokens": 545536439.0, + "step": 21082 + }, + { + "epoch": 2.3152866242038215, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3583931922912598, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7280325889587402, + "num_tokens": 545562043.0, + "step": 21083 + }, + { + "epoch": 2.3153964419064352, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2241570949554443, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7312759160995483, + "num_tokens": 545589760.0, + "step": 21084 + }, + { + "epoch": 2.315506259609049, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3401660919189453, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7408601641654968, + "num_tokens": 545614440.0, + "step": 21085 + }, + { + "epoch": 2.3156160773116627, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3311872482299805, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.734774649143219, + "num_tokens": 545639962.0, + "step": 21086 + }, + { + "epoch": 2.3157258950142765, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.6604511737823486, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7165675163269043, + "num_tokens": 545663324.0, + "step": 21087 + }, + { + "epoch": 2.31583571271689, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.622800588607788, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7513324022293091, + "num_tokens": 545685668.0, + "step": 21088 + }, + { + "epoch": 2.3159455304195036, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4818153381347656, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7278296947479248, + "num_tokens": 545709631.0, + "step": 21089 + }, + { + "epoch": 2.3160553481221173, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.514793634414673, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.746183454990387, + "num_tokens": 545733222.0, + "step": 21090 + }, + { + "epoch": 2.316165165824731, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.351407051086426, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7177589535713196, + "num_tokens": 545759942.0, + "step": 21091 + }, + { + "epoch": 2.316274983527345, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3215432167053223, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.746152400970459, + "num_tokens": 545785642.0, + "step": 21092 + }, + { + "epoch": 2.316384801229958, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2803781032562256, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7374194264411926, + "num_tokens": 545814865.0, + "step": 21093 + }, + { + "epoch": 2.316494618932572, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.500138282775879, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7232791185379028, + "num_tokens": 545838748.0, + "step": 21094 + }, + { + "epoch": 2.3166044366351857, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 1.938615083694458, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7146459817886353, + "num_tokens": 545876189.0, + "step": 21095 + }, + { + "epoch": 2.3167142543377994, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.2080981731414795, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7134280204772949, + "num_tokens": 545906308.0, + "step": 21096 + }, + { + "epoch": 2.3168240720404127, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3469135761260986, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7354931831359863, + "num_tokens": 545931183.0, + "step": 21097 + }, + { + "epoch": 2.3169338897430265, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.234771728515625, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7488225102424622, + "num_tokens": 545958608.0, + "step": 21098 + }, + { + "epoch": 2.3170437074456403, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.1425671577453613, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7241487503051758, + "num_tokens": 545988236.0, + "step": 21099 + }, + { + "epoch": 2.317153525148254, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.282726764678955, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7248983979225159, + "num_tokens": 546016863.0, + "step": 21100 + }, + { + "epoch": 2.3172633428508673, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.6236472129821777, + "learning_rate": 1e-06, + "loss": 0.7527, + "mean_token_accuracy": 0.7575992345809937, + "num_tokens": 546038858.0, + "step": 21101 + }, + { + "epoch": 2.317373160553481, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3779773712158203, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7435813546180725, + "num_tokens": 546064374.0, + "step": 21102 + }, + { + "epoch": 2.317482978256095, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3043859004974365, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7394083738327026, + "num_tokens": 546091560.0, + "step": 21103 + }, + { + "epoch": 2.3175927959587086, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.223900079727173, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7355490922927856, + "num_tokens": 546121413.0, + "step": 21104 + }, + { + "epoch": 2.3177026136613224, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3811371326446533, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7065465450286865, + "num_tokens": 546146179.0, + "step": 21105 + }, + { + "epoch": 2.3178124313639357, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.464179277420044, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7487630248069763, + "num_tokens": 546169669.0, + "step": 21106 + }, + { + "epoch": 2.3179222490665494, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.275275468826294, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7085909843444824, + "num_tokens": 546196400.0, + "step": 21107 + }, + { + "epoch": 2.318032066769163, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.613276720046997, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7341783046722412, + "num_tokens": 546219794.0, + "step": 21108 + }, + { + "epoch": 2.318141884471777, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.090726852416992, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.710729718208313, + "num_tokens": 546252721.0, + "step": 21109 + }, + { + "epoch": 2.3182517021743907, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.477625608444214, + "learning_rate": 1e-06, + "loss": 0.765, + "mean_token_accuracy": 0.7607305645942688, + "num_tokens": 546275903.0, + "step": 21110 + }, + { + "epoch": 2.318361519877004, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.8688583374023438, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7415030002593994, + "num_tokens": 546296258.0, + "step": 21111 + }, + { + "epoch": 2.3184713375796178, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4738361835479736, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7239997386932373, + "num_tokens": 546320414.0, + "step": 21112 + }, + { + "epoch": 2.3185811552822315, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.4885406494140625, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7535490393638611, + "num_tokens": 546341684.0, + "step": 21113 + }, + { + "epoch": 2.3186909729848453, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.5523507595062256, + "learning_rate": 1e-06, + "loss": 0.7425, + "mean_token_accuracy": 0.7584068775177002, + "num_tokens": 546362874.0, + "step": 21114 + }, + { + "epoch": 2.318800790687459, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.3583590984344482, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7182630300521851, + "num_tokens": 546390052.0, + "step": 21115 + }, + { + "epoch": 2.3189106083900723, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.18648099899292, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7219433784484863, + "num_tokens": 546419092.0, + "step": 21116 + }, + { + "epoch": 2.319020426092686, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.57525897026062, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7320430278778076, + "num_tokens": 546441330.0, + "step": 21117 + }, + { + "epoch": 2.3191302437953, + "ewc_loss": 0.00018405914306640625, + "grad_norm": 2.6350338459014893, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7343176007270813, + "num_tokens": 546462439.0, + "step": 21118 + }, + { + "epoch": 2.3192400614979136, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4151058197021484, + "learning_rate": 1e-06, + "loss": 0.7412, + "mean_token_accuracy": 0.758162260055542, + "num_tokens": 546487977.0, + "step": 21119 + }, + { + "epoch": 2.319349879200527, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.989346742630005, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7498809099197388, + "num_tokens": 546505219.0, + "step": 21120 + }, + { + "epoch": 2.3194596969031407, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.1500673294067383, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.733796238899231, + "num_tokens": 546533972.0, + "step": 21121 + }, + { + "epoch": 2.3195695146057544, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4495105743408203, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7163541316986084, + "num_tokens": 546559537.0, + "step": 21122 + }, + { + "epoch": 2.319679332308368, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4982235431671143, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7129335999488831, + "num_tokens": 546583319.0, + "step": 21123 + }, + { + "epoch": 2.319789150010982, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.25146484375, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7298789024353027, + "num_tokens": 546609774.0, + "step": 21124 + }, + { + "epoch": 2.3198989677135953, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2145485877990723, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7282116413116455, + "num_tokens": 546637470.0, + "step": 21125 + }, + { + "epoch": 2.320008785416209, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3309977054595947, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.745242714881897, + "num_tokens": 546663033.0, + "step": 21126 + }, + { + "epoch": 2.320118603118823, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.6185617446899414, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6927710771560669, + "num_tokens": 546686081.0, + "step": 21127 + }, + { + "epoch": 2.3202284208214365, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.420226573944092, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7445802092552185, + "num_tokens": 546711171.0, + "step": 21128 + }, + { + "epoch": 2.32033823852405, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5922691822052, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7139737606048584, + "num_tokens": 546733502.0, + "step": 21129 + }, + { + "epoch": 2.3204480562266636, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2476253509521484, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7182259559631348, + "num_tokens": 546762358.0, + "step": 21130 + }, + { + "epoch": 2.3205578739292774, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4515626430511475, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7296265363693237, + "num_tokens": 546787678.0, + "step": 21131 + }, + { + "epoch": 2.320667691631891, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.131352186203003, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7230484485626221, + "num_tokens": 546818176.0, + "step": 21132 + }, + { + "epoch": 2.320777509334505, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4573638439178467, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7340105772018433, + "num_tokens": 546842222.0, + "step": 21133 + }, + { + "epoch": 2.320887327037118, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.426992416381836, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7303122282028198, + "num_tokens": 546866707.0, + "step": 21134 + }, + { + "epoch": 2.320997144739732, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4199557304382324, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7415086627006531, + "num_tokens": 546891770.0, + "step": 21135 + }, + { + "epoch": 2.3211069624423457, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.7100918292999268, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7259196639060974, + "num_tokens": 546915119.0, + "step": 21136 + }, + { + "epoch": 2.3212167801449595, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.223046064376831, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7146725654602051, + "num_tokens": 546946013.0, + "step": 21137 + }, + { + "epoch": 2.3213265978475732, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.448373317718506, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7454865574836731, + "num_tokens": 546971462.0, + "step": 21138 + }, + { + "epoch": 2.3214364155501865, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.260361909866333, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7269382476806641, + "num_tokens": 547000981.0, + "step": 21139 + }, + { + "epoch": 2.3215462332528003, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.541550397872925, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7440166473388672, + "num_tokens": 547025222.0, + "step": 21140 + }, + { + "epoch": 2.321656050955414, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.53953218460083, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7308322191238403, + "num_tokens": 547049261.0, + "step": 21141 + }, + { + "epoch": 2.321765868658028, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.456505298614502, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7455177307128906, + "num_tokens": 547071238.0, + "step": 21142 + }, + { + "epoch": 2.3218756863606416, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2047922611236572, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7177751660346985, + "num_tokens": 547102348.0, + "step": 21143 + }, + { + "epoch": 2.321985504063255, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2798268795013428, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.720707893371582, + "num_tokens": 547130826.0, + "step": 21144 + }, + { + "epoch": 2.3220953217658686, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.7296292781829834, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7386579513549805, + "num_tokens": 547152390.0, + "step": 21145 + }, + { + "epoch": 2.3222051394684824, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.406461238861084, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7249880433082581, + "num_tokens": 547176962.0, + "step": 21146 + }, + { + "epoch": 2.322314957171096, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.345231771469116, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7308365106582642, + "num_tokens": 547201818.0, + "step": 21147 + }, + { + "epoch": 2.3224247748737095, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.344979763031006, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7388970255851746, + "num_tokens": 547227966.0, + "step": 21148 + }, + { + "epoch": 2.322534592576323, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.309507369995117, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7450077533721924, + "num_tokens": 547254343.0, + "step": 21149 + }, + { + "epoch": 2.322644410278937, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.407562494277954, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7379924058914185, + "num_tokens": 547281730.0, + "step": 21150 + }, + { + "epoch": 2.3227542279815507, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.1619648933410645, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7223342657089233, + "num_tokens": 547310001.0, + "step": 21151 + }, + { + "epoch": 2.322864045684164, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.7292473316192627, + "learning_rate": 1e-06, + "loss": 0.793, + "mean_token_accuracy": 0.748479962348938, + "num_tokens": 547331369.0, + "step": 21152 + }, + { + "epoch": 2.322973863386778, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3201332092285156, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7302915453910828, + "num_tokens": 547357263.0, + "step": 21153 + }, + { + "epoch": 2.3230836810893916, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.249439001083374, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7110366821289062, + "num_tokens": 547386568.0, + "step": 21154 + }, + { + "epoch": 2.3231934987920053, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 1.9650659561157227, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6852006316184998, + "num_tokens": 547422711.0, + "step": 21155 + }, + { + "epoch": 2.323303316494619, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3448779582977295, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7153369188308716, + "num_tokens": 547448999.0, + "step": 21156 + }, + { + "epoch": 2.3234131341972324, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4162402153015137, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7342200875282288, + "num_tokens": 547473272.0, + "step": 21157 + }, + { + "epoch": 2.323522951899846, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.811654567718506, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7369906902313232, + "num_tokens": 547492875.0, + "step": 21158 + }, + { + "epoch": 2.32363276960246, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.1049416065216064, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7184650897979736, + "num_tokens": 547526202.0, + "step": 21159 + }, + { + "epoch": 2.3237425873050737, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4600815773010254, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7308259606361389, + "num_tokens": 547551740.0, + "step": 21160 + }, + { + "epoch": 2.3238524050076874, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5837066173553467, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7188822031021118, + "num_tokens": 547576042.0, + "step": 21161 + }, + { + "epoch": 2.3239622227103007, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.209099054336548, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7155839204788208, + "num_tokens": 547605542.0, + "step": 21162 + }, + { + "epoch": 2.3240720404129145, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.0803639888763428, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7040771842002869, + "num_tokens": 547636849.0, + "step": 21163 + }, + { + "epoch": 2.3241818581155282, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5383102893829346, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7191107869148254, + "num_tokens": 547661953.0, + "step": 21164 + }, + { + "epoch": 2.324291675818142, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4768996238708496, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7439025640487671, + "num_tokens": 547685654.0, + "step": 21165 + }, + { + "epoch": 2.3244014935207558, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.342517375946045, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7145548462867737, + "num_tokens": 547712398.0, + "step": 21166 + }, + { + "epoch": 2.324511311223369, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.372036933898926, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7337709665298462, + "num_tokens": 547737811.0, + "step": 21167 + }, + { + "epoch": 2.324621128925983, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5099964141845703, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7296103239059448, + "num_tokens": 547762265.0, + "step": 21168 + }, + { + "epoch": 2.3247309466285966, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.6234912872314453, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7421283721923828, + "num_tokens": 547783120.0, + "step": 21169 + }, + { + "epoch": 2.3248407643312103, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.1356170177459717, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.712257981300354, + "num_tokens": 547814700.0, + "step": 21170 + }, + { + "epoch": 2.324950582033824, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.355543375015259, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7108618021011353, + "num_tokens": 547842383.0, + "step": 21171 + }, + { + "epoch": 2.3250603997364374, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.41693115234375, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7205989360809326, + "num_tokens": 547868389.0, + "step": 21172 + }, + { + "epoch": 2.325170217439051, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2379517555236816, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7197451591491699, + "num_tokens": 547898639.0, + "step": 21173 + }, + { + "epoch": 2.325280035141665, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.347264528274536, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7351202964782715, + "num_tokens": 547925825.0, + "step": 21174 + }, + { + "epoch": 2.3253898528442787, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5342185497283936, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7326246500015259, + "num_tokens": 547949645.0, + "step": 21175 + }, + { + "epoch": 2.325499670546892, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.63783860206604, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7176526784896851, + "num_tokens": 547974075.0, + "step": 21176 + }, + { + "epoch": 2.3256094882495058, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3826816082000732, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7296886444091797, + "num_tokens": 548000066.0, + "step": 21177 + }, + { + "epoch": 2.3257193059521195, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2485921382904053, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7101717591285706, + "num_tokens": 548028726.0, + "step": 21178 + }, + { + "epoch": 2.3258291236547333, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.514359712600708, + "learning_rate": 1e-06, + "loss": 0.7188, + "mean_token_accuracy": 0.7605031728744507, + "num_tokens": 548050967.0, + "step": 21179 + }, + { + "epoch": 2.3259389413573466, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.473050117492676, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7283721566200256, + "num_tokens": 548075738.0, + "step": 21180 + }, + { + "epoch": 2.3260487590599603, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.6133313179016113, + "learning_rate": 1e-06, + "loss": 0.7302, + "mean_token_accuracy": 0.7673159837722778, + "num_tokens": 548097259.0, + "step": 21181 + }, + { + "epoch": 2.326158576762574, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.198892831802368, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.715187668800354, + "num_tokens": 548128313.0, + "step": 21182 + }, + { + "epoch": 2.326268394465188, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.410463333129883, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7243500351905823, + "num_tokens": 548155143.0, + "step": 21183 + }, + { + "epoch": 2.3263782121678016, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.7003958225250244, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7382134199142456, + "num_tokens": 548176848.0, + "step": 21184 + }, + { + "epoch": 2.326488029870415, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.801417350769043, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.743111252784729, + "num_tokens": 548196503.0, + "step": 21185 + }, + { + "epoch": 2.3265978475730287, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.889932155609131, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7415053248405457, + "num_tokens": 548214433.0, + "step": 21186 + }, + { + "epoch": 2.3267076652756424, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4490015506744385, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7124450206756592, + "num_tokens": 548242208.0, + "step": 21187 + }, + { + "epoch": 2.326817482978256, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.398467779159546, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7299774885177612, + "num_tokens": 548267871.0, + "step": 21188 + }, + { + "epoch": 2.32692730068087, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.913855791091919, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7366923093795776, + "num_tokens": 548285796.0, + "step": 21189 + }, + { + "epoch": 2.3270371183834833, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.418391704559326, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7248037457466125, + "num_tokens": 548314029.0, + "step": 21190 + }, + { + "epoch": 2.327146936086097, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4179937839508057, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7574217319488525, + "num_tokens": 548339195.0, + "step": 21191 + }, + { + "epoch": 2.3272567537887108, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4307861328125, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7266450524330139, + "num_tokens": 548363915.0, + "step": 21192 + }, + { + "epoch": 2.3273665714913245, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.2535579204559326, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.72429358959198, + "num_tokens": 548392975.0, + "step": 21193 + }, + { + "epoch": 2.3274763891939383, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.1228389739990234, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7052785158157349, + "num_tokens": 548423278.0, + "step": 21194 + }, + { + "epoch": 2.3275862068965516, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3906102180480957, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7352256774902344, + "num_tokens": 548451510.0, + "step": 21195 + }, + { + "epoch": 2.3276960245991654, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3546841144561768, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7392377853393555, + "num_tokens": 548477218.0, + "step": 21196 + }, + { + "epoch": 2.327805842301779, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3466477394104004, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7475371360778809, + "num_tokens": 548503026.0, + "step": 21197 + }, + { + "epoch": 2.327915660004393, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.653270721435547, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7391581535339355, + "num_tokens": 548525537.0, + "step": 21198 + }, + { + "epoch": 2.328025477707006, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3450639247894287, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7157481908798218, + "num_tokens": 548554097.0, + "step": 21199 + }, + { + "epoch": 2.32813529540962, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.4771885871887207, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.734538197517395, + "num_tokens": 548578518.0, + "step": 21200 + }, + { + "epoch": 2.3282451131122337, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.372393846511841, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7330926060676575, + "num_tokens": 548603344.0, + "step": 21201 + }, + { + "epoch": 2.3283549308148475, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.251887798309326, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7465004920959473, + "num_tokens": 548627826.0, + "step": 21202 + }, + { + "epoch": 2.3284647485174608, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3975119590759277, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7410457730293274, + "num_tokens": 548652117.0, + "step": 21203 + }, + { + "epoch": 2.3285745662200745, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.436368227005005, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7472084760665894, + "num_tokens": 548677237.0, + "step": 21204 + }, + { + "epoch": 2.3286843839226883, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.0972344875335693, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7260587215423584, + "num_tokens": 548709854.0, + "step": 21205 + }, + { + "epoch": 2.328794201625302, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3493363857269287, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7203587889671326, + "num_tokens": 548736894.0, + "step": 21206 + }, + { + "epoch": 2.328904019327916, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.510089635848999, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.739117443561554, + "num_tokens": 548758761.0, + "step": 21207 + }, + { + "epoch": 2.329013837030529, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.415069580078125, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7234640121459961, + "num_tokens": 548783443.0, + "step": 21208 + }, + { + "epoch": 2.329123654733143, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.39420747756958, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7179940938949585, + "num_tokens": 548809750.0, + "step": 21209 + }, + { + "epoch": 2.3292334724357566, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.147360324859619, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7263906598091125, + "num_tokens": 548839854.0, + "step": 21210 + }, + { + "epoch": 2.3293432901383704, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.379042863845825, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7547693848609924, + "num_tokens": 548864134.0, + "step": 21211 + }, + { + "epoch": 2.329453107840984, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.317018508911133, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.703754186630249, + "num_tokens": 548893718.0, + "step": 21212 + }, + { + "epoch": 2.3295629255435975, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3793835639953613, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7184597253799438, + "num_tokens": 548920479.0, + "step": 21213 + }, + { + "epoch": 2.329672743246211, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.6237990856170654, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7258331179618835, + "num_tokens": 548944177.0, + "step": 21214 + }, + { + "epoch": 2.329782560948825, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5539815425872803, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7157922983169556, + "num_tokens": 548969981.0, + "step": 21215 + }, + { + "epoch": 2.3298923786514387, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.666923761367798, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7177100777626038, + "num_tokens": 548991299.0, + "step": 21216 + }, + { + "epoch": 2.3300021963540525, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.5686140060424805, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7350472211837769, + "num_tokens": 549013991.0, + "step": 21217 + }, + { + "epoch": 2.330112014056666, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.360092878341675, + "learning_rate": 1e-06, + "loss": 0.7541, + "mean_token_accuracy": 0.7612103819847107, + "num_tokens": 549038032.0, + "step": 21218 + }, + { + "epoch": 2.3302218317592795, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.3844830989837646, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7071822285652161, + "num_tokens": 549064922.0, + "step": 21219 + }, + { + "epoch": 2.3303316494618933, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.0464296340942383, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6976900100708008, + "num_tokens": 549100841.0, + "step": 21220 + }, + { + "epoch": 2.330441467164507, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.521353244781494, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7168207168579102, + "num_tokens": 549125312.0, + "step": 21221 + }, + { + "epoch": 2.330551284867121, + "ewc_loss": 0.0001850128173828125, + "grad_norm": 2.8132622241973877, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7374616265296936, + "num_tokens": 549144977.0, + "step": 21222 + }, + { + "epoch": 2.330661102569734, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1030118465423584, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6970361471176147, + "num_tokens": 549178168.0, + "step": 21223 + }, + { + "epoch": 2.330770920272348, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4092962741851807, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7365713119506836, + "num_tokens": 549203052.0, + "step": 21224 + }, + { + "epoch": 2.3308807379749616, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4566867351531982, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7583061456680298, + "num_tokens": 549226087.0, + "step": 21225 + }, + { + "epoch": 2.3309905556775754, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 3.4386210441589355, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7539622783660889, + "num_tokens": 549251485.0, + "step": 21226 + }, + { + "epoch": 2.3311003733801887, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.270864725112915, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7170476317405701, + "num_tokens": 549280757.0, + "step": 21227 + }, + { + "epoch": 2.3312101910828025, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1218502521514893, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.727936863899231, + "num_tokens": 549311095.0, + "step": 21228 + }, + { + "epoch": 2.3313200087854162, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3992390632629395, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7349646091461182, + "num_tokens": 549336923.0, + "step": 21229 + }, + { + "epoch": 2.33142982648803, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.206406354904175, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.715241014957428, + "num_tokens": 549367414.0, + "step": 21230 + }, + { + "epoch": 2.3315396441906433, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5702953338623047, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7406207919120789, + "num_tokens": 549390980.0, + "step": 21231 + }, + { + "epoch": 2.331649461893257, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4364914894104004, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.737155020236969, + "num_tokens": 549415582.0, + "step": 21232 + }, + { + "epoch": 2.331759279595871, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3350284099578857, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7078267335891724, + "num_tokens": 549442023.0, + "step": 21233 + }, + { + "epoch": 2.3318690972984846, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2570321559906006, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.7426886558532715, + "num_tokens": 549468923.0, + "step": 21234 + }, + { + "epoch": 2.3319789150010983, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6241559982299805, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.714533269405365, + "num_tokens": 549493366.0, + "step": 21235 + }, + { + "epoch": 2.3320887327037116, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.196030378341675, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7235302925109863, + "num_tokens": 549521336.0, + "step": 21236 + }, + { + "epoch": 2.3321985504063254, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.400738000869751, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7144050598144531, + "num_tokens": 549547172.0, + "step": 21237 + }, + { + "epoch": 2.332308368108939, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.794548511505127, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.7531150579452515, + "num_tokens": 549566702.0, + "step": 21238 + }, + { + "epoch": 2.332418185811553, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.373619794845581, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7487086057662964, + "num_tokens": 549591628.0, + "step": 21239 + }, + { + "epoch": 2.3325280035141667, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2788429260253906, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7175983190536499, + "num_tokens": 549618880.0, + "step": 21240 + }, + { + "epoch": 2.33263782121678, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 1.949660062789917, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7260317802429199, + "num_tokens": 549652810.0, + "step": 21241 + }, + { + "epoch": 2.3327476389193937, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.7956578731536865, + "learning_rate": 1e-06, + "loss": 0.7135, + "mean_token_accuracy": 0.7710287570953369, + "num_tokens": 549671444.0, + "step": 21242 + }, + { + "epoch": 2.3328574566220075, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.811457872390747, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7509145140647888, + "num_tokens": 549691769.0, + "step": 21243 + }, + { + "epoch": 2.3329672743246213, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.533287763595581, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.729030966758728, + "num_tokens": 549715034.0, + "step": 21244 + }, + { + "epoch": 2.333077092027235, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2755558490753174, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7345253229141235, + "num_tokens": 549740360.0, + "step": 21245 + }, + { + "epoch": 2.3331869097298483, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.321035623550415, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.756413459777832, + "num_tokens": 549766285.0, + "step": 21246 + }, + { + "epoch": 2.333296727432462, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1649184226989746, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7089112997055054, + "num_tokens": 549796952.0, + "step": 21247 + }, + { + "epoch": 2.333406545135076, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.709512948989868, + "learning_rate": 1e-06, + "loss": 0.7816, + "mean_token_accuracy": 0.7500312328338623, + "num_tokens": 549816546.0, + "step": 21248 + }, + { + "epoch": 2.3335163628376896, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3754160404205322, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7342631816864014, + "num_tokens": 549840980.0, + "step": 21249 + }, + { + "epoch": 2.333626180540303, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.550320625305176, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7094833254814148, + "num_tokens": 549865581.0, + "step": 21250 + }, + { + "epoch": 2.3337359982429167, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.427879810333252, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7277160286903381, + "num_tokens": 549893505.0, + "step": 21251 + }, + { + "epoch": 2.3338458159455304, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.7238106727600098, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7373764514923096, + "num_tokens": 549913166.0, + "step": 21252 + }, + { + "epoch": 2.333955633648144, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5149636268615723, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.730292797088623, + "num_tokens": 549936725.0, + "step": 21253 + }, + { + "epoch": 2.334065451350758, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2136828899383545, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7343287467956543, + "num_tokens": 549965484.0, + "step": 21254 + }, + { + "epoch": 2.3341752690533712, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2881863117218018, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7318400144577026, + "num_tokens": 549990882.0, + "step": 21255 + }, + { + "epoch": 2.334285086755985, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3013241291046143, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7070558667182922, + "num_tokens": 550019734.0, + "step": 21256 + }, + { + "epoch": 2.3343949044585988, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6186490058898926, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7494927048683167, + "num_tokens": 550040980.0, + "step": 21257 + }, + { + "epoch": 2.3345047221612125, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3291969299316406, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7187551259994507, + "num_tokens": 550065783.0, + "step": 21258 + }, + { + "epoch": 2.334614539863826, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.155113935470581, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7481574416160583, + "num_tokens": 550094010.0, + "step": 21259 + }, + { + "epoch": 2.3347243575664396, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.7365236282348633, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7239400148391724, + "num_tokens": 550114344.0, + "step": 21260 + }, + { + "epoch": 2.3348341752690533, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.453793525695801, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7253565788269043, + "num_tokens": 550142710.0, + "step": 21261 + }, + { + "epoch": 2.334943992971667, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3465707302093506, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7395789623260498, + "num_tokens": 550168823.0, + "step": 21262 + }, + { + "epoch": 2.335053810674281, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.360071897506714, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7328848838806152, + "num_tokens": 550193547.0, + "step": 21263 + }, + { + "epoch": 2.335163628376894, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.331434488296509, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7331393957138062, + "num_tokens": 550219685.0, + "step": 21264 + }, + { + "epoch": 2.335273446079508, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1941726207733154, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7136473655700684, + "num_tokens": 550248353.0, + "step": 21265 + }, + { + "epoch": 2.3353832637821217, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1612749099731445, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7411676645278931, + "num_tokens": 550278798.0, + "step": 21266 + }, + { + "epoch": 2.3354930814847354, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4063096046447754, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7166508436203003, + "num_tokens": 550306111.0, + "step": 21267 + }, + { + "epoch": 2.335602899187349, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.269237995147705, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.74180006980896, + "num_tokens": 550333461.0, + "step": 21268 + }, + { + "epoch": 2.3357127168899625, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5716474056243896, + "learning_rate": 1e-06, + "loss": 0.678, + "mean_token_accuracy": 0.7775651812553406, + "num_tokens": 550354430.0, + "step": 21269 + }, + { + "epoch": 2.3358225345925763, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4997076988220215, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7277987599372864, + "num_tokens": 550379375.0, + "step": 21270 + }, + { + "epoch": 2.33593235229519, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.162214517593384, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7164812684059143, + "num_tokens": 550408939.0, + "step": 21271 + }, + { + "epoch": 2.336042169997804, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6744823455810547, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7355057597160339, + "num_tokens": 550430170.0, + "step": 21272 + }, + { + "epoch": 2.3361519877004175, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4850680828094482, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7545812129974365, + "num_tokens": 550453836.0, + "step": 21273 + }, + { + "epoch": 2.336261805403031, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.0809595584869385, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7202794551849365, + "num_tokens": 550485340.0, + "step": 21274 + }, + { + "epoch": 2.3363716231056446, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.0797293186187744, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7042649984359741, + "num_tokens": 550520121.0, + "step": 21275 + }, + { + "epoch": 2.3364814408082584, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5448243618011475, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7379501461982727, + "num_tokens": 550545685.0, + "step": 21276 + }, + { + "epoch": 2.336591258510872, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2085187435150146, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6946054697036743, + "num_tokens": 550576946.0, + "step": 21277 + }, + { + "epoch": 2.3367010762134854, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.56200909614563, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7490748167037964, + "num_tokens": 550598771.0, + "step": 21278 + }, + { + "epoch": 2.336810893916099, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.372756004333496, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7224529385566711, + "num_tokens": 550625923.0, + "step": 21279 + }, + { + "epoch": 2.336920711618713, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2209720611572266, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7212879657745361, + "num_tokens": 550654256.0, + "step": 21280 + }, + { + "epoch": 2.3370305293213267, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.162522554397583, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7170456051826477, + "num_tokens": 550684929.0, + "step": 21281 + }, + { + "epoch": 2.33714034702394, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.323604106903076, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7201634645462036, + "num_tokens": 550712460.0, + "step": 21282 + }, + { + "epoch": 2.337250164726554, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2469089031219482, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7317379713058472, + "num_tokens": 550740829.0, + "step": 21283 + }, + { + "epoch": 2.3373599824291675, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3896613121032715, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7300567626953125, + "num_tokens": 550767310.0, + "step": 21284 + }, + { + "epoch": 2.3374698001317813, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2172622680664062, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7421458959579468, + "num_tokens": 550793297.0, + "step": 21285 + }, + { + "epoch": 2.337579617834395, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.551847457885742, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7385814189910889, + "num_tokens": 550814235.0, + "step": 21286 + }, + { + "epoch": 2.3376894355370084, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.41877818107605, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7536134123802185, + "num_tokens": 550838545.0, + "step": 21287 + }, + { + "epoch": 2.337799253239622, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.597261905670166, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7282443046569824, + "num_tokens": 550861019.0, + "step": 21288 + }, + { + "epoch": 2.337909070942236, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4814863204956055, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.732103705406189, + "num_tokens": 550884329.0, + "step": 21289 + }, + { + "epoch": 2.3380188886448496, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.14296555519104, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7186828851699829, + "num_tokens": 550914916.0, + "step": 21290 + }, + { + "epoch": 2.3381287063474634, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6394920349121094, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7424759864807129, + "num_tokens": 550937653.0, + "step": 21291 + }, + { + "epoch": 2.3382385240500767, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3380134105682373, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7186881899833679, + "num_tokens": 550964748.0, + "step": 21292 + }, + { + "epoch": 2.3383483417526905, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4930684566497803, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7334623336791992, + "num_tokens": 550988746.0, + "step": 21293 + }, + { + "epoch": 2.338458159455304, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.350226640701294, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7438430190086365, + "num_tokens": 551012690.0, + "step": 21294 + }, + { + "epoch": 2.338567977157918, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.0380659103393555, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7369809150695801, + "num_tokens": 551046199.0, + "step": 21295 + }, + { + "epoch": 2.3386777948605317, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3562138080596924, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7224651575088501, + "num_tokens": 551072392.0, + "step": 21296 + }, + { + "epoch": 2.338787612563145, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4831392765045166, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7173667550086975, + "num_tokens": 551096181.0, + "step": 21297 + }, + { + "epoch": 2.338897430265759, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.325737237930298, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.6970545649528503, + "num_tokens": 551126784.0, + "step": 21298 + }, + { + "epoch": 2.3390072479683726, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.177325487136841, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7164695262908936, + "num_tokens": 551157975.0, + "step": 21299 + }, + { + "epoch": 2.3391170656709863, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.185288429260254, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7134495377540588, + "num_tokens": 551188296.0, + "step": 21300 + }, + { + "epoch": 2.3392268833735996, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5344858169555664, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.7481224536895752, + "num_tokens": 551209808.0, + "step": 21301 + }, + { + "epoch": 2.3393367010762134, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2084712982177734, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7277751564979553, + "num_tokens": 551237982.0, + "step": 21302 + }, + { + "epoch": 2.339446518778827, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3470516204833984, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7349346876144409, + "num_tokens": 551263064.0, + "step": 21303 + }, + { + "epoch": 2.339556336481441, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3922319412231445, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7398108243942261, + "num_tokens": 551290330.0, + "step": 21304 + }, + { + "epoch": 2.3396661541840547, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.126138925552368, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7371435165405273, + "num_tokens": 551324588.0, + "step": 21305 + }, + { + "epoch": 2.339775971886668, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5971853733062744, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7280758619308472, + "num_tokens": 551346696.0, + "step": 21306 + }, + { + "epoch": 2.3398857895892817, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3336479663848877, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7070226073265076, + "num_tokens": 551374637.0, + "step": 21307 + }, + { + "epoch": 2.3399956072918955, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2601144313812256, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7260125875473022, + "num_tokens": 551402293.0, + "step": 21308 + }, + { + "epoch": 2.3401054249945092, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.318650007247925, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7173920273780823, + "num_tokens": 551430867.0, + "step": 21309 + }, + { + "epoch": 2.3402152426971226, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5000076293945312, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7122589349746704, + "num_tokens": 551456102.0, + "step": 21310 + }, + { + "epoch": 2.3403250603997363, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.560481548309326, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7276815176010132, + "num_tokens": 551479576.0, + "step": 21311 + }, + { + "epoch": 2.34043487810235, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1528077125549316, + "learning_rate": 1e-06, + "loss": 0.7582, + "mean_token_accuracy": 0.7639869451522827, + "num_tokens": 551508248.0, + "step": 21312 + }, + { + "epoch": 2.340544695804964, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6459763050079346, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7268649339675903, + "num_tokens": 551529376.0, + "step": 21313 + }, + { + "epoch": 2.3406545135075776, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3038716316223145, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7396820783615112, + "num_tokens": 551554610.0, + "step": 21314 + }, + { + "epoch": 2.340764331210191, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6282799243927, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7307789325714111, + "num_tokens": 551579653.0, + "step": 21315 + }, + { + "epoch": 2.3408741489128047, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.398834705352783, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7371957302093506, + "num_tokens": 551603025.0, + "step": 21316 + }, + { + "epoch": 2.3409839666154184, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3296728134155273, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7251406311988831, + "num_tokens": 551631007.0, + "step": 21317 + }, + { + "epoch": 2.341093784318032, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2272372245788574, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.6968454122543335, + "num_tokens": 551662782.0, + "step": 21318 + }, + { + "epoch": 2.341203602020646, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.802999496459961, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7313909530639648, + "num_tokens": 551683837.0, + "step": 21319 + }, + { + "epoch": 2.3413134197232592, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5616023540496826, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7560505270957947, + "num_tokens": 551705496.0, + "step": 21320 + }, + { + "epoch": 2.341423237425873, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.669334888458252, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7248559594154358, + "num_tokens": 551728394.0, + "step": 21321 + }, + { + "epoch": 2.3415330551284868, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.666883945465088, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7440935373306274, + "num_tokens": 551748347.0, + "step": 21322 + }, + { + "epoch": 2.3416428728311005, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5018699169158936, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.740899384021759, + "num_tokens": 551771301.0, + "step": 21323 + }, + { + "epoch": 2.3417526905337143, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6171960830688477, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7416313886642456, + "num_tokens": 551791860.0, + "step": 21324 + }, + { + "epoch": 2.3418625082363276, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2553324699401855, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7306289076805115, + "num_tokens": 551820556.0, + "step": 21325 + }, + { + "epoch": 2.3419723259389413, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.4332199096679688, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7389087677001953, + "num_tokens": 551845083.0, + "step": 21326 + }, + { + "epoch": 2.342082143641555, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5568532943725586, + "learning_rate": 1e-06, + "loss": 0.7584, + "mean_token_accuracy": 0.7563755512237549, + "num_tokens": 551866820.0, + "step": 21327 + }, + { + "epoch": 2.342191961344169, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.423733711242676, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7383115291595459, + "num_tokens": 551890731.0, + "step": 21328 + }, + { + "epoch": 2.342301779046782, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1887476444244385, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7211006879806519, + "num_tokens": 551919352.0, + "step": 21329 + }, + { + "epoch": 2.342411596749396, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.442864418029785, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7213546633720398, + "num_tokens": 551946592.0, + "step": 21330 + }, + { + "epoch": 2.3425214144520097, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.491753578186035, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7197854518890381, + "num_tokens": 551972924.0, + "step": 21331 + }, + { + "epoch": 2.3426312321546234, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5544776916503906, + "learning_rate": 1e-06, + "loss": 0.7774, + "mean_token_accuracy": 0.7555409669876099, + "num_tokens": 551994666.0, + "step": 21332 + }, + { + "epoch": 2.3427410498572367, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3929686546325684, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7161646485328674, + "num_tokens": 552021422.0, + "step": 21333 + }, + { + "epoch": 2.3428508675598505, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.227020740509033, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7262066602706909, + "num_tokens": 552051876.0, + "step": 21334 + }, + { + "epoch": 2.3429606852624643, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5379867553710938, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.742152214050293, + "num_tokens": 552074281.0, + "step": 21335 + }, + { + "epoch": 2.343070502965078, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3151726722717285, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7330831289291382, + "num_tokens": 552101939.0, + "step": 21336 + }, + { + "epoch": 2.3431803206676918, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.183458089828491, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7343215942382812, + "num_tokens": 552133240.0, + "step": 21337 + }, + { + "epoch": 2.343290138370305, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2281410694122314, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7453203797340393, + "num_tokens": 552160652.0, + "step": 21338 + }, + { + "epoch": 2.343399956072919, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3189966678619385, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7220814228057861, + "num_tokens": 552188326.0, + "step": 21339 + }, + { + "epoch": 2.3435097737755326, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5140774250030518, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7504769563674927, + "num_tokens": 552211141.0, + "step": 21340 + }, + { + "epoch": 2.3436195914781464, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.6181447505950928, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7307288646697998, + "num_tokens": 552235120.0, + "step": 21341 + }, + { + "epoch": 2.34372940918076, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5076639652252197, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7451682686805725, + "num_tokens": 552257858.0, + "step": 21342 + }, + { + "epoch": 2.3438392268833734, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2612228393554688, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7385508418083191, + "num_tokens": 552286546.0, + "step": 21343 + }, + { + "epoch": 2.343949044585987, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.36796236038208, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7530645132064819, + "num_tokens": 552308667.0, + "step": 21344 + }, + { + "epoch": 2.344058862288601, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.141918659210205, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6908336877822876, + "num_tokens": 552342909.0, + "step": 21345 + }, + { + "epoch": 2.3441686799912147, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3028225898742676, + "learning_rate": 1e-06, + "loss": 0.7798, + "mean_token_accuracy": 0.7520727515220642, + "num_tokens": 552368696.0, + "step": 21346 + }, + { + "epoch": 2.3442784976938285, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.2925899028778076, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7059879899024963, + "num_tokens": 552395639.0, + "step": 21347 + }, + { + "epoch": 2.3443883153964418, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.5198187828063965, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7177682518959045, + "num_tokens": 552420544.0, + "step": 21348 + }, + { + "epoch": 2.3444981330990555, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.8012654781341553, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7429624795913696, + "num_tokens": 552439038.0, + "step": 21349 + }, + { + "epoch": 2.3446079508016693, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.3767170906066895, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7343486547470093, + "num_tokens": 552464579.0, + "step": 21350 + }, + { + "epoch": 2.344717768504283, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.1728460788726807, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7238640189170837, + "num_tokens": 552495383.0, + "step": 21351 + }, + { + "epoch": 2.344827586206897, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.301651954650879, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7215181589126587, + "num_tokens": 552520115.0, + "step": 21352 + }, + { + "epoch": 2.34493740390951, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.6432979106903076, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7331339120864868, + "num_tokens": 552541490.0, + "step": 21353 + }, + { + "epoch": 2.345047221612124, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.0935182571411133, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7247685790061951, + "num_tokens": 552571728.0, + "step": 21354 + }, + { + "epoch": 2.3451570393147376, + "ewc_loss": 0.00018596649169921875, + "grad_norm": 2.211301803588867, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7333559989929199, + "num_tokens": 552599496.0, + "step": 21355 + }, + { + "epoch": 2.3452668570173514, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.358799695968628, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7185003757476807, + "num_tokens": 552627444.0, + "step": 21356 + }, + { + "epoch": 2.3453766747199647, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.503787040710449, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7317993640899658, + "num_tokens": 552651759.0, + "step": 21357 + }, + { + "epoch": 2.3454864924225785, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.104315996170044, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7333729267120361, + "num_tokens": 552684943.0, + "step": 21358 + }, + { + "epoch": 2.345596310125192, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.485229969024658, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7285006046295166, + "num_tokens": 552708085.0, + "step": 21359 + }, + { + "epoch": 2.345706127827806, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.243621587753296, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7139306664466858, + "num_tokens": 552738251.0, + "step": 21360 + }, + { + "epoch": 2.3458159455304193, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.146486282348633, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7370191812515259, + "num_tokens": 552770263.0, + "step": 21361 + }, + { + "epoch": 2.345925763233033, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1688711643218994, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7151426076889038, + "num_tokens": 552799588.0, + "step": 21362 + }, + { + "epoch": 2.346035580935647, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2881011962890625, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7277218103408813, + "num_tokens": 552825885.0, + "step": 21363 + }, + { + "epoch": 2.3461453986382605, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.644912004470825, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7223501205444336, + "num_tokens": 552849550.0, + "step": 21364 + }, + { + "epoch": 2.3462552163408743, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.8009305000305176, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7477783560752869, + "num_tokens": 552870271.0, + "step": 21365 + }, + { + "epoch": 2.3463650340434876, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1474082469940186, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7427272200584412, + "num_tokens": 552897487.0, + "step": 21366 + }, + { + "epoch": 2.3464748517461014, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.452162265777588, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7145150899887085, + "num_tokens": 552931433.0, + "step": 21367 + }, + { + "epoch": 2.346584669448715, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.6367692947387695, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.735813558101654, + "num_tokens": 552952389.0, + "step": 21368 + }, + { + "epoch": 2.346694487151329, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2517786026000977, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7281062006950378, + "num_tokens": 552980525.0, + "step": 21369 + }, + { + "epoch": 2.3468043048539426, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2455272674560547, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.727374255657196, + "num_tokens": 553006900.0, + "step": 21370 + }, + { + "epoch": 2.346914122556556, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.334834098815918, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7233227491378784, + "num_tokens": 553032119.0, + "step": 21371 + }, + { + "epoch": 2.3470239402591697, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.747943639755249, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7523493766784668, + "num_tokens": 553051757.0, + "step": 21372 + }, + { + "epoch": 2.3471337579617835, + "ewc_loss": 0.000186920166015625, + "grad_norm": 3.020035982131958, + "learning_rate": 1e-06, + "loss": 0.7621, + "mean_token_accuracy": 0.7542965412139893, + "num_tokens": 553068911.0, + "step": 21373 + }, + { + "epoch": 2.3472435756643972, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4342129230499268, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7418466806411743, + "num_tokens": 553091522.0, + "step": 21374 + }, + { + "epoch": 2.347353393367011, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.7365658283233643, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.724785566329956, + "num_tokens": 553114261.0, + "step": 21375 + }, + { + "epoch": 2.3474632110696243, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4047932624816895, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7217801213264465, + "num_tokens": 553138696.0, + "step": 21376 + }, + { + "epoch": 2.347573028772238, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.522480010986328, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7104794383049011, + "num_tokens": 553164347.0, + "step": 21377 + }, + { + "epoch": 2.347682846474852, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.0890822410583496, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7046338319778442, + "num_tokens": 553196295.0, + "step": 21378 + }, + { + "epoch": 2.3477926641774656, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2374792098999023, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7183969616889954, + "num_tokens": 553224688.0, + "step": 21379 + }, + { + "epoch": 2.347902481880079, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4217774868011475, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.734933614730835, + "num_tokens": 553248386.0, + "step": 21380 + }, + { + "epoch": 2.3480122995826926, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3684725761413574, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7183297872543335, + "num_tokens": 553275638.0, + "step": 21381 + }, + { + "epoch": 2.3481221172853064, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.7093217372894287, + "learning_rate": 1e-06, + "loss": 0.7262, + "mean_token_accuracy": 0.7595661878585815, + "num_tokens": 553294839.0, + "step": 21382 + }, + { + "epoch": 2.34823193498792, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5989043712615967, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7526414394378662, + "num_tokens": 553316714.0, + "step": 21383 + }, + { + "epoch": 2.3483417526905335, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.159304618835449, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7206140756607056, + "num_tokens": 553346341.0, + "step": 21384 + }, + { + "epoch": 2.3484515703931472, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5632104873657227, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7234536409378052, + "num_tokens": 553368820.0, + "step": 21385 + }, + { + "epoch": 2.348561388095761, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5869858264923096, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7515652179718018, + "num_tokens": 553390215.0, + "step": 21386 + }, + { + "epoch": 2.3486712057983747, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.394099235534668, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7254511117935181, + "num_tokens": 553416577.0, + "step": 21387 + }, + { + "epoch": 2.3487810235009885, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2727699279785156, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7434649467468262, + "num_tokens": 553443738.0, + "step": 21388 + }, + { + "epoch": 2.348890841203602, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.6165595054626465, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7205635905265808, + "num_tokens": 553465023.0, + "step": 21389 + }, + { + "epoch": 2.3490006589062156, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5099904537200928, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7101802825927734, + "num_tokens": 553489967.0, + "step": 21390 + }, + { + "epoch": 2.3491104766088293, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.336012125015259, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7438822388648987, + "num_tokens": 553516728.0, + "step": 21391 + }, + { + "epoch": 2.349220294311443, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.510495662689209, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7440100312232971, + "num_tokens": 553538784.0, + "step": 21392 + }, + { + "epoch": 2.349330112014057, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1882846355438232, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.6925976872444153, + "num_tokens": 553568410.0, + "step": 21393 + }, + { + "epoch": 2.34943992971667, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.222370147705078, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7482786178588867, + "num_tokens": 553597423.0, + "step": 21394 + }, + { + "epoch": 2.349549747419284, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5910756587982178, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7325121164321899, + "num_tokens": 553619223.0, + "step": 21395 + }, + { + "epoch": 2.3496595651218977, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1047487258911133, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7172455191612244, + "num_tokens": 553648604.0, + "step": 21396 + }, + { + "epoch": 2.3497693828245114, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.7867515087127686, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7411143779754639, + "num_tokens": 553668209.0, + "step": 21397 + }, + { + "epoch": 2.349879200527125, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3502819538116455, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7317674160003662, + "num_tokens": 553693394.0, + "step": 21398 + }, + { + "epoch": 2.3499890182297385, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.643017292022705, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7342572212219238, + "num_tokens": 553714622.0, + "step": 21399 + }, + { + "epoch": 2.3500988359323522, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4130613803863525, + "learning_rate": 1e-06, + "loss": 0.7626, + "mean_token_accuracy": 0.7525095343589783, + "num_tokens": 553738769.0, + "step": 21400 + }, + { + "epoch": 2.350208653634966, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5577750205993652, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7056047916412354, + "num_tokens": 553762917.0, + "step": 21401 + }, + { + "epoch": 2.3503184713375798, + "ewc_loss": 0.000186920166015625, + "grad_norm": 1.9914451837539673, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7226420044898987, + "num_tokens": 553798724.0, + "step": 21402 + }, + { + "epoch": 2.3504282890401935, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.761965274810791, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7355441451072693, + "num_tokens": 553819757.0, + "step": 21403 + }, + { + "epoch": 2.350538106742807, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3507800102233887, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.732394814491272, + "num_tokens": 553846996.0, + "step": 21404 + }, + { + "epoch": 2.3506479244454206, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.5069777965545654, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7221384048461914, + "num_tokens": 553871149.0, + "step": 21405 + }, + { + "epoch": 2.3507577421480343, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.172841787338257, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7183307409286499, + "num_tokens": 553904582.0, + "step": 21406 + }, + { + "epoch": 2.350867559850648, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.6716959476470947, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7481521368026733, + "num_tokens": 553925932.0, + "step": 21407 + }, + { + "epoch": 2.3509773775532614, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4449150562286377, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7493129372596741, + "num_tokens": 553950760.0, + "step": 21408 + }, + { + "epoch": 2.351087195255875, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.2010793685913086, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.710185706615448, + "num_tokens": 553980795.0, + "step": 21409 + }, + { + "epoch": 2.351197012958489, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3054392337799072, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.7552660703659058, + "num_tokens": 554007473.0, + "step": 21410 + }, + { + "epoch": 2.3513068306611027, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1824324131011963, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7139374613761902, + "num_tokens": 554036706.0, + "step": 21411 + }, + { + "epoch": 2.351416648363716, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.7742366790771484, + "learning_rate": 1e-06, + "loss": 0.7406, + "mean_token_accuracy": 0.7659330368041992, + "num_tokens": 554055256.0, + "step": 21412 + }, + { + "epoch": 2.3515264660663298, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.4773333072662354, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7313778400421143, + "num_tokens": 554081341.0, + "step": 21413 + }, + { + "epoch": 2.3516362837689435, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.1175084114074707, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7275692224502563, + "num_tokens": 554110544.0, + "step": 21414 + }, + { + "epoch": 2.3517461014715573, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3060755729675293, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7297698855400085, + "num_tokens": 554136803.0, + "step": 21415 + }, + { + "epoch": 2.351855919174171, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.169733762741089, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7226229310035706, + "num_tokens": 554164323.0, + "step": 21416 + }, + { + "epoch": 2.3519657368767843, + "ewc_loss": 0.000186920166015625, + "grad_norm": 2.3052515983581543, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7446314692497253, + "num_tokens": 554188916.0, + "step": 21417 + }, + { + "epoch": 2.352075554579398, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.566676139831543, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7359145879745483, + "num_tokens": 554210815.0, + "step": 21418 + }, + { + "epoch": 2.352185372282012, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.6040518283843994, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7365392446517944, + "num_tokens": 554233278.0, + "step": 21419 + }, + { + "epoch": 2.3522951899846256, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5645549297332764, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7256569266319275, + "num_tokens": 554256319.0, + "step": 21420 + }, + { + "epoch": 2.3524050076872394, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.770871639251709, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7244437336921692, + "num_tokens": 554275103.0, + "step": 21421 + }, + { + "epoch": 2.3525148253898527, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.319304943084717, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7195726037025452, + "num_tokens": 554302972.0, + "step": 21422 + }, + { + "epoch": 2.3526246430924664, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.482224225997925, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7309821844100952, + "num_tokens": 554326597.0, + "step": 21423 + }, + { + "epoch": 2.35273446079508, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4350829124450684, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7118464708328247, + "num_tokens": 554350438.0, + "step": 21424 + }, + { + "epoch": 2.352844278497694, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.661330461502075, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.730143666267395, + "num_tokens": 554372166.0, + "step": 21425 + }, + { + "epoch": 2.3529540962003077, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5413742065429688, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7406012415885925, + "num_tokens": 554394105.0, + "step": 21426 + }, + { + "epoch": 2.353063913902921, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.425600290298462, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7223389744758606, + "num_tokens": 554419285.0, + "step": 21427 + }, + { + "epoch": 2.353173731605535, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.51434063911438, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7392592430114746, + "num_tokens": 554442757.0, + "step": 21428 + }, + { + "epoch": 2.3532835493081485, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3930559158325195, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7126051783561707, + "num_tokens": 554468867.0, + "step": 21429 + }, + { + "epoch": 2.3533933670107623, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.21968674659729, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7250233292579651, + "num_tokens": 554499879.0, + "step": 21430 + }, + { + "epoch": 2.3535031847133756, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.463986396789551, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7360201478004456, + "num_tokens": 554524143.0, + "step": 21431 + }, + { + "epoch": 2.3536130024159894, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.585205078125, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7371089458465576, + "num_tokens": 554546452.0, + "step": 21432 + }, + { + "epoch": 2.353722820118603, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3752799034118652, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7406229376792908, + "num_tokens": 554571337.0, + "step": 21433 + }, + { + "epoch": 2.353832637821217, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.539607048034668, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7326977252960205, + "num_tokens": 554595132.0, + "step": 21434 + }, + { + "epoch": 2.3539424555238306, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3864939212799072, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7231650352478027, + "num_tokens": 554622249.0, + "step": 21435 + }, + { + "epoch": 2.354052273226444, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.0694022178649902, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7167585492134094, + "num_tokens": 554654011.0, + "step": 21436 + }, + { + "epoch": 2.3541620909290577, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2972280979156494, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7321362495422363, + "num_tokens": 554681248.0, + "step": 21437 + }, + { + "epoch": 2.3542719086316715, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3489766120910645, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7427998781204224, + "num_tokens": 554704839.0, + "step": 21438 + }, + { + "epoch": 2.354381726334285, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.190188407897949, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7252285480499268, + "num_tokens": 554736566.0, + "step": 21439 + }, + { + "epoch": 2.3544915440368985, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5538861751556396, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7315764427185059, + "num_tokens": 554759453.0, + "step": 21440 + }, + { + "epoch": 2.3546013617395123, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3476932048797607, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7286527752876282, + "num_tokens": 554785923.0, + "step": 21441 + }, + { + "epoch": 2.354711179442126, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.6426737308502197, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7223668694496155, + "num_tokens": 554807141.0, + "step": 21442 + }, + { + "epoch": 2.35482099714474, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3960976600646973, + "learning_rate": 1e-06, + "loss": 0.7445, + "mean_token_accuracy": 0.7658316493034363, + "num_tokens": 554830120.0, + "step": 21443 + }, + { + "epoch": 2.3549308148473536, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.357642889022827, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7328806519508362, + "num_tokens": 554854882.0, + "step": 21444 + }, + { + "epoch": 2.355040632549967, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5150606632232666, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7019666433334351, + "num_tokens": 554879697.0, + "step": 21445 + }, + { + "epoch": 2.3551504502525806, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4898946285247803, + "learning_rate": 1e-06, + "loss": 0.7642, + "mean_token_accuracy": 0.7549340128898621, + "num_tokens": 554901742.0, + "step": 21446 + }, + { + "epoch": 2.3552602679551944, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3765695095062256, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7271019816398621, + "num_tokens": 554927477.0, + "step": 21447 + }, + { + "epoch": 2.355370085657808, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.9004886150360107, + "learning_rate": 1e-06, + "loss": 0.7519, + "mean_token_accuracy": 0.7649606466293335, + "num_tokens": 554945121.0, + "step": 21448 + }, + { + "epoch": 2.355479903360422, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4428272247314453, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7219170331954956, + "num_tokens": 554971482.0, + "step": 21449 + }, + { + "epoch": 2.355589721063035, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7474329471588135, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7269675731658936, + "num_tokens": 554991702.0, + "step": 21450 + }, + { + "epoch": 2.355699538765649, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.510697364807129, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7475059032440186, + "num_tokens": 555012775.0, + "step": 21451 + }, + { + "epoch": 2.3558093564682627, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2835206985473633, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7248789072036743, + "num_tokens": 555042151.0, + "step": 21452 + }, + { + "epoch": 2.3559191741708765, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.215798854827881, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7295529842376709, + "num_tokens": 555069862.0, + "step": 21453 + }, + { + "epoch": 2.3560289918734902, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5948431491851807, + "learning_rate": 1e-06, + "loss": 0.7976, + "mean_token_accuracy": 0.7471498250961304, + "num_tokens": 555092922.0, + "step": 21454 + }, + { + "epoch": 2.3561388095761036, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2439632415771484, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7383540272712708, + "num_tokens": 555119169.0, + "step": 21455 + }, + { + "epoch": 2.3562486272787173, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3175156116485596, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7265537977218628, + "num_tokens": 555145956.0, + "step": 21456 + }, + { + "epoch": 2.356358444981331, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2420644760131836, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7176060676574707, + "num_tokens": 555170733.0, + "step": 21457 + }, + { + "epoch": 2.356468262683945, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.632155656814575, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7556486129760742, + "num_tokens": 555190957.0, + "step": 21458 + }, + { + "epoch": 2.356578080386558, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.776719331741333, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7331634163856506, + "num_tokens": 555211908.0, + "step": 21459 + }, + { + "epoch": 2.356687898089172, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4478213787078857, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7258744239807129, + "num_tokens": 555236327.0, + "step": 21460 + }, + { + "epoch": 2.3567977157917857, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.545797824859619, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7363601326942444, + "num_tokens": 555260806.0, + "step": 21461 + }, + { + "epoch": 2.3569075334943994, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3465054035186768, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7243814468383789, + "num_tokens": 555290160.0, + "step": 21462 + }, + { + "epoch": 2.3570173511970127, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.739436388015747, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7368594408035278, + "num_tokens": 555310957.0, + "step": 21463 + }, + { + "epoch": 2.3571271688996265, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4940905570983887, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7265214920043945, + "num_tokens": 555335161.0, + "step": 21464 + }, + { + "epoch": 2.3572369866022402, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2627806663513184, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7224042415618896, + "num_tokens": 555362095.0, + "step": 21465 + }, + { + "epoch": 2.357346804304854, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.5094494819641113, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7438493967056274, + "num_tokens": 555384986.0, + "step": 21466 + }, + { + "epoch": 2.3574566220074678, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.532932996749878, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7378606796264648, + "num_tokens": 555410007.0, + "step": 21467 + }, + { + "epoch": 2.357566439710081, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.1786763668060303, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7222604155540466, + "num_tokens": 555441338.0, + "step": 21468 + }, + { + "epoch": 2.357676257412695, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2612130641937256, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7149434089660645, + "num_tokens": 555470582.0, + "step": 21469 + }, + { + "epoch": 2.3577860751153086, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.0827648639678955, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7309892177581787, + "num_tokens": 555503886.0, + "step": 21470 + }, + { + "epoch": 2.3578958928179223, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4584531784057617, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7339529395103455, + "num_tokens": 555528617.0, + "step": 21471 + }, + { + "epoch": 2.358005710520536, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3156707286834717, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7392900586128235, + "num_tokens": 555555610.0, + "step": 21472 + }, + { + "epoch": 2.3581155282231494, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.328542947769165, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7038942575454712, + "num_tokens": 555581989.0, + "step": 21473 + }, + { + "epoch": 2.358225345925763, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.879009962081909, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.731399416923523, + "num_tokens": 555604084.0, + "step": 21474 + }, + { + "epoch": 2.358335163628377, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2564899921417236, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7436540126800537, + "num_tokens": 555631310.0, + "step": 21475 + }, + { + "epoch": 2.3584449813309907, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.248213052749634, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7198880910873413, + "num_tokens": 555661471.0, + "step": 21476 + }, + { + "epoch": 2.3585547990336044, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.2741873264312744, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7506344318389893, + "num_tokens": 555686611.0, + "step": 21477 + }, + { + "epoch": 2.3586646167362177, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.619227409362793, + "learning_rate": 1e-06, + "loss": 0.7962, + "mean_token_accuracy": 0.7439213395118713, + "num_tokens": 555706852.0, + "step": 21478 + }, + { + "epoch": 2.3587744344388315, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3670294284820557, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7555726170539856, + "num_tokens": 555731573.0, + "step": 21479 + }, + { + "epoch": 2.3588842521414453, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.397994041442871, + "learning_rate": 1e-06, + "loss": 0.7655, + "mean_token_accuracy": 0.752701461315155, + "num_tokens": 555754399.0, + "step": 21480 + }, + { + "epoch": 2.358994069844059, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4474315643310547, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7170953750610352, + "num_tokens": 555779355.0, + "step": 21481 + }, + { + "epoch": 2.3591038875466728, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.543104887008667, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7423462867736816, + "num_tokens": 555801736.0, + "step": 21482 + }, + { + "epoch": 2.359213705249286, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.433321475982666, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7269901037216187, + "num_tokens": 555828488.0, + "step": 21483 + }, + { + "epoch": 2.3593235229519, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.4446358680725098, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.755279004573822, + "num_tokens": 555853675.0, + "step": 21484 + }, + { + "epoch": 2.3594333406545136, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.469212293624878, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7417236566543579, + "num_tokens": 555877499.0, + "step": 21485 + }, + { + "epoch": 2.3595431583571274, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.228233814239502, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7267061471939087, + "num_tokens": 555904816.0, + "step": 21486 + }, + { + "epoch": 2.3596529760597407, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.438681125640869, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7406055927276611, + "num_tokens": 555930791.0, + "step": 21487 + }, + { + "epoch": 2.3597627937623544, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4874229431152344, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7234137058258057, + "num_tokens": 555954763.0, + "step": 21488 + }, + { + "epoch": 2.359872611464968, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4731366634368896, + "learning_rate": 1e-06, + "loss": 0.7744, + "mean_token_accuracy": 0.758530855178833, + "num_tokens": 555978907.0, + "step": 21489 + }, + { + "epoch": 2.359982429167582, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.3449156284332275, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.728503406047821, + "num_tokens": 556004729.0, + "step": 21490 + }, + { + "epoch": 2.3600922468701953, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3555803298950195, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7233231067657471, + "num_tokens": 556031476.0, + "step": 21491 + }, + { + "epoch": 2.360202064572809, + "ewc_loss": 0.00018787384033203125, + "grad_norm": 2.498075008392334, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7336812019348145, + "num_tokens": 556054552.0, + "step": 21492 + }, + { + "epoch": 2.3603118822754228, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6402809619903564, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7137261629104614, + "num_tokens": 556079224.0, + "step": 21493 + }, + { + "epoch": 2.3604216999780365, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.702808141708374, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7448554039001465, + "num_tokens": 556099862.0, + "step": 21494 + }, + { + "epoch": 2.3605315176806503, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0781445503234863, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.71128249168396, + "num_tokens": 556131139.0, + "step": 21495 + }, + { + "epoch": 2.3606413353832636, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.422412157058716, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7420148253440857, + "num_tokens": 556154231.0, + "step": 21496 + }, + { + "epoch": 2.3607511530858774, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.184972047805786, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7243826389312744, + "num_tokens": 556185699.0, + "step": 21497 + }, + { + "epoch": 2.360860970788491, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1782782077789307, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7398738861083984, + "num_tokens": 556213295.0, + "step": 21498 + }, + { + "epoch": 2.360970788491105, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4922127723693848, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7176938056945801, + "num_tokens": 556236909.0, + "step": 21499 + }, + { + "epoch": 2.3610806061937186, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6250803470611572, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.738038182258606, + "num_tokens": 556260911.0, + "step": 21500 + }, + { + "epoch": 2.361190423896332, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.543002128601074, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.748602032661438, + "num_tokens": 556284155.0, + "step": 21501 + }, + { + "epoch": 2.3613002415989457, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.377135753631592, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7249516844749451, + "num_tokens": 556311007.0, + "step": 21502 + }, + { + "epoch": 2.3614100593015594, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.9352962970733643, + "learning_rate": 1e-06, + "loss": 0.71, + "mean_token_accuracy": 0.7728406190872192, + "num_tokens": 556327874.0, + "step": 21503 + }, + { + "epoch": 2.361519877004173, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.305710554122925, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7344576716423035, + "num_tokens": 556354663.0, + "step": 21504 + }, + { + "epoch": 2.361629694706787, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4217097759246826, + "learning_rate": 1e-06, + "loss": 0.7672, + "mean_token_accuracy": 0.7619701027870178, + "num_tokens": 556380131.0, + "step": 21505 + }, + { + "epoch": 2.3617395124094003, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0431244373321533, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7332528829574585, + "num_tokens": 556412535.0, + "step": 21506 + }, + { + "epoch": 2.361849330112014, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3185298442840576, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7252079248428345, + "num_tokens": 556438374.0, + "step": 21507 + }, + { + "epoch": 2.361959147814628, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.512037754058838, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.729277491569519, + "num_tokens": 556461616.0, + "step": 21508 + }, + { + "epoch": 2.3620689655172415, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4843242168426514, + "learning_rate": 1e-06, + "loss": 0.7962, + "mean_token_accuracy": 0.7396543025970459, + "num_tokens": 556483612.0, + "step": 21509 + }, + { + "epoch": 2.362178783219855, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3782172203063965, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.743802547454834, + "num_tokens": 556508354.0, + "step": 21510 + }, + { + "epoch": 2.3622886009224686, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4344677925109863, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7359191179275513, + "num_tokens": 556533953.0, + "step": 21511 + }, + { + "epoch": 2.3623984186250824, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.371105194091797, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7293730974197388, + "num_tokens": 556559964.0, + "step": 21512 + }, + { + "epoch": 2.362508236327696, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2156460285186768, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7200567722320557, + "num_tokens": 556591345.0, + "step": 21513 + }, + { + "epoch": 2.3626180540303094, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.830831289291382, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7353188395500183, + "num_tokens": 556610964.0, + "step": 21514 + }, + { + "epoch": 2.362727871732923, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.386594533920288, + "learning_rate": 1e-06, + "loss": 0.7922, + "mean_token_accuracy": 0.7498438358306885, + "num_tokens": 556635293.0, + "step": 21515 + }, + { + "epoch": 2.362837689435537, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2701778411865234, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7100131511688232, + "num_tokens": 556663735.0, + "step": 21516 + }, + { + "epoch": 2.3629475071381507, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.693612575531006, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7325855493545532, + "num_tokens": 556686187.0, + "step": 21517 + }, + { + "epoch": 2.3630573248407645, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3074636459350586, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7246593236923218, + "num_tokens": 556714669.0, + "step": 21518 + }, + { + "epoch": 2.363167142543378, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.285353660583496, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.726710319519043, + "num_tokens": 556741696.0, + "step": 21519 + }, + { + "epoch": 2.3632769602459915, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4724843502044678, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.73780357837677, + "num_tokens": 556763765.0, + "step": 21520 + }, + { + "epoch": 2.3633867779486053, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2927327156066895, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7248884439468384, + "num_tokens": 556790514.0, + "step": 21521 + }, + { + "epoch": 2.363496595651219, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5465986728668213, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7303314208984375, + "num_tokens": 556813619.0, + "step": 21522 + }, + { + "epoch": 2.363606413353833, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2479162216186523, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7427868843078613, + "num_tokens": 556841250.0, + "step": 21523 + }, + { + "epoch": 2.363716231056446, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.534672260284424, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.748455822467804, + "num_tokens": 556863528.0, + "step": 21524 + }, + { + "epoch": 2.36382604875906, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2119052410125732, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7258651256561279, + "num_tokens": 556892962.0, + "step": 21525 + }, + { + "epoch": 2.3639358664616736, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2737889289855957, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7225223779678345, + "num_tokens": 556920673.0, + "step": 21526 + }, + { + "epoch": 2.3640456841642874, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.590423822402954, + "learning_rate": 1e-06, + "loss": 0.7532, + "mean_token_accuracy": 0.7579083442687988, + "num_tokens": 556943449.0, + "step": 21527 + }, + { + "epoch": 2.364155501866901, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.215780258178711, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7181055545806885, + "num_tokens": 556972204.0, + "step": 21528 + }, + { + "epoch": 2.3642653195695145, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2948899269104004, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7060988545417786, + "num_tokens": 557002786.0, + "step": 21529 + }, + { + "epoch": 2.3643751372721282, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.571629047393799, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7078779935836792, + "num_tokens": 557028432.0, + "step": 21530 + }, + { + "epoch": 2.364484954974742, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.498098134994507, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.710591733455658, + "num_tokens": 557053635.0, + "step": 21531 + }, + { + "epoch": 2.3645947726773557, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4588022232055664, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7360249757766724, + "num_tokens": 557080238.0, + "step": 21532 + }, + { + "epoch": 2.3647045903799695, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4651176929473877, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7391620874404907, + "num_tokens": 557102780.0, + "step": 21533 + }, + { + "epoch": 2.364814408082583, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.29128098487854, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7203329801559448, + "num_tokens": 557131178.0, + "step": 21534 + }, + { + "epoch": 2.3649242257851966, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.609825849533081, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7408875226974487, + "num_tokens": 557153394.0, + "step": 21535 + }, + { + "epoch": 2.3650340434878103, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.274012565612793, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7287306189537048, + "num_tokens": 557180490.0, + "step": 21536 + }, + { + "epoch": 2.365143861190424, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1751976013183594, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7282623052597046, + "num_tokens": 557213714.0, + "step": 21537 + }, + { + "epoch": 2.3652536788930374, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.415088653564453, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7387352585792542, + "num_tokens": 557238581.0, + "step": 21538 + }, + { + "epoch": 2.365363496595651, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3518855571746826, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7179641723632812, + "num_tokens": 557264846.0, + "step": 21539 + }, + { + "epoch": 2.365473314298265, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.382711172103882, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7245599031448364, + "num_tokens": 557291147.0, + "step": 21540 + }, + { + "epoch": 2.3655831320008787, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.581102132797241, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7330069541931152, + "num_tokens": 557312834.0, + "step": 21541 + }, + { + "epoch": 2.365692949703492, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4076712131500244, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7441740036010742, + "num_tokens": 557339964.0, + "step": 21542 + }, + { + "epoch": 2.3658027674061057, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.343698024749756, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.742830753326416, + "num_tokens": 557366144.0, + "step": 21543 + }, + { + "epoch": 2.3659125851087195, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.362851858139038, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7224183082580566, + "num_tokens": 557391270.0, + "step": 21544 + }, + { + "epoch": 2.3660224028113332, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3217597007751465, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7292768955230713, + "num_tokens": 557417639.0, + "step": 21545 + }, + { + "epoch": 2.366132220513947, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4809799194335938, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7341212630271912, + "num_tokens": 557444990.0, + "step": 21546 + }, + { + "epoch": 2.3662420382165603, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.629650115966797, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7042980194091797, + "num_tokens": 557467022.0, + "step": 21547 + }, + { + "epoch": 2.366351855919174, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2923521995544434, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7214304208755493, + "num_tokens": 557495147.0, + "step": 21548 + }, + { + "epoch": 2.366461673621788, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.470088481903076, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7397514581680298, + "num_tokens": 557519017.0, + "step": 21549 + }, + { + "epoch": 2.3665714913244016, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4104549884796143, + "learning_rate": 1e-06, + "loss": 0.75, + "mean_token_accuracy": 0.7651922106742859, + "num_tokens": 557543080.0, + "step": 21550 + }, + { + "epoch": 2.3666813090270153, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.345618963241577, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7297227382659912, + "num_tokens": 557569638.0, + "step": 21551 + }, + { + "epoch": 2.3667911267296287, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1714723110198975, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7432358264923096, + "num_tokens": 557598536.0, + "step": 21552 + }, + { + "epoch": 2.3669009444322424, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.021143913269043, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7177760004997253, + "num_tokens": 557634436.0, + "step": 21553 + }, + { + "epoch": 2.367010762134856, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.8177332878112793, + "learning_rate": 1e-06, + "loss": 0.783, + "mean_token_accuracy": 0.7510706186294556, + "num_tokens": 557654703.0, + "step": 21554 + }, + { + "epoch": 2.36712057983747, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3175463676452637, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7244616150856018, + "num_tokens": 557683104.0, + "step": 21555 + }, + { + "epoch": 2.3672303975400837, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2201616764068604, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7144954204559326, + "num_tokens": 557715197.0, + "step": 21556 + }, + { + "epoch": 2.367340215242697, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.278170585632324, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7300068736076355, + "num_tokens": 557742148.0, + "step": 21557 + }, + { + "epoch": 2.3674500329453108, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.189321517944336, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7244969606399536, + "num_tokens": 557769927.0, + "step": 21558 + }, + { + "epoch": 2.3675598506479245, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5022528171539307, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.730121374130249, + "num_tokens": 557793071.0, + "step": 21559 + }, + { + "epoch": 2.3676696683505383, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1781816482543945, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7473270893096924, + "num_tokens": 557822869.0, + "step": 21560 + }, + { + "epoch": 2.3677794860531516, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2350358963012695, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.731187105178833, + "num_tokens": 557852209.0, + "step": 21561 + }, + { + "epoch": 2.3678893037557653, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4386167526245117, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7229024767875671, + "num_tokens": 557876654.0, + "step": 21562 + }, + { + "epoch": 2.367999121458379, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1347529888153076, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7390490770339966, + "num_tokens": 557906940.0, + "step": 21563 + }, + { + "epoch": 2.368108939160993, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.174187183380127, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.722187876701355, + "num_tokens": 557938472.0, + "step": 21564 + }, + { + "epoch": 2.368218756863606, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.635059356689453, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7276415824890137, + "num_tokens": 557960421.0, + "step": 21565 + }, + { + "epoch": 2.36832857456622, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.679624080657959, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7421846985816956, + "num_tokens": 557981998.0, + "step": 21566 + }, + { + "epoch": 2.3684383922688337, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5484566688537598, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7380270957946777, + "num_tokens": 558004848.0, + "step": 21567 + }, + { + "epoch": 2.3685482099714474, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.527188301086426, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7418701648712158, + "num_tokens": 558028237.0, + "step": 21568 + }, + { + "epoch": 2.368658027674061, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2259342670440674, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7222911715507507, + "num_tokens": 558058594.0, + "step": 21569 + }, + { + "epoch": 2.3687678453766745, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.317756414413452, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7531577944755554, + "num_tokens": 558084242.0, + "step": 21570 + }, + { + "epoch": 2.3688776630792883, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.131026268005371, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7003591060638428, + "num_tokens": 558115959.0, + "step": 21571 + }, + { + "epoch": 2.368987480781902, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3552610874176025, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7207602262496948, + "num_tokens": 558141050.0, + "step": 21572 + }, + { + "epoch": 2.369097298484516, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.13523268699646, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6918342709541321, + "num_tokens": 558173423.0, + "step": 21573 + }, + { + "epoch": 2.3692071161871295, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3673782348632812, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7357696890830994, + "num_tokens": 558197401.0, + "step": 21574 + }, + { + "epoch": 2.369316933889743, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6197102069854736, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7392077445983887, + "num_tokens": 558220872.0, + "step": 21575 + }, + { + "epoch": 2.3694267515923566, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4146389961242676, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7296780347824097, + "num_tokens": 558245340.0, + "step": 21576 + }, + { + "epoch": 2.3695365692949704, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3504817485809326, + "learning_rate": 1e-06, + "loss": 0.7801, + "mean_token_accuracy": 0.7504611611366272, + "num_tokens": 558269052.0, + "step": 21577 + }, + { + "epoch": 2.369646386997584, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.486910581588745, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7367203235626221, + "num_tokens": 558296135.0, + "step": 21578 + }, + { + "epoch": 2.369756204700198, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5506346225738525, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7236270308494568, + "num_tokens": 558319289.0, + "step": 21579 + }, + { + "epoch": 2.369866022402811, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6051783561706543, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.743094801902771, + "num_tokens": 558341681.0, + "step": 21580 + }, + { + "epoch": 2.369975840105425, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2540011405944824, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.707324206829071, + "num_tokens": 558369106.0, + "step": 21581 + }, + { + "epoch": 2.3700856578080387, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2347400188446045, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7321344614028931, + "num_tokens": 558398945.0, + "step": 21582 + }, + { + "epoch": 2.3701954755106525, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7380318641662598, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7420355081558228, + "num_tokens": 558422580.0, + "step": 21583 + }, + { + "epoch": 2.370305293213266, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2193801403045654, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7381946444511414, + "num_tokens": 558450847.0, + "step": 21584 + }, + { + "epoch": 2.3704151109158795, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.33725643157959, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7323102951049805, + "num_tokens": 558475667.0, + "step": 21585 + }, + { + "epoch": 2.3705249286184933, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2390353679656982, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7405695915222168, + "num_tokens": 558504062.0, + "step": 21586 + }, + { + "epoch": 2.370634746321107, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.534250259399414, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7375901937484741, + "num_tokens": 558527171.0, + "step": 21587 + }, + { + "epoch": 2.370744564023721, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.8120365142822266, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7419177293777466, + "num_tokens": 558546295.0, + "step": 21588 + }, + { + "epoch": 2.370854381726334, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.537984609603882, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7206941843032837, + "num_tokens": 558568236.0, + "step": 21589 + }, + { + "epoch": 2.370964199428948, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.179147958755493, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7034777402877808, + "num_tokens": 558598369.0, + "step": 21590 + }, + { + "epoch": 2.3710740171315616, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6420204639434814, + "learning_rate": 1e-06, + "loss": 0.7684, + "mean_token_accuracy": 0.7603543400764465, + "num_tokens": 558620381.0, + "step": 21591 + }, + { + "epoch": 2.3711838348341754, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6186885833740234, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7385354042053223, + "num_tokens": 558642173.0, + "step": 21592 + }, + { + "epoch": 2.3712936525367887, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4593777656555176, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7238370776176453, + "num_tokens": 558669174.0, + "step": 21593 + }, + { + "epoch": 2.3714034702394025, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.549473524093628, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.737199068069458, + "num_tokens": 558691184.0, + "step": 21594 + }, + { + "epoch": 2.371513287942016, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.930976629257202, + "learning_rate": 1e-06, + "loss": 0.7479, + "mean_token_accuracy": 0.7574672698974609, + "num_tokens": 558707598.0, + "step": 21595 + }, + { + "epoch": 2.37162310564463, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.048494815826416, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6925206184387207, + "num_tokens": 558744396.0, + "step": 21596 + }, + { + "epoch": 2.3717329233472437, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.464090347290039, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7429225444793701, + "num_tokens": 558769673.0, + "step": 21597 + }, + { + "epoch": 2.371842741049857, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4624292850494385, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7317423820495605, + "num_tokens": 558795749.0, + "step": 21598 + }, + { + "epoch": 2.371952558752471, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1258022785186768, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7461700439453125, + "num_tokens": 558826193.0, + "step": 21599 + }, + { + "epoch": 2.3720623764550846, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.489988088607788, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7486075162887573, + "num_tokens": 558850118.0, + "step": 21600 + }, + { + "epoch": 2.3721721941576983, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7150986194610596, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7300046682357788, + "num_tokens": 558870899.0, + "step": 21601 + }, + { + "epoch": 2.372282011860312, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.304042100906372, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7228716611862183, + "num_tokens": 558897874.0, + "step": 21602 + }, + { + "epoch": 2.3723918295629254, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.271209239959717, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7202867865562439, + "num_tokens": 558925023.0, + "step": 21603 + }, + { + "epoch": 2.372501647265539, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4196314811706543, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7346128821372986, + "num_tokens": 558950567.0, + "step": 21604 + }, + { + "epoch": 2.372611464968153, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4277563095092773, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7425166368484497, + "num_tokens": 558974871.0, + "step": 21605 + }, + { + "epoch": 2.3727212826707667, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.188831090927124, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7183075547218323, + "num_tokens": 559005912.0, + "step": 21606 + }, + { + "epoch": 2.3728311003733804, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.276918649673462, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.728848934173584, + "num_tokens": 559032318.0, + "step": 21607 + }, + { + "epoch": 2.3729409180759937, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.9019057750701904, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7493772506713867, + "num_tokens": 559050923.0, + "step": 21608 + }, + { + "epoch": 2.3730507357786075, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0361034870147705, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7507061958312988, + "num_tokens": 559081779.0, + "step": 21609 + }, + { + "epoch": 2.3731605534812212, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6839957237243652, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7190171480178833, + "num_tokens": 559101853.0, + "step": 21610 + }, + { + "epoch": 2.373270371183835, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.658971071243286, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7127091288566589, + "num_tokens": 559125924.0, + "step": 21611 + }, + { + "epoch": 2.3733801888864483, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.196744680404663, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7422658205032349, + "num_tokens": 559155653.0, + "step": 21612 + }, + { + "epoch": 2.373490006589062, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6740260124206543, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7319933176040649, + "num_tokens": 559179068.0, + "step": 21613 + }, + { + "epoch": 2.373599824291676, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3057162761688232, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7225189805030823, + "num_tokens": 559206875.0, + "step": 21614 + }, + { + "epoch": 2.3737096419942896, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2085092067718506, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7276977896690369, + "num_tokens": 559235853.0, + "step": 21615 + }, + { + "epoch": 2.3738194596969033, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.752851724624634, + "learning_rate": 1e-06, + "loss": 0.74, + "mean_token_accuracy": 0.7600823640823364, + "num_tokens": 559256191.0, + "step": 21616 + }, + { + "epoch": 2.3739292773995166, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6260061264038086, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7164203524589539, + "num_tokens": 559278183.0, + "step": 21617 + }, + { + "epoch": 2.3740390951021304, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.633312940597534, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7392421960830688, + "num_tokens": 559299964.0, + "step": 21618 + }, + { + "epoch": 2.374148912804744, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.360464334487915, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7197138071060181, + "num_tokens": 559324445.0, + "step": 21619 + }, + { + "epoch": 2.374258730507358, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.400742530822754, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7255642414093018, + "num_tokens": 559349323.0, + "step": 21620 + }, + { + "epoch": 2.3743685482099712, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2600326538085938, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.724113941192627, + "num_tokens": 559378716.0, + "step": 21621 + }, + { + "epoch": 2.374478365912585, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0166666507720947, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7079252004623413, + "num_tokens": 559412640.0, + "step": 21622 + }, + { + "epoch": 2.3745881836151987, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.809022903442383, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7458570003509521, + "num_tokens": 559431472.0, + "step": 21623 + }, + { + "epoch": 2.3746980013178125, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.567894697189331, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7438068389892578, + "num_tokens": 559453465.0, + "step": 21624 + }, + { + "epoch": 2.3748078190204263, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.485337972640991, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7164047956466675, + "num_tokens": 559479660.0, + "step": 21625 + }, + { + "epoch": 2.3749176367230396, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.130737781524658, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7429084777832031, + "num_tokens": 559509927.0, + "step": 21626 + }, + { + "epoch": 2.3750274544256533, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.391252279281616, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7263244986534119, + "num_tokens": 559535535.0, + "step": 21627 + }, + { + "epoch": 2.375137272128267, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.291518449783325, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.738836407661438, + "num_tokens": 559563278.0, + "step": 21628 + }, + { + "epoch": 2.375247089830881, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2989726066589355, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7292671799659729, + "num_tokens": 559588563.0, + "step": 21629 + }, + { + "epoch": 2.3753569075334946, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2655727863311768, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7101790904998779, + "num_tokens": 559619051.0, + "step": 21630 + }, + { + "epoch": 2.375466725236108, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.612053394317627, + "learning_rate": 1e-06, + "loss": 0.7552, + "mean_token_accuracy": 0.7557536959648132, + "num_tokens": 559639925.0, + "step": 21631 + }, + { + "epoch": 2.3755765429387217, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1745996475219727, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7507770657539368, + "num_tokens": 559671210.0, + "step": 21632 + }, + { + "epoch": 2.3756863606413354, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5848867893218994, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7289615273475647, + "num_tokens": 559694236.0, + "step": 21633 + }, + { + "epoch": 2.375796178343949, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7731773853302, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.721717119216919, + "num_tokens": 559714548.0, + "step": 21634 + }, + { + "epoch": 2.375905996046563, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.224931001663208, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7181971669197083, + "num_tokens": 559744762.0, + "step": 21635 + }, + { + "epoch": 2.3760158137491763, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.356027841567993, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7307363748550415, + "num_tokens": 559770635.0, + "step": 21636 + }, + { + "epoch": 2.37612563145179, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7584640979766846, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7319715023040771, + "num_tokens": 559791610.0, + "step": 21637 + }, + { + "epoch": 2.3762354491544038, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.589664936065674, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7292719483375549, + "num_tokens": 559814186.0, + "step": 21638 + }, + { + "epoch": 2.3763452668570175, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6022419929504395, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7218155860900879, + "num_tokens": 559835190.0, + "step": 21639 + }, + { + "epoch": 2.376455084559631, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.104259729385376, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7199544906616211, + "num_tokens": 559866682.0, + "step": 21640 + }, + { + "epoch": 2.3765649022622446, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1849045753479004, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7233718633651733, + "num_tokens": 559895962.0, + "step": 21641 + }, + { + "epoch": 2.3766747199648584, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.512769937515259, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7549772262573242, + "num_tokens": 559918730.0, + "step": 21642 + }, + { + "epoch": 2.376784537667472, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4597666263580322, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7271320223808289, + "num_tokens": 559942407.0, + "step": 21643 + }, + { + "epoch": 2.3768943553700854, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4273948669433594, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.71164470911026, + "num_tokens": 559969006.0, + "step": 21644 + }, + { + "epoch": 2.377004173072699, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.501544713973999, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7301663160324097, + "num_tokens": 559992947.0, + "step": 21645 + }, + { + "epoch": 2.377113990775313, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.343472480773926, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7069031596183777, + "num_tokens": 560020395.0, + "step": 21646 + }, + { + "epoch": 2.3772238084779267, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.228286027908325, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.736691415309906, + "num_tokens": 560048467.0, + "step": 21647 + }, + { + "epoch": 2.3773336261805404, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5870742797851562, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.733069658279419, + "num_tokens": 560070449.0, + "step": 21648 + }, + { + "epoch": 2.3774434438831538, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.048954963684082, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7061038017272949, + "num_tokens": 560102206.0, + "step": 21649 + }, + { + "epoch": 2.3775532615857675, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4396913051605225, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7286016941070557, + "num_tokens": 560124865.0, + "step": 21650 + }, + { + "epoch": 2.3776630792883813, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.423673391342163, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7417792677879333, + "num_tokens": 560146976.0, + "step": 21651 + }, + { + "epoch": 2.377772896990995, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6231658458709717, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7449017763137817, + "num_tokens": 560166722.0, + "step": 21652 + }, + { + "epoch": 2.377882714693609, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5895047187805176, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7363580465316772, + "num_tokens": 560189172.0, + "step": 21653 + }, + { + "epoch": 2.377992532396222, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4054739475250244, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7190030813217163, + "num_tokens": 560214066.0, + "step": 21654 + }, + { + "epoch": 2.378102350098836, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1155829429626465, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7365438342094421, + "num_tokens": 560246232.0, + "step": 21655 + }, + { + "epoch": 2.3782121678014496, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3651015758514404, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7342977523803711, + "num_tokens": 560272804.0, + "step": 21656 + }, + { + "epoch": 2.3783219855040634, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.490182876586914, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.733830988407135, + "num_tokens": 560296297.0, + "step": 21657 + }, + { + "epoch": 2.378431803206677, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3631653785705566, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7172964811325073, + "num_tokens": 560319923.0, + "step": 21658 + }, + { + "epoch": 2.3785416209092904, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4498825073242188, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7248460054397583, + "num_tokens": 560345057.0, + "step": 21659 + }, + { + "epoch": 2.378651438611904, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6964781284332275, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7307565212249756, + "num_tokens": 560365551.0, + "step": 21660 + }, + { + "epoch": 2.378761256314518, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.357133626937866, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.755394697189331, + "num_tokens": 560390724.0, + "step": 21661 + }, + { + "epoch": 2.3788710740171317, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.630593776702881, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7198571562767029, + "num_tokens": 560412951.0, + "step": 21662 + }, + { + "epoch": 2.3789808917197455, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.499022960662842, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.728535532951355, + "num_tokens": 560437061.0, + "step": 21663 + }, + { + "epoch": 2.379090709422359, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4678866863250732, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.7569209337234497, + "num_tokens": 560459323.0, + "step": 21664 + }, + { + "epoch": 2.3792005271249725, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7133326530456543, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7216099500656128, + "num_tokens": 560480023.0, + "step": 21665 + }, + { + "epoch": 2.3793103448275863, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.51346492767334, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7189313173294067, + "num_tokens": 560507057.0, + "step": 21666 + }, + { + "epoch": 2.3794201625302, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.415044069290161, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.719396710395813, + "num_tokens": 560531806.0, + "step": 21667 + }, + { + "epoch": 2.3795299802328134, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0163049697875977, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7341949939727783, + "num_tokens": 560562967.0, + "step": 21668 + }, + { + "epoch": 2.379639797935427, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.452871322631836, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.745008647441864, + "num_tokens": 560587734.0, + "step": 21669 + }, + { + "epoch": 2.379749615638041, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.532719612121582, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7227888703346252, + "num_tokens": 560612771.0, + "step": 21670 + }, + { + "epoch": 2.3798594333406546, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.811474323272705, + "learning_rate": 1e-06, + "loss": 0.6937, + "mean_token_accuracy": 0.7838517427444458, + "num_tokens": 560631954.0, + "step": 21671 + }, + { + "epoch": 2.379969251043268, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1032445430755615, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.6989171504974365, + "num_tokens": 560664443.0, + "step": 21672 + }, + { + "epoch": 2.3800790687458817, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.425260305404663, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7296581268310547, + "num_tokens": 560689391.0, + "step": 21673 + }, + { + "epoch": 2.3801888864484955, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.243957281112671, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7214407920837402, + "num_tokens": 560717244.0, + "step": 21674 + }, + { + "epoch": 2.3802987041511092, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.493288278579712, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.745369017124176, + "num_tokens": 560738910.0, + "step": 21675 + }, + { + "epoch": 2.380408521853723, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.118155002593994, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7269729971885681, + "num_tokens": 560769723.0, + "step": 21676 + }, + { + "epoch": 2.3805183395563363, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 1.8793714046478271, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7190647125244141, + "num_tokens": 560806610.0, + "step": 21677 + }, + { + "epoch": 2.38062815725895, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.542276382446289, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7276257276535034, + "num_tokens": 560829502.0, + "step": 21678 + }, + { + "epoch": 2.380737974961564, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.287813186645508, + "learning_rate": 1e-06, + "loss": 0.7661, + "mean_token_accuracy": 0.7542627453804016, + "num_tokens": 560855250.0, + "step": 21679 + }, + { + "epoch": 2.3808477926641776, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.640048027038574, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7482160329818726, + "num_tokens": 560876357.0, + "step": 21680 + }, + { + "epoch": 2.3809576103667913, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4800968170166016, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7136515378952026, + "num_tokens": 560902717.0, + "step": 21681 + }, + { + "epoch": 2.3810674280694046, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.20009446144104, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7406898736953735, + "num_tokens": 560930540.0, + "step": 21682 + }, + { + "epoch": 2.3811772457720184, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4662671089172363, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7424512505531311, + "num_tokens": 560954786.0, + "step": 21683 + }, + { + "epoch": 2.381287063474632, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2849788665771484, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7315330505371094, + "num_tokens": 560981734.0, + "step": 21684 + }, + { + "epoch": 2.381396881177246, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.426754951477051, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7384594678878784, + "num_tokens": 561004489.0, + "step": 21685 + }, + { + "epoch": 2.3815066988798597, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.687039613723755, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.746268093585968, + "num_tokens": 561024652.0, + "step": 21686 + }, + { + "epoch": 2.381616516582473, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4210424423217773, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7371012568473816, + "num_tokens": 561049401.0, + "step": 21687 + }, + { + "epoch": 2.3817263342850867, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.630567789077759, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7446396350860596, + "num_tokens": 561071127.0, + "step": 21688 + }, + { + "epoch": 2.3818361519877005, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4111809730529785, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7112772464752197, + "num_tokens": 561095712.0, + "step": 21689 + }, + { + "epoch": 2.3819459696903142, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.567101240158081, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.73964524269104, + "num_tokens": 561118285.0, + "step": 21690 + }, + { + "epoch": 2.3820557873929276, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2144670486450195, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7128076553344727, + "num_tokens": 561147901.0, + "step": 21691 + }, + { + "epoch": 2.3821656050955413, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.198735237121582, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7240915298461914, + "num_tokens": 561176893.0, + "step": 21692 + }, + { + "epoch": 2.382275422798155, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1980698108673096, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6902046203613281, + "num_tokens": 561207332.0, + "step": 21693 + }, + { + "epoch": 2.382385240500769, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6470816135406494, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7393378019332886, + "num_tokens": 561228511.0, + "step": 21694 + }, + { + "epoch": 2.382495058203382, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2271111011505127, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7235047817230225, + "num_tokens": 561256955.0, + "step": 21695 + }, + { + "epoch": 2.382604875905996, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5496785640716553, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7355453968048096, + "num_tokens": 561279513.0, + "step": 21696 + }, + { + "epoch": 2.3827146936086097, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2351772785186768, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7216310501098633, + "num_tokens": 561307164.0, + "step": 21697 + }, + { + "epoch": 2.3828245113112234, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1340560913085938, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6934736967086792, + "num_tokens": 561337569.0, + "step": 21698 + }, + { + "epoch": 2.382934329013837, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.226398468017578, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7198558449745178, + "num_tokens": 561367097.0, + "step": 21699 + }, + { + "epoch": 2.3830441467164505, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.278543710708618, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7233881950378418, + "num_tokens": 561394499.0, + "step": 21700 + }, + { + "epoch": 2.3831539644190642, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3765971660614014, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7297557592391968, + "num_tokens": 561422943.0, + "step": 21701 + }, + { + "epoch": 2.383263782121678, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.601529836654663, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7417879104614258, + "num_tokens": 561446362.0, + "step": 21702 + }, + { + "epoch": 2.3833735998242918, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4727566242218018, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7186764478683472, + "num_tokens": 561471673.0, + "step": 21703 + }, + { + "epoch": 2.3834834175269055, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.9196979999542236, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7349002361297607, + "num_tokens": 561491792.0, + "step": 21704 + }, + { + "epoch": 2.383593235229519, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.145519256591797, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7076764106750488, + "num_tokens": 561523490.0, + "step": 21705 + }, + { + "epoch": 2.3837030529321326, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5259087085723877, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7298824787139893, + "num_tokens": 561546471.0, + "step": 21706 + }, + { + "epoch": 2.3838128706347463, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.564539670944214, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.75138258934021, + "num_tokens": 561568689.0, + "step": 21707 + }, + { + "epoch": 2.38392268833736, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3643174171447754, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7242493629455566, + "num_tokens": 561595060.0, + "step": 21708 + }, + { + "epoch": 2.384032506039974, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.39461350440979, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7460200786590576, + "num_tokens": 561619795.0, + "step": 21709 + }, + { + "epoch": 2.384142323742587, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 3.0705504417419434, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7455234527587891, + "num_tokens": 561636235.0, + "step": 21710 + }, + { + "epoch": 2.384252141445201, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.100024700164795, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7495126724243164, + "num_tokens": 561665324.0, + "step": 21711 + }, + { + "epoch": 2.3843619591478147, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5306107997894287, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7110919952392578, + "num_tokens": 561686989.0, + "step": 21712 + }, + { + "epoch": 2.3844717768504284, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4145233631134033, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7192679643630981, + "num_tokens": 561713475.0, + "step": 21713 + }, + { + "epoch": 2.384581594553042, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2182798385620117, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.730177640914917, + "num_tokens": 561740807.0, + "step": 21714 + }, + { + "epoch": 2.3846914122556555, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4854252338409424, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7445379495620728, + "num_tokens": 561764139.0, + "step": 21715 + }, + { + "epoch": 2.3848012299582693, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2100658416748047, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7476047277450562, + "num_tokens": 561791526.0, + "step": 21716 + }, + { + "epoch": 2.384911047660883, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6799323558807373, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7332677841186523, + "num_tokens": 561813069.0, + "step": 21717 + }, + { + "epoch": 2.385020865363497, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2979414463043213, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.740574300289154, + "num_tokens": 561838720.0, + "step": 21718 + }, + { + "epoch": 2.38513068306611, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.450042724609375, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7393219470977783, + "num_tokens": 561864126.0, + "step": 21719 + }, + { + "epoch": 2.385240500768724, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4182357788085938, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7326624393463135, + "num_tokens": 561889744.0, + "step": 21720 + }, + { + "epoch": 2.3853503184713376, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.324730396270752, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7298603057861328, + "num_tokens": 561914889.0, + "step": 21721 + }, + { + "epoch": 2.3854601361739514, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4396607875823975, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.73045814037323, + "num_tokens": 561939297.0, + "step": 21722 + }, + { + "epoch": 2.3855699538765647, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.360827922821045, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7310848832130432, + "num_tokens": 561965447.0, + "step": 21723 + }, + { + "epoch": 2.3856797715791784, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.279524087905884, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7006233930587769, + "num_tokens": 561994561.0, + "step": 21724 + }, + { + "epoch": 2.385789589281792, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 1.995108962059021, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7446221113204956, + "num_tokens": 562028541.0, + "step": 21725 + }, + { + "epoch": 2.385899406984406, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.1722946166992188, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.733107328414917, + "num_tokens": 562059217.0, + "step": 21726 + }, + { + "epoch": 2.3860092246870197, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.367220640182495, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.739306628704071, + "num_tokens": 562085241.0, + "step": 21727 + }, + { + "epoch": 2.386119042389633, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.49530291557312, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.746590256690979, + "num_tokens": 562109433.0, + "step": 21728 + }, + { + "epoch": 2.3862288600922468, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.7693092823028564, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7354649305343628, + "num_tokens": 562130464.0, + "step": 21729 + }, + { + "epoch": 2.3863386777948605, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5128066539764404, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7428877353668213, + "num_tokens": 562154334.0, + "step": 21730 + }, + { + "epoch": 2.3864484954974743, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3430237770080566, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.749640166759491, + "num_tokens": 562179353.0, + "step": 21731 + }, + { + "epoch": 2.386558313200088, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2454938888549805, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7324837446212769, + "num_tokens": 562209174.0, + "step": 21732 + }, + { + "epoch": 2.3866681309027014, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2883780002593994, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7441447973251343, + "num_tokens": 562238300.0, + "step": 21733 + }, + { + "epoch": 2.386777948605315, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2762045860290527, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7182660102844238, + "num_tokens": 562267018.0, + "step": 21734 + }, + { + "epoch": 2.386887766307929, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0979580879211426, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7563124895095825, + "num_tokens": 562296651.0, + "step": 21735 + }, + { + "epoch": 2.3869975840105426, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4920523166656494, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7203084230422974, + "num_tokens": 562322156.0, + "step": 21736 + }, + { + "epoch": 2.3871074017131564, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4244132041931152, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7368733882904053, + "num_tokens": 562344786.0, + "step": 21737 + }, + { + "epoch": 2.3872172194157697, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.48667573928833, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7317157983779907, + "num_tokens": 562368680.0, + "step": 21738 + }, + { + "epoch": 2.3873270371183835, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2458200454711914, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7096606492996216, + "num_tokens": 562396086.0, + "step": 21739 + }, + { + "epoch": 2.387436854820997, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.283700704574585, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7303693294525146, + "num_tokens": 562424322.0, + "step": 21740 + }, + { + "epoch": 2.387546672523611, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.419459104537964, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7066411972045898, + "num_tokens": 562449818.0, + "step": 21741 + }, + { + "epoch": 2.3876564902262243, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2099053859710693, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7285977005958557, + "num_tokens": 562479687.0, + "step": 21742 + }, + { + "epoch": 2.387766307928838, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.616940975189209, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7526777982711792, + "num_tokens": 562500708.0, + "step": 21743 + }, + { + "epoch": 2.387876125631452, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.566209316253662, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7166345119476318, + "num_tokens": 562524041.0, + "step": 21744 + }, + { + "epoch": 2.3879859433340656, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4440724849700928, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7260761260986328, + "num_tokens": 562551157.0, + "step": 21745 + }, + { + "epoch": 2.388095761036679, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.050372838973999, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7384529113769531, + "num_tokens": 562583952.0, + "step": 21746 + }, + { + "epoch": 2.3882055787392926, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.587110996246338, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.740415632724762, + "num_tokens": 562606927.0, + "step": 21747 + }, + { + "epoch": 2.3883153964419064, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.739442825317383, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7289769649505615, + "num_tokens": 562627104.0, + "step": 21748 + }, + { + "epoch": 2.38842521414452, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.622619390487671, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7252243757247925, + "num_tokens": 562649867.0, + "step": 21749 + }, + { + "epoch": 2.388535031847134, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.641122341156006, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7140775918960571, + "num_tokens": 562672739.0, + "step": 21750 + }, + { + "epoch": 2.388644849549747, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.415022373199463, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7137220501899719, + "num_tokens": 562699666.0, + "step": 21751 + }, + { + "epoch": 2.388754667252361, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6588001251220703, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.747104823589325, + "num_tokens": 562719402.0, + "step": 21752 + }, + { + "epoch": 2.3888644849549747, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.574209213256836, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7236313819885254, + "num_tokens": 562743460.0, + "step": 21753 + }, + { + "epoch": 2.3889743026575885, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5405561923980713, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7293833494186401, + "num_tokens": 562766855.0, + "step": 21754 + }, + { + "epoch": 2.3890841203602022, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.548752546310425, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7371655702590942, + "num_tokens": 562789578.0, + "step": 21755 + }, + { + "epoch": 2.3891939380628155, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.262582778930664, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7397385239601135, + "num_tokens": 562816256.0, + "step": 21756 + }, + { + "epoch": 2.3893037557654293, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4556362628936768, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6954455375671387, + "num_tokens": 562842188.0, + "step": 21757 + }, + { + "epoch": 2.389413573468043, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2822325229644775, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7339096069335938, + "num_tokens": 562870493.0, + "step": 21758 + }, + { + "epoch": 2.389523391170657, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3641469478607178, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7323474287986755, + "num_tokens": 562896126.0, + "step": 21759 + }, + { + "epoch": 2.3896332088732706, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6924710273742676, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7462975978851318, + "num_tokens": 562916736.0, + "step": 21760 + }, + { + "epoch": 2.389743026575884, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.088428258895874, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7189136743545532, + "num_tokens": 562948370.0, + "step": 21761 + }, + { + "epoch": 2.3898528442784976, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.248889446258545, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.74539715051651, + "num_tokens": 562974389.0, + "step": 21762 + }, + { + "epoch": 2.3899626619811114, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2302191257476807, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7307133674621582, + "num_tokens": 563001419.0, + "step": 21763 + }, + { + "epoch": 2.390072479683725, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.4198005199432373, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7104206085205078, + "num_tokens": 563028210.0, + "step": 21764 + }, + { + "epoch": 2.390182297386339, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.324028491973877, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7413250207901001, + "num_tokens": 563052099.0, + "step": 21765 + }, + { + "epoch": 2.3902921150889522, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.6571218967437744, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7189337015151978, + "num_tokens": 563074501.0, + "step": 21766 + }, + { + "epoch": 2.390401932791566, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.130913496017456, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7279065847396851, + "num_tokens": 563105438.0, + "step": 21767 + }, + { + "epoch": 2.3905117504941797, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.558220624923706, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7321285009384155, + "num_tokens": 563127829.0, + "step": 21768 + }, + { + "epoch": 2.3906215681967935, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.5405614376068115, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7359977960586548, + "num_tokens": 563150649.0, + "step": 21769 + }, + { + "epoch": 2.390731385899407, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.388826608657837, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7217617034912109, + "num_tokens": 563174548.0, + "step": 21770 + }, + { + "epoch": 2.3908412036020206, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.0040178298950195, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7107496857643127, + "num_tokens": 563211815.0, + "step": 21771 + }, + { + "epoch": 2.3909510213046343, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.2113001346588135, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7154018878936768, + "num_tokens": 563240838.0, + "step": 21772 + }, + { + "epoch": 2.391060839007248, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.3669328689575195, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7323247194290161, + "num_tokens": 563266701.0, + "step": 21773 + }, + { + "epoch": 2.3911706567098614, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.456482410430908, + "learning_rate": 1e-06, + "loss": 0.7717, + "mean_token_accuracy": 0.7543628215789795, + "num_tokens": 563290876.0, + "step": 21774 + }, + { + "epoch": 2.391280474412475, + "ewc_loss": 0.0001888275146484375, + "grad_norm": 2.825922727584839, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7488906979560852, + "num_tokens": 563310841.0, + "step": 21775 + }, + { + "epoch": 2.391390292115089, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5041286945343018, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7482870817184448, + "num_tokens": 563335458.0, + "step": 21776 + }, + { + "epoch": 2.3915001098177027, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2364346981048584, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.75395667552948, + "num_tokens": 563362466.0, + "step": 21777 + }, + { + "epoch": 2.3916099275203164, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.1756882667541504, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.73279869556427, + "num_tokens": 563391871.0, + "step": 21778 + }, + { + "epoch": 2.3917197452229297, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4098503589630127, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7238199710845947, + "num_tokens": 563417280.0, + "step": 21779 + }, + { + "epoch": 2.3918295629255435, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.210089683532715, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7128691673278809, + "num_tokens": 563446860.0, + "step": 21780 + }, + { + "epoch": 2.3919393806281573, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.211932420730591, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7280048131942749, + "num_tokens": 563476694.0, + "step": 21781 + }, + { + "epoch": 2.392049198330771, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.408134937286377, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7463435530662537, + "num_tokens": 563501967.0, + "step": 21782 + }, + { + "epoch": 2.3921590160333848, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.20178484916687, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7217875719070435, + "num_tokens": 563533006.0, + "step": 21783 + }, + { + "epoch": 2.392268833735998, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4528987407684326, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7539628744125366, + "num_tokens": 563556085.0, + "step": 21784 + }, + { + "epoch": 2.392378651438612, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.278015375137329, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7344011068344116, + "num_tokens": 563583818.0, + "step": 21785 + }, + { + "epoch": 2.3924884691412256, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2600839138031006, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7209347486495972, + "num_tokens": 563611700.0, + "step": 21786 + }, + { + "epoch": 2.3925982868438394, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.6439363956451416, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7272264957427979, + "num_tokens": 563635091.0, + "step": 21787 + }, + { + "epoch": 2.392708104546453, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.275477170944214, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7319707870483398, + "num_tokens": 563664243.0, + "step": 21788 + }, + { + "epoch": 2.3928179222490664, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.6308202743530273, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7403779029846191, + "num_tokens": 563687430.0, + "step": 21789 + }, + { + "epoch": 2.39292773995168, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2606310844421387, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7254819869995117, + "num_tokens": 563716757.0, + "step": 21790 + }, + { + "epoch": 2.393037557654294, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5083062648773193, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.745391309261322, + "num_tokens": 563739681.0, + "step": 21791 + }, + { + "epoch": 2.3931473753569077, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4229679107666016, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7413467764854431, + "num_tokens": 563763013.0, + "step": 21792 + }, + { + "epoch": 2.393257193059521, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.372637987136841, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7089897394180298, + "num_tokens": 563789203.0, + "step": 21793 + }, + { + "epoch": 2.3933670107621348, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4877398014068604, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7361292839050293, + "num_tokens": 563813546.0, + "step": 21794 + }, + { + "epoch": 2.3934768284647485, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4599146842956543, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7200429439544678, + "num_tokens": 563840301.0, + "step": 21795 + }, + { + "epoch": 2.3935866461673623, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.370537757873535, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7218647599220276, + "num_tokens": 563867125.0, + "step": 21796 + }, + { + "epoch": 2.393696463869976, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.616478681564331, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7323852777481079, + "num_tokens": 563888995.0, + "step": 21797 + }, + { + "epoch": 2.3938062815725893, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.285386800765991, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7366552352905273, + "num_tokens": 563914370.0, + "step": 21798 + }, + { + "epoch": 2.393916099275203, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.325547456741333, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7411292195320129, + "num_tokens": 563941515.0, + "step": 21799 + }, + { + "epoch": 2.394025916977817, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2708988189697266, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.748194694519043, + "num_tokens": 563967176.0, + "step": 21800 + }, + { + "epoch": 2.3941357346804306, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5648725032806396, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7460170984268188, + "num_tokens": 563989351.0, + "step": 21801 + }, + { + "epoch": 2.394245552383044, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2766411304473877, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7373108267784119, + "num_tokens": 564015982.0, + "step": 21802 + }, + { + "epoch": 2.3943553700856577, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.18047833442688, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7389257550239563, + "num_tokens": 564043171.0, + "step": 21803 + }, + { + "epoch": 2.3944651877882714, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.18452525138855, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.718966007232666, + "num_tokens": 564073133.0, + "step": 21804 + }, + { + "epoch": 2.394575005490885, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.307749032974243, + "learning_rate": 1e-06, + "loss": 0.777, + "mean_token_accuracy": 0.761350691318512, + "num_tokens": 564099305.0, + "step": 21805 + }, + { + "epoch": 2.394684823193499, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.280061721801758, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7455266714096069, + "num_tokens": 564124842.0, + "step": 21806 + }, + { + "epoch": 2.3947946408961123, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.1830968856811523, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7092180848121643, + "num_tokens": 564158542.0, + "step": 21807 + }, + { + "epoch": 2.394904458598726, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.327650308609009, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7284640669822693, + "num_tokens": 564185746.0, + "step": 21808 + }, + { + "epoch": 2.39501427630134, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4637157917022705, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7370485067367554, + "num_tokens": 564208325.0, + "step": 21809 + }, + { + "epoch": 2.3951240940039535, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2000510692596436, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7280359268188477, + "num_tokens": 564237384.0, + "step": 21810 + }, + { + "epoch": 2.3952339117065673, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4799413681030273, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7180798649787903, + "num_tokens": 564263098.0, + "step": 21811 + }, + { + "epoch": 2.3953437294091806, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.438081979751587, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7131472229957581, + "num_tokens": 564287499.0, + "step": 21812 + }, + { + "epoch": 2.3954535471117944, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.7442855834960938, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7371739149093628, + "num_tokens": 564307598.0, + "step": 21813 + }, + { + "epoch": 2.395563364814408, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2298264503479004, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7104693651199341, + "num_tokens": 564338375.0, + "step": 21814 + }, + { + "epoch": 2.395673182517022, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.440753936767578, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7350726127624512, + "num_tokens": 564361749.0, + "step": 21815 + }, + { + "epoch": 2.3957830002196356, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3714232444763184, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7257437109947205, + "num_tokens": 564387269.0, + "step": 21816 + }, + { + "epoch": 2.395892817922249, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.398318290710449, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7185046076774597, + "num_tokens": 564411993.0, + "step": 21817 + }, + { + "epoch": 2.3960026356248627, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.289473533630371, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7393861413002014, + "num_tokens": 564437303.0, + "step": 21818 + }, + { + "epoch": 2.3961124533274765, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.1785356998443604, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7333524823188782, + "num_tokens": 564467197.0, + "step": 21819 + }, + { + "epoch": 2.3962222710300902, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.05007266998291, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7220585346221924, + "num_tokens": 564501039.0, + "step": 21820 + }, + { + "epoch": 2.3963320887327035, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2081165313720703, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.724250853061676, + "num_tokens": 564529656.0, + "step": 21821 + }, + { + "epoch": 2.3964419064353173, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.0746803283691406, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7335571646690369, + "num_tokens": 564560411.0, + "step": 21822 + }, + { + "epoch": 2.396551724137931, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.152148723602295, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.734134316444397, + "num_tokens": 564589332.0, + "step": 21823 + }, + { + "epoch": 2.396661541840545, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2198264598846436, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7410176992416382, + "num_tokens": 564617369.0, + "step": 21824 + }, + { + "epoch": 2.396771359543158, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5926995277404785, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.724503219127655, + "num_tokens": 564640388.0, + "step": 21825 + }, + { + "epoch": 2.396881177245772, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.506166934967041, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7127103805541992, + "num_tokens": 564664886.0, + "step": 21826 + }, + { + "epoch": 2.3969909949483856, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3376364707946777, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7180395126342773, + "num_tokens": 564693525.0, + "step": 21827 + }, + { + "epoch": 2.3971008126509994, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3767123222351074, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7226830720901489, + "num_tokens": 564721609.0, + "step": 21828 + }, + { + "epoch": 2.397210630353613, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.371821165084839, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.730711042881012, + "num_tokens": 564747590.0, + "step": 21829 + }, + { + "epoch": 2.3973204480562265, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.0975167751312256, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7023729681968689, + "num_tokens": 564778965.0, + "step": 21830 + }, + { + "epoch": 2.39743026575884, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5163114070892334, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7404628396034241, + "num_tokens": 564800943.0, + "step": 21831 + }, + { + "epoch": 2.397540083461454, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4121880531311035, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7084323167800903, + "num_tokens": 564826490.0, + "step": 21832 + }, + { + "epoch": 2.3976499011640677, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.422614336013794, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.717877209186554, + "num_tokens": 564852336.0, + "step": 21833 + }, + { + "epoch": 2.3977597188666815, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3807406425476074, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7401455044746399, + "num_tokens": 564876595.0, + "step": 21834 + }, + { + "epoch": 2.397869536569295, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3070578575134277, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7279828190803528, + "num_tokens": 564906207.0, + "step": 21835 + }, + { + "epoch": 2.3979793542719086, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.7959017753601074, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.742159366607666, + "num_tokens": 564925607.0, + "step": 21836 + }, + { + "epoch": 2.3980891719745223, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4653964042663574, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7452787160873413, + "num_tokens": 564948981.0, + "step": 21837 + }, + { + "epoch": 2.398198989677136, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.100698947906494, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7180531620979309, + "num_tokens": 564980853.0, + "step": 21838 + }, + { + "epoch": 2.39830880737975, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.19895076751709, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7105754017829895, + "num_tokens": 565010521.0, + "step": 21839 + }, + { + "epoch": 2.398418625082363, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.7272989749908447, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7539976835250854, + "num_tokens": 565030037.0, + "step": 21840 + }, + { + "epoch": 2.398528442784977, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4692318439483643, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.724092960357666, + "num_tokens": 565056268.0, + "step": 21841 + }, + { + "epoch": 2.3986382604875907, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3524081707000732, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7160800695419312, + "num_tokens": 565082378.0, + "step": 21842 + }, + { + "epoch": 2.3987480781902044, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.912959337234497, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7462075352668762, + "num_tokens": 565100237.0, + "step": 21843 + }, + { + "epoch": 2.398857895892818, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3550360202789307, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7169749736785889, + "num_tokens": 565125361.0, + "step": 21844 + }, + { + "epoch": 2.3989677135954315, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.6193103790283203, + "learning_rate": 1e-06, + "loss": 0.7397, + "mean_token_accuracy": 0.7579860687255859, + "num_tokens": 565146362.0, + "step": 21845 + }, + { + "epoch": 2.3990775312980452, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5865907669067383, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.750151515007019, + "num_tokens": 565168609.0, + "step": 21846 + }, + { + "epoch": 2.399187349000659, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.503798723220825, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7196511626243591, + "num_tokens": 565192456.0, + "step": 21847 + }, + { + "epoch": 2.3992971667032728, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2723517417907715, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7163534164428711, + "num_tokens": 565221728.0, + "step": 21848 + }, + { + "epoch": 2.399406984405886, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.1730170249938965, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7375050187110901, + "num_tokens": 565253334.0, + "step": 21849 + }, + { + "epoch": 2.3995168021085, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3454768657684326, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7170061469078064, + "num_tokens": 565280534.0, + "step": 21850 + }, + { + "epoch": 2.3996266198111136, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3216421604156494, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7297399044036865, + "num_tokens": 565307008.0, + "step": 21851 + }, + { + "epoch": 2.3997364375137273, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4867336750030518, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7403287887573242, + "num_tokens": 565329287.0, + "step": 21852 + }, + { + "epoch": 2.3998462552163407, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.195446014404297, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7430741786956787, + "num_tokens": 565358958.0, + "step": 21853 + }, + { + "epoch": 2.3999560729189544, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3183627128601074, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7462811470031738, + "num_tokens": 565388515.0, + "step": 21854 + }, + { + "epoch": 2.400065890621568, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.6047394275665283, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7316594123840332, + "num_tokens": 565411671.0, + "step": 21855 + }, + { + "epoch": 2.400175708324182, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4896416664123535, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.74379563331604, + "num_tokens": 565434986.0, + "step": 21856 + }, + { + "epoch": 2.4002855260267957, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3539609909057617, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7247534394264221, + "num_tokens": 565460261.0, + "step": 21857 + }, + { + "epoch": 2.400395343729409, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2385165691375732, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7182368636131287, + "num_tokens": 565489292.0, + "step": 21858 + }, + { + "epoch": 2.4005051614320227, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2971911430358887, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7569829225540161, + "num_tokens": 565512923.0, + "step": 21859 + }, + { + "epoch": 2.4006149791346365, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.292250394821167, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.6977986693382263, + "num_tokens": 565542048.0, + "step": 21860 + }, + { + "epoch": 2.4007247968372503, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.223710775375366, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7147796154022217, + "num_tokens": 565571858.0, + "step": 21861 + }, + { + "epoch": 2.400834614539864, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3820085525512695, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.719562828540802, + "num_tokens": 565597292.0, + "step": 21862 + }, + { + "epoch": 2.4009444322424773, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2982866764068604, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7409747242927551, + "num_tokens": 565622850.0, + "step": 21863 + }, + { + "epoch": 2.401054249945091, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5369858741760254, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7231101989746094, + "num_tokens": 565645293.0, + "step": 21864 + }, + { + "epoch": 2.401164067647705, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3478119373321533, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7697604298591614, + "num_tokens": 565673669.0, + "step": 21865 + }, + { + "epoch": 2.4012738853503186, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.167546272277832, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7090385556221008, + "num_tokens": 565703567.0, + "step": 21866 + }, + { + "epoch": 2.4013837030529324, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.341588020324707, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7167773842811584, + "num_tokens": 565728942.0, + "step": 21867 + }, + { + "epoch": 2.4014935207555457, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.0864322185516357, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7291545867919922, + "num_tokens": 565760420.0, + "step": 21868 + }, + { + "epoch": 2.4016033384581594, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.6728076934814453, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7429351806640625, + "num_tokens": 565782604.0, + "step": 21869 + }, + { + "epoch": 2.401713156160773, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.696922540664673, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7202833890914917, + "num_tokens": 565803210.0, + "step": 21870 + }, + { + "epoch": 2.401822973863387, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2311744689941406, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7284002304077148, + "num_tokens": 565830695.0, + "step": 21871 + }, + { + "epoch": 2.4019327915660003, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.410444974899292, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7303342223167419, + "num_tokens": 565855945.0, + "step": 21872 + }, + { + "epoch": 2.402042609268614, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.7586889266967773, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7420790195465088, + "num_tokens": 565875323.0, + "step": 21873 + }, + { + "epoch": 2.4021524269712278, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3179872035980225, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7048115730285645, + "num_tokens": 565904005.0, + "step": 21874 + }, + { + "epoch": 2.4022622446738415, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.1566476821899414, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7176313400268555, + "num_tokens": 565933325.0, + "step": 21875 + }, + { + "epoch": 2.402372062376455, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.820399284362793, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7302058935165405, + "num_tokens": 565952176.0, + "step": 21876 + }, + { + "epoch": 2.4024818800790686, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.493586540222168, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7444523572921753, + "num_tokens": 565975579.0, + "step": 21877 + }, + { + "epoch": 2.4025916977816824, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.211019515991211, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.715474545955658, + "num_tokens": 566003823.0, + "step": 21878 + }, + { + "epoch": 2.402701515484296, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3392887115478516, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7373674511909485, + "num_tokens": 566031231.0, + "step": 21879 + }, + { + "epoch": 2.40281133318691, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5984835624694824, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7357973456382751, + "num_tokens": 566053359.0, + "step": 21880 + }, + { + "epoch": 2.402921150889523, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3263399600982666, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7362289428710938, + "num_tokens": 566081184.0, + "step": 21881 + }, + { + "epoch": 2.403030968592137, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4322094917297363, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7211652994155884, + "num_tokens": 566107003.0, + "step": 21882 + }, + { + "epoch": 2.4031407862947507, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3856654167175293, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.719489574432373, + "num_tokens": 566131996.0, + "step": 21883 + }, + { + "epoch": 2.4032506039973645, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.311532974243164, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7410027384757996, + "num_tokens": 566156993.0, + "step": 21884 + }, + { + "epoch": 2.403360421699978, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3938403129577637, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7153390645980835, + "num_tokens": 566181627.0, + "step": 21885 + }, + { + "epoch": 2.4034702394025915, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.504688024520874, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7277363538742065, + "num_tokens": 566207643.0, + "step": 21886 + }, + { + "epoch": 2.4035800571052053, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.412869930267334, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7269430756568909, + "num_tokens": 566232623.0, + "step": 21887 + }, + { + "epoch": 2.403689874807819, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.391146421432495, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7224392294883728, + "num_tokens": 566259132.0, + "step": 21888 + }, + { + "epoch": 2.403799692510433, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5104947090148926, + "learning_rate": 1e-06, + "loss": 0.7855, + "mean_token_accuracy": 0.7488954663276672, + "num_tokens": 566283076.0, + "step": 21889 + }, + { + "epoch": 2.4039095102130466, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.4432456493377686, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7412117719650269, + "num_tokens": 566307400.0, + "step": 21890 + }, + { + "epoch": 2.40401932791566, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.88234281539917, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7374825477600098, + "num_tokens": 566326041.0, + "step": 21891 + }, + { + "epoch": 2.4041291456182736, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2576026916503906, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7309462428092957, + "num_tokens": 566355322.0, + "step": 21892 + }, + { + "epoch": 2.4042389633208874, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.297363758087158, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.731032133102417, + "num_tokens": 566381855.0, + "step": 21893 + }, + { + "epoch": 2.404348781023501, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.3957314491271973, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.746660053730011, + "num_tokens": 566405157.0, + "step": 21894 + }, + { + "epoch": 2.404458598726115, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.2272160053253174, + "learning_rate": 1e-06, + "loss": 0.773, + "mean_token_accuracy": 0.7536176443099976, + "num_tokens": 566432532.0, + "step": 21895 + }, + { + "epoch": 2.404568416428728, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.868943452835083, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7444730997085571, + "num_tokens": 566452057.0, + "step": 21896 + }, + { + "epoch": 2.404678234131342, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.7054734230041504, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7429002523422241, + "num_tokens": 566472312.0, + "step": 21897 + }, + { + "epoch": 2.4047880518339557, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5341732501983643, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7385606169700623, + "num_tokens": 566498063.0, + "step": 21898 + }, + { + "epoch": 2.4048978695365695, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.5379319190979004, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7334135174751282, + "num_tokens": 566520570.0, + "step": 21899 + }, + { + "epoch": 2.405007687239183, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.164249897003174, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7276854515075684, + "num_tokens": 566547892.0, + "step": 21900 + }, + { + "epoch": 2.4051175049417965, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.524291515350342, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7403183579444885, + "num_tokens": 566569457.0, + "step": 21901 + }, + { + "epoch": 2.4052273226444103, + "ewc_loss": 0.00018978118896484375, + "grad_norm": 2.100367307662964, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7230725884437561, + "num_tokens": 566601355.0, + "step": 21902 + }, + { + "epoch": 2.405337140347024, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.544328212738037, + "learning_rate": 1e-06, + "loss": 0.7887, + "mean_token_accuracy": 0.7441402673721313, + "num_tokens": 566622944.0, + "step": 21903 + }, + { + "epoch": 2.4054469580496374, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.331540107727051, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7165807485580444, + "num_tokens": 566649726.0, + "step": 21904 + }, + { + "epoch": 2.405556775752251, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4252073764801025, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7141245007514954, + "num_tokens": 566676527.0, + "step": 21905 + }, + { + "epoch": 2.405666593454865, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2422266006469727, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7344117164611816, + "num_tokens": 566703545.0, + "step": 21906 + }, + { + "epoch": 2.4057764111574786, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2847936153411865, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7491722106933594, + "num_tokens": 566730234.0, + "step": 21907 + }, + { + "epoch": 2.4058862288600924, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.265249252319336, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7307612299919128, + "num_tokens": 566755985.0, + "step": 21908 + }, + { + "epoch": 2.4059960465627057, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.419969081878662, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7271744608879089, + "num_tokens": 566780924.0, + "step": 21909 + }, + { + "epoch": 2.4061058642653195, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3891189098358154, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7422735691070557, + "num_tokens": 566805480.0, + "step": 21910 + }, + { + "epoch": 2.4062156819679332, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4911205768585205, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6944512128829956, + "num_tokens": 566832750.0, + "step": 21911 + }, + { + "epoch": 2.406325499670547, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4915051460266113, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7384223937988281, + "num_tokens": 566855160.0, + "step": 21912 + }, + { + "epoch": 2.4064353173731607, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.571808099746704, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.7399557828903198, + "num_tokens": 566876672.0, + "step": 21913 + }, + { + "epoch": 2.406545135075774, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2567968368530273, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7143071293830872, + "num_tokens": 566906330.0, + "step": 21914 + }, + { + "epoch": 2.406654952778388, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.481468915939331, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7176320552825928, + "num_tokens": 566930922.0, + "step": 21915 + }, + { + "epoch": 2.4067647704810016, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.249783515930176, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7461479902267456, + "num_tokens": 566957596.0, + "step": 21916 + }, + { + "epoch": 2.4068745881836153, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3952600955963135, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7290328741073608, + "num_tokens": 566982957.0, + "step": 21917 + }, + { + "epoch": 2.406984405886229, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4267494678497314, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7307470440864563, + "num_tokens": 567006482.0, + "step": 21918 + }, + { + "epoch": 2.4070942235888424, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.62337327003479, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7367623448371887, + "num_tokens": 567032029.0, + "step": 21919 + }, + { + "epoch": 2.407204041291456, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4597487449645996, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7344033718109131, + "num_tokens": 567057629.0, + "step": 21920 + }, + { + "epoch": 2.40731385899407, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.296959161758423, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7240568399429321, + "num_tokens": 567085534.0, + "step": 21921 + }, + { + "epoch": 2.4074236766966837, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.1063809394836426, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7173819541931152, + "num_tokens": 567118239.0, + "step": 21922 + }, + { + "epoch": 2.407533494399297, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5483558177948, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7436611652374268, + "num_tokens": 567141682.0, + "step": 21923 + }, + { + "epoch": 2.4076433121019107, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.0926034450531006, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7298740148544312, + "num_tokens": 567174548.0, + "step": 21924 + }, + { + "epoch": 2.4077531298045245, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4845473766326904, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7429351806640625, + "num_tokens": 567197996.0, + "step": 21925 + }, + { + "epoch": 2.4078629475071383, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.0445921421051025, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7230396866798401, + "num_tokens": 567231224.0, + "step": 21926 + }, + { + "epoch": 2.4079727652097516, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.399723529815674, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7237876653671265, + "num_tokens": 567257919.0, + "step": 21927 + }, + { + "epoch": 2.4080825829123653, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5199012756347656, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7249152660369873, + "num_tokens": 567283142.0, + "step": 21928 + }, + { + "epoch": 2.408192400614979, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.348667621612549, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7163154482841492, + "num_tokens": 567308445.0, + "step": 21929 + }, + { + "epoch": 2.408302218317593, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.171818971633911, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7278785109519958, + "num_tokens": 567337026.0, + "step": 21930 + }, + { + "epoch": 2.4084120360202066, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.458606004714966, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7094814777374268, + "num_tokens": 567362751.0, + "step": 21931 + }, + { + "epoch": 2.40852185372282, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2787206172943115, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7355819940567017, + "num_tokens": 567389259.0, + "step": 21932 + }, + { + "epoch": 2.4086316714254337, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4441239833831787, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7325284481048584, + "num_tokens": 567413614.0, + "step": 21933 + }, + { + "epoch": 2.4087414891280474, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 1.9637179374694824, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7335579991340637, + "num_tokens": 567449268.0, + "step": 21934 + }, + { + "epoch": 2.408851306830661, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.672776699066162, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7509903907775879, + "num_tokens": 567469816.0, + "step": 21935 + }, + { + "epoch": 2.408961124533275, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4523372650146484, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7245476841926575, + "num_tokens": 567495710.0, + "step": 21936 + }, + { + "epoch": 2.4090709422358882, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5695059299468994, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7427433729171753, + "num_tokens": 567518594.0, + "step": 21937 + }, + { + "epoch": 2.409180759938502, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3458197116851807, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7337694764137268, + "num_tokens": 567545070.0, + "step": 21938 + }, + { + "epoch": 2.4092905776411158, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2912473678588867, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7386131286621094, + "num_tokens": 567572442.0, + "step": 21939 + }, + { + "epoch": 2.4094003953437295, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.521308660507202, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7216871976852417, + "num_tokens": 567597535.0, + "step": 21940 + }, + { + "epoch": 2.4095102130463433, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.356703519821167, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7048594951629639, + "num_tokens": 567625135.0, + "step": 21941 + }, + { + "epoch": 2.4096200307489566, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.536165237426758, + "learning_rate": 1e-06, + "loss": 0.778, + "mean_token_accuracy": 0.747105598449707, + "num_tokens": 567647853.0, + "step": 21942 + }, + { + "epoch": 2.4097298484515703, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2256176471710205, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7185648679733276, + "num_tokens": 567676905.0, + "step": 21943 + }, + { + "epoch": 2.409839666154184, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.588184118270874, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7520195841789246, + "num_tokens": 567698642.0, + "step": 21944 + }, + { + "epoch": 2.409949483856798, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.224893093109131, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7367659211158752, + "num_tokens": 567729030.0, + "step": 21945 + }, + { + "epoch": 2.4100593015594116, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.7148308753967285, + "learning_rate": 1e-06, + "loss": 0.7673, + "mean_token_accuracy": 0.7528388500213623, + "num_tokens": 567747945.0, + "step": 21946 + }, + { + "epoch": 2.410169119262025, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.350142002105713, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7275949716567993, + "num_tokens": 567773337.0, + "step": 21947 + }, + { + "epoch": 2.4102789369646387, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.471562385559082, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7139116525650024, + "num_tokens": 567798204.0, + "step": 21948 + }, + { + "epoch": 2.4103887546672524, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4034698009490967, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7222166657447815, + "num_tokens": 567824427.0, + "step": 21949 + }, + { + "epoch": 2.410498572369866, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.176091194152832, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7499039173126221, + "num_tokens": 567853385.0, + "step": 21950 + }, + { + "epoch": 2.4106083900724795, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.7066707611083984, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7371118664741516, + "num_tokens": 567874455.0, + "step": 21951 + }, + { + "epoch": 2.4107182077750933, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.1643295288085938, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7300585508346558, + "num_tokens": 567904011.0, + "step": 21952 + }, + { + "epoch": 2.410828025477707, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4059371948242188, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7115772366523743, + "num_tokens": 567929546.0, + "step": 21953 + }, + { + "epoch": 2.410937843180321, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3188512325286865, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7652559876441956, + "num_tokens": 567955058.0, + "step": 21954 + }, + { + "epoch": 2.411047660882934, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.337839126586914, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7438082695007324, + "num_tokens": 567982081.0, + "step": 21955 + }, + { + "epoch": 2.411157478585548, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.397883415222168, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7329371571540833, + "num_tokens": 568007640.0, + "step": 21956 + }, + { + "epoch": 2.4112672962881616, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5113561153411865, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7280057668685913, + "num_tokens": 568031630.0, + "step": 21957 + }, + { + "epoch": 2.4113771139907754, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.454528331756592, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7331738471984863, + "num_tokens": 568054579.0, + "step": 21958 + }, + { + "epoch": 2.411486931693389, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.503894567489624, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7473796010017395, + "num_tokens": 568078160.0, + "step": 21959 + }, + { + "epoch": 2.4115967493960024, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.386516571044922, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.72371506690979, + "num_tokens": 568102673.0, + "step": 21960 + }, + { + "epoch": 2.411706567098616, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2644596099853516, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.732917070388794, + "num_tokens": 568129707.0, + "step": 21961 + }, + { + "epoch": 2.41181638480123, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.140497922897339, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7345218062400818, + "num_tokens": 568159400.0, + "step": 21962 + }, + { + "epoch": 2.4119262025038437, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3480653762817383, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.72380131483078, + "num_tokens": 568184122.0, + "step": 21963 + }, + { + "epoch": 2.4120360202064575, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4847543239593506, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7179213762283325, + "num_tokens": 568207931.0, + "step": 21964 + }, + { + "epoch": 2.412145837909071, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.090003728866577, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.730705738067627, + "num_tokens": 568239815.0, + "step": 21965 + }, + { + "epoch": 2.4122556556116845, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 1.9342981576919556, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7137515544891357, + "num_tokens": 568277237.0, + "step": 21966 + }, + { + "epoch": 2.4123654733142983, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.0936055183410645, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7187373638153076, + "num_tokens": 568311920.0, + "step": 21967 + }, + { + "epoch": 2.412475291016912, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.7245233058929443, + "learning_rate": 1e-06, + "loss": 0.6836, + "mean_token_accuracy": 0.772631824016571, + "num_tokens": 568330760.0, + "step": 21968 + }, + { + "epoch": 2.412585108719526, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5127389430999756, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7172861695289612, + "num_tokens": 568355507.0, + "step": 21969 + }, + { + "epoch": 2.412694926422139, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.3332371711730957, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7200106382369995, + "num_tokens": 568382024.0, + "step": 21970 + }, + { + "epoch": 2.412804744124753, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.6233692169189453, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7278152108192444, + "num_tokens": 568405147.0, + "step": 21971 + }, + { + "epoch": 2.4129145618273666, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.251554489135742, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7243548035621643, + "num_tokens": 568433274.0, + "step": 21972 + }, + { + "epoch": 2.4130243795299804, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4821927547454834, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7327326536178589, + "num_tokens": 568457178.0, + "step": 21973 + }, + { + "epoch": 2.4131341972325937, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.0559747219085693, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7461187839508057, + "num_tokens": 568487821.0, + "step": 21974 + }, + { + "epoch": 2.4132440149352075, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.571596622467041, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7260819673538208, + "num_tokens": 568511891.0, + "step": 21975 + }, + { + "epoch": 2.413353832637821, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.627444267272949, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7291455268859863, + "num_tokens": 568533065.0, + "step": 21976 + }, + { + "epoch": 2.413463650340435, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2472469806671143, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7086442708969116, + "num_tokens": 568562977.0, + "step": 21977 + }, + { + "epoch": 2.4135734680430487, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2868824005126953, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7168319225311279, + "num_tokens": 568588649.0, + "step": 21978 + }, + { + "epoch": 2.413683285745662, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.761712074279785, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7316941022872925, + "num_tokens": 568609408.0, + "step": 21979 + }, + { + "epoch": 2.413793103448276, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2873728275299072, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7129014134407043, + "num_tokens": 568635159.0, + "step": 21980 + }, + { + "epoch": 2.4139029211508896, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.348681688308716, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.708105206489563, + "num_tokens": 568662805.0, + "step": 21981 + }, + { + "epoch": 2.4140127388535033, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.476759195327759, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7409734725952148, + "num_tokens": 568686176.0, + "step": 21982 + }, + { + "epoch": 2.4141225565561166, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.7916152477264404, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7219557762145996, + "num_tokens": 568706853.0, + "step": 21983 + }, + { + "epoch": 2.4142323742587304, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.1848928928375244, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7169832587242126, + "num_tokens": 568736656.0, + "step": 21984 + }, + { + "epoch": 2.414342191961344, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.532209634780884, + "learning_rate": 1e-06, + "loss": 0.7503, + "mean_token_accuracy": 0.7600522041320801, + "num_tokens": 568758166.0, + "step": 21985 + }, + { + "epoch": 2.414452009663958, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.570432424545288, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7407577037811279, + "num_tokens": 568778611.0, + "step": 21986 + }, + { + "epoch": 2.4145618273665717, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.7405941486358643, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7377375364303589, + "num_tokens": 568799349.0, + "step": 21987 + }, + { + "epoch": 2.414671645069185, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.1762640476226807, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7134414911270142, + "num_tokens": 568830941.0, + "step": 21988 + }, + { + "epoch": 2.4147814627717987, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.0946762561798096, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7126404643058777, + "num_tokens": 568861571.0, + "step": 21989 + }, + { + "epoch": 2.4148912804744125, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2414610385894775, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7381207942962646, + "num_tokens": 568889511.0, + "step": 21990 + }, + { + "epoch": 2.4150010981770262, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2414166927337646, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7450956106185913, + "num_tokens": 568916393.0, + "step": 21991 + }, + { + "epoch": 2.41511091587964, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2713725566864014, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7222185730934143, + "num_tokens": 568943616.0, + "step": 21992 + }, + { + "epoch": 2.4152207335822533, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.601154088973999, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7422521710395813, + "num_tokens": 568966047.0, + "step": 21993 + }, + { + "epoch": 2.415330551284867, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.706477642059326, + "learning_rate": 1e-06, + "loss": 0.757, + "mean_token_accuracy": 0.7572566270828247, + "num_tokens": 568987775.0, + "step": 21994 + }, + { + "epoch": 2.415440368987481, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.1302649974823, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7270488739013672, + "num_tokens": 569017181.0, + "step": 21995 + }, + { + "epoch": 2.4155501866900946, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.2554750442504883, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7453383803367615, + "num_tokens": 569043843.0, + "step": 21996 + }, + { + "epoch": 2.4156600043927083, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4853358268737793, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7238105535507202, + "num_tokens": 569070037.0, + "step": 21997 + }, + { + "epoch": 2.4157698220953217, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.738361120223999, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7416946887969971, + "num_tokens": 569090111.0, + "step": 21998 + }, + { + "epoch": 2.4158796397979354, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 3.1858980655670166, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.71436607837677, + "num_tokens": 569108485.0, + "step": 21999 + }, + { + "epoch": 2.415989457500549, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4411003589630127, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7141103744506836, + "num_tokens": 569134939.0, + "step": 22000 + }, + { + "epoch": 2.416099275203163, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.300771713256836, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7358652353286743, + "num_tokens": 569163020.0, + "step": 22001 + }, + { + "epoch": 2.4162090929057762, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.467190980911255, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7291617393493652, + "num_tokens": 569188176.0, + "step": 22002 + }, + { + "epoch": 2.41631891060839, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.164299964904785, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7414170503616333, + "num_tokens": 569216999.0, + "step": 22003 + }, + { + "epoch": 2.4164287283110037, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.077535390853882, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7321583032608032, + "num_tokens": 569247007.0, + "step": 22004 + }, + { + "epoch": 2.4165385460136175, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.8356776237487793, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7478147745132446, + "num_tokens": 569264811.0, + "step": 22005 + }, + { + "epoch": 2.416648363716231, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.589606523513794, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7238757610321045, + "num_tokens": 569289177.0, + "step": 22006 + }, + { + "epoch": 2.4167581814188446, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.089787483215332, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7170668244361877, + "num_tokens": 569320467.0, + "step": 22007 + }, + { + "epoch": 2.4168679991214583, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.5917654037475586, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7281171679496765, + "num_tokens": 569344687.0, + "step": 22008 + }, + { + "epoch": 2.416977816824072, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.05869460105896, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7242022752761841, + "num_tokens": 569376539.0, + "step": 22009 + }, + { + "epoch": 2.417087634526686, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.4203603267669678, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7484798431396484, + "num_tokens": 569399611.0, + "step": 22010 + }, + { + "epoch": 2.417197452229299, + "ewc_loss": 0.00019168853759765625, + "grad_norm": 2.607285737991333, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7084506154060364, + "num_tokens": 569425410.0, + "step": 22011 + }, + { + "epoch": 2.417307269931913, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.563230276107788, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7376433610916138, + "num_tokens": 569447821.0, + "step": 22012 + }, + { + "epoch": 2.4174170876345267, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.608485460281372, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7226923704147339, + "num_tokens": 569471760.0, + "step": 22013 + }, + { + "epoch": 2.4175269053371404, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4426538944244385, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7334105968475342, + "num_tokens": 569495548.0, + "step": 22014 + }, + { + "epoch": 2.417636723039754, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.472151279449463, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7271650433540344, + "num_tokens": 569518643.0, + "step": 22015 + }, + { + "epoch": 2.4177465407423675, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.794801712036133, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.737891674041748, + "num_tokens": 569538571.0, + "step": 22016 + }, + { + "epoch": 2.4178563584449813, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.121065139770508, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7207884788513184, + "num_tokens": 569569174.0, + "step": 22017 + }, + { + "epoch": 2.417966176147595, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3242170810699463, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7137582302093506, + "num_tokens": 569595496.0, + "step": 22018 + }, + { + "epoch": 2.4180759938502088, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.054624557495117, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7117782831192017, + "num_tokens": 569629327.0, + "step": 22019 + }, + { + "epoch": 2.4181858115528225, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.308767795562744, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7539496421813965, + "num_tokens": 569654413.0, + "step": 22020 + }, + { + "epoch": 2.418295629255436, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5599441528320312, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7537320852279663, + "num_tokens": 569676924.0, + "step": 22021 + }, + { + "epoch": 2.4184054469580496, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.258737564086914, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7300184369087219, + "num_tokens": 569706225.0, + "step": 22022 + }, + { + "epoch": 2.4185152646606634, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.41741943359375, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7359789609909058, + "num_tokens": 569729987.0, + "step": 22023 + }, + { + "epoch": 2.418625082363277, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.643065929412842, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7171831130981445, + "num_tokens": 569751708.0, + "step": 22024 + }, + { + "epoch": 2.418734900065891, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.553105115890503, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7306408882141113, + "num_tokens": 569775611.0, + "step": 22025 + }, + { + "epoch": 2.418844717768504, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.058823585510254, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.694900631904602, + "num_tokens": 569811089.0, + "step": 22026 + }, + { + "epoch": 2.418954535471118, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3328402042388916, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7255258560180664, + "num_tokens": 569836572.0, + "step": 22027 + }, + { + "epoch": 2.4190643531737317, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.809568166732788, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7467923760414124, + "num_tokens": 569856953.0, + "step": 22028 + }, + { + "epoch": 2.4191741708763455, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2925453186035156, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7478860020637512, + "num_tokens": 569883376.0, + "step": 22029 + }, + { + "epoch": 2.4192839885789588, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4920480251312256, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7261608839035034, + "num_tokens": 569907999.0, + "step": 22030 + }, + { + "epoch": 2.4193938062815725, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2025699615478516, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7261915802955627, + "num_tokens": 569938111.0, + "step": 22031 + }, + { + "epoch": 2.4195036239841863, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2805063724517822, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7157322764396667, + "num_tokens": 569966184.0, + "step": 22032 + }, + { + "epoch": 2.4196134416868, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.8243491649627686, + "learning_rate": 1e-06, + "loss": 0.73, + "mean_token_accuracy": 0.766616702079773, + "num_tokens": 569985005.0, + "step": 22033 + }, + { + "epoch": 2.4197232593894134, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4395980834960938, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7235725522041321, + "num_tokens": 570010300.0, + "step": 22034 + }, + { + "epoch": 2.419833077092027, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.52748703956604, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7399531602859497, + "num_tokens": 570032860.0, + "step": 22035 + }, + { + "epoch": 2.419942894794641, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2760064601898193, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7473005652427673, + "num_tokens": 570061059.0, + "step": 22036 + }, + { + "epoch": 2.4200527124972546, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.306941509246826, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7388013601303101, + "num_tokens": 570087969.0, + "step": 22037 + }, + { + "epoch": 2.4201625301998684, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.304164409637451, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7078287601470947, + "num_tokens": 570115084.0, + "step": 22038 + }, + { + "epoch": 2.4202723479024817, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1047003269195557, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.728107213973999, + "num_tokens": 570143869.0, + "step": 22039 + }, + { + "epoch": 2.4203821656050954, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1221346855163574, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7258478999137878, + "num_tokens": 570173988.0, + "step": 22040 + }, + { + "epoch": 2.420491983307709, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2658867835998535, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7142211198806763, + "num_tokens": 570202500.0, + "step": 22041 + }, + { + "epoch": 2.420601801010323, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6330881118774414, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7305609583854675, + "num_tokens": 570226143.0, + "step": 22042 + }, + { + "epoch": 2.4207116187129367, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5896987915039062, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7411292791366577, + "num_tokens": 570246966.0, + "step": 22043 + }, + { + "epoch": 2.42082143641555, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.139861583709717, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7490397691726685, + "num_tokens": 570276122.0, + "step": 22044 + }, + { + "epoch": 2.420931254118164, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4594571590423584, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7262173295021057, + "num_tokens": 570299470.0, + "step": 22045 + }, + { + "epoch": 2.4210410718207775, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3313536643981934, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7326003909111023, + "num_tokens": 570326235.0, + "step": 22046 + }, + { + "epoch": 2.4211508895233913, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.624976634979248, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.7446032166481018, + "num_tokens": 570346722.0, + "step": 22047 + }, + { + "epoch": 2.421260707226005, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.7379581928253174, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.717788577079773, + "num_tokens": 570368090.0, + "step": 22048 + }, + { + "epoch": 2.4213705249286184, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.647857904434204, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7357593774795532, + "num_tokens": 570391613.0, + "step": 22049 + }, + { + "epoch": 2.421480342631232, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6539316177368164, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.747846245765686, + "num_tokens": 570411032.0, + "step": 22050 + }, + { + "epoch": 2.421590160333846, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.409898281097412, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7221564054489136, + "num_tokens": 570435697.0, + "step": 22051 + }, + { + "epoch": 2.4216999780364596, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6519052982330322, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7234551906585693, + "num_tokens": 570457632.0, + "step": 22052 + }, + { + "epoch": 2.421809795739073, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.378145456314087, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7496763467788696, + "num_tokens": 570482577.0, + "step": 22053 + }, + { + "epoch": 2.4219196134416867, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.163602590560913, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7182374596595764, + "num_tokens": 570515113.0, + "step": 22054 + }, + { + "epoch": 2.4220294311443005, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1518259048461914, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7214947938919067, + "num_tokens": 570546724.0, + "step": 22055 + }, + { + "epoch": 2.4221392488469142, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.376311779022217, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7355698347091675, + "num_tokens": 570572622.0, + "step": 22056 + }, + { + "epoch": 2.4222490665495275, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.7417476177215576, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7525821924209595, + "num_tokens": 570592893.0, + "step": 22057 + }, + { + "epoch": 2.4223588842521413, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.405818223953247, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7248595952987671, + "num_tokens": 570617906.0, + "step": 22058 + }, + { + "epoch": 2.422468701954755, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6168859004974365, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7304973602294922, + "num_tokens": 570640269.0, + "step": 22059 + }, + { + "epoch": 2.422578519657369, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.8064324855804443, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.747386634349823, + "num_tokens": 570661526.0, + "step": 22060 + }, + { + "epoch": 2.4226883373599826, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.141467809677124, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7171183824539185, + "num_tokens": 570692755.0, + "step": 22061 + }, + { + "epoch": 2.422798155062596, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.62394642829895, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7348204255104065, + "num_tokens": 570715163.0, + "step": 22062 + }, + { + "epoch": 2.4229079727652096, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1286611557006836, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7113102674484253, + "num_tokens": 570746436.0, + "step": 22063 + }, + { + "epoch": 2.4230177904678234, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6822221279144287, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7180685997009277, + "num_tokens": 570768185.0, + "step": 22064 + }, + { + "epoch": 2.423127608170437, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4565987586975098, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7338171005249023, + "num_tokens": 570793782.0, + "step": 22065 + }, + { + "epoch": 2.423237425873051, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3398196697235107, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7478429079055786, + "num_tokens": 570817523.0, + "step": 22066 + }, + { + "epoch": 2.4233472435756642, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2044403553009033, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.724277675151825, + "num_tokens": 570847349.0, + "step": 22067 + }, + { + "epoch": 2.423457061278278, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.20308518409729, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.743539571762085, + "num_tokens": 570874899.0, + "step": 22068 + }, + { + "epoch": 2.4235668789808917, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.354018449783325, + "learning_rate": 1e-06, + "loss": 0.6813, + "mean_token_accuracy": 0.7756552696228027, + "num_tokens": 570898312.0, + "step": 22069 + }, + { + "epoch": 2.4236766966835055, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4945614337921143, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7279913425445557, + "num_tokens": 570923526.0, + "step": 22070 + }, + { + "epoch": 2.4237865143861193, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.11483097076416, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7336567640304565, + "num_tokens": 570955857.0, + "step": 22071 + }, + { + "epoch": 2.4238963320887326, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2705440521240234, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7463625073432922, + "num_tokens": 570982408.0, + "step": 22072 + }, + { + "epoch": 2.4240061497913463, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3165664672851562, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.727338969707489, + "num_tokens": 571012542.0, + "step": 22073 + }, + { + "epoch": 2.42411596749396, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.36321759223938, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7137952446937561, + "num_tokens": 571038782.0, + "step": 22074 + }, + { + "epoch": 2.424225785196574, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.0126681327819824, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7469592094421387, + "num_tokens": 571071042.0, + "step": 22075 + }, + { + "epoch": 2.4243356028991876, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3402159214019775, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7284747958183289, + "num_tokens": 571097481.0, + "step": 22076 + }, + { + "epoch": 2.424445420601801, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3786864280700684, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7322640419006348, + "num_tokens": 571123289.0, + "step": 22077 + }, + { + "epoch": 2.4245552383044147, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.360074758529663, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.732975959777832, + "num_tokens": 571150458.0, + "step": 22078 + }, + { + "epoch": 2.4246650560070284, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4564661979675293, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7536032199859619, + "num_tokens": 571174379.0, + "step": 22079 + }, + { + "epoch": 2.424774873709642, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.227520704269409, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7337934970855713, + "num_tokens": 571203623.0, + "step": 22080 + }, + { + "epoch": 2.4248846914122555, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.097179412841797, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7078182697296143, + "num_tokens": 571236688.0, + "step": 22081 + }, + { + "epoch": 2.4249945091148692, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1022040843963623, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7200111150741577, + "num_tokens": 571266124.0, + "step": 22082 + }, + { + "epoch": 2.425104326817483, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.523937225341797, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7173001170158386, + "num_tokens": 571289727.0, + "step": 22083 + }, + { + "epoch": 2.4252141445200968, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.217090368270874, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7320522665977478, + "num_tokens": 571317444.0, + "step": 22084 + }, + { + "epoch": 2.42532396222271, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3288493156433105, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7202224135398865, + "num_tokens": 571346632.0, + "step": 22085 + }, + { + "epoch": 2.425433779925324, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3296337127685547, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7239241003990173, + "num_tokens": 571374381.0, + "step": 22086 + }, + { + "epoch": 2.4255435976279376, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2708935737609863, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.721939206123352, + "num_tokens": 571400742.0, + "step": 22087 + }, + { + "epoch": 2.4256534153305513, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.335327386856079, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.747159481048584, + "num_tokens": 571427455.0, + "step": 22088 + }, + { + "epoch": 2.425763233033165, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4499595165252686, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7174513936042786, + "num_tokens": 571452419.0, + "step": 22089 + }, + { + "epoch": 2.4258730507357784, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3092782497406006, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7302498817443848, + "num_tokens": 571479041.0, + "step": 22090 + }, + { + "epoch": 2.425982868438392, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.482957363128662, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7443075180053711, + "num_tokens": 571501627.0, + "step": 22091 + }, + { + "epoch": 2.426092686141006, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3501851558685303, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7277094721794128, + "num_tokens": 571527922.0, + "step": 22092 + }, + { + "epoch": 2.4262025038436197, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.484921932220459, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7535240054130554, + "num_tokens": 571552399.0, + "step": 22093 + }, + { + "epoch": 2.4263123215462334, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4158389568328857, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7341543436050415, + "num_tokens": 571577848.0, + "step": 22094 + }, + { + "epoch": 2.4264221392488468, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.194490909576416, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7180868983268738, + "num_tokens": 571607660.0, + "step": 22095 + }, + { + "epoch": 2.4265319569514605, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.355882406234741, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7364808320999146, + "num_tokens": 571633207.0, + "step": 22096 + }, + { + "epoch": 2.4266417746540743, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2370119094848633, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7192451357841492, + "num_tokens": 571660053.0, + "step": 22097 + }, + { + "epoch": 2.426751592356688, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.199221134185791, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7325721979141235, + "num_tokens": 571689648.0, + "step": 22098 + }, + { + "epoch": 2.426861410059302, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2836506366729736, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7068772315979004, + "num_tokens": 571718070.0, + "step": 22099 + }, + { + "epoch": 2.426971227761915, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 1.965111255645752, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7274266481399536, + "num_tokens": 571755031.0, + "step": 22100 + }, + { + "epoch": 2.427081045464529, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4124953746795654, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7311408519744873, + "num_tokens": 571779782.0, + "step": 22101 + }, + { + "epoch": 2.4271908631671426, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6768946647644043, + "learning_rate": 1e-06, + "loss": 0.7863, + "mean_token_accuracy": 0.7489227652549744, + "num_tokens": 571800827.0, + "step": 22102 + }, + { + "epoch": 2.4273006808697564, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4264564514160156, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7242158651351929, + "num_tokens": 571823948.0, + "step": 22103 + }, + { + "epoch": 2.4274104985723697, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.7929370403289795, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7465652823448181, + "num_tokens": 571844500.0, + "step": 22104 + }, + { + "epoch": 2.4275203162749834, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3340630531311035, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7244690656661987, + "num_tokens": 571871878.0, + "step": 22105 + }, + { + "epoch": 2.427630133977597, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3196043968200684, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7173150181770325, + "num_tokens": 571899440.0, + "step": 22106 + }, + { + "epoch": 2.427739951680211, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.343942165374756, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7323006987571716, + "num_tokens": 571930948.0, + "step": 22107 + }, + { + "epoch": 2.4278497693828243, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1339328289031982, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7215983867645264, + "num_tokens": 571960027.0, + "step": 22108 + }, + { + "epoch": 2.427959587085438, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.313598871231079, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.697388768196106, + "num_tokens": 571989873.0, + "step": 22109 + }, + { + "epoch": 2.4280694047880518, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.330885171890259, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.725082516670227, + "num_tokens": 572017896.0, + "step": 22110 + }, + { + "epoch": 2.4281792224906655, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5228312015533447, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7303626537322998, + "num_tokens": 572042412.0, + "step": 22111 + }, + { + "epoch": 2.4282890401932793, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2984042167663574, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7439838647842407, + "num_tokens": 572068675.0, + "step": 22112 + }, + { + "epoch": 2.4283988578958926, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4781076908111572, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7293963432312012, + "num_tokens": 572092858.0, + "step": 22113 + }, + { + "epoch": 2.4285086755985064, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2479982376098633, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7367867827415466, + "num_tokens": 572122554.0, + "step": 22114 + }, + { + "epoch": 2.42861849330112, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.540123462677002, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7221629619598389, + "num_tokens": 572145919.0, + "step": 22115 + }, + { + "epoch": 2.428728311003734, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1910150051116943, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7210352420806885, + "num_tokens": 572174593.0, + "step": 22116 + }, + { + "epoch": 2.4288381287063476, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.758505344390869, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7372901439666748, + "num_tokens": 572194188.0, + "step": 22117 + }, + { + "epoch": 2.428947946408961, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.7482335567474365, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7373859882354736, + "num_tokens": 572213760.0, + "step": 22118 + }, + { + "epoch": 2.4290577641115747, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3229880332946777, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7204453349113464, + "num_tokens": 572240846.0, + "step": 22119 + }, + { + "epoch": 2.4291675818141885, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.825063467025757, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7371883988380432, + "num_tokens": 572260410.0, + "step": 22120 + }, + { + "epoch": 2.429277399516802, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3102357387542725, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7137376070022583, + "num_tokens": 572288503.0, + "step": 22121 + }, + { + "epoch": 2.429387217219416, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.308433771133423, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7248847484588623, + "num_tokens": 572314762.0, + "step": 22122 + }, + { + "epoch": 2.4294970349220293, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4855570793151855, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7277531623840332, + "num_tokens": 572338449.0, + "step": 22123 + }, + { + "epoch": 2.429606852624643, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1499054431915283, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7270762324333191, + "num_tokens": 572371814.0, + "step": 22124 + }, + { + "epoch": 2.429716670327257, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.364987850189209, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7170918583869934, + "num_tokens": 572401330.0, + "step": 22125 + }, + { + "epoch": 2.4298264880298706, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.106215715408325, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7431910037994385, + "num_tokens": 572431993.0, + "step": 22126 + }, + { + "epoch": 2.4299363057324843, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1730599403381348, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.749963641166687, + "num_tokens": 572461968.0, + "step": 22127 + }, + { + "epoch": 2.4300461234350976, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.399463415145874, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7108508348464966, + "num_tokens": 572487597.0, + "step": 22128 + }, + { + "epoch": 2.4301559411377114, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2698113918304443, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7389092445373535, + "num_tokens": 572513688.0, + "step": 22129 + }, + { + "epoch": 2.430265758840325, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.455212116241455, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7243885397911072, + "num_tokens": 572537892.0, + "step": 22130 + }, + { + "epoch": 2.430375576542939, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.491297483444214, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7399536371231079, + "num_tokens": 572562504.0, + "step": 22131 + }, + { + "epoch": 2.430485394245552, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5138401985168457, + "learning_rate": 1e-06, + "loss": 0.7183, + "mean_token_accuracy": 0.7670009136199951, + "num_tokens": 572585240.0, + "step": 22132 + }, + { + "epoch": 2.430595211948166, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.596912145614624, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7272913455963135, + "num_tokens": 572608644.0, + "step": 22133 + }, + { + "epoch": 2.4307050296507797, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3886237144470215, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.744957685470581, + "num_tokens": 572634180.0, + "step": 22134 + }, + { + "epoch": 2.4308148473533935, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.692356586456299, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7323088049888611, + "num_tokens": 572655918.0, + "step": 22135 + }, + { + "epoch": 2.430924665056007, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.2387378215789795, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7396328449249268, + "num_tokens": 572683866.0, + "step": 22136 + }, + { + "epoch": 2.4310344827586206, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5964555740356445, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.747103750705719, + "num_tokens": 572706081.0, + "step": 22137 + }, + { + "epoch": 2.4311443004612343, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5020737648010254, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7478152513504028, + "num_tokens": 572729135.0, + "step": 22138 + }, + { + "epoch": 2.431254118163848, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.5729875564575195, + "learning_rate": 1e-06, + "loss": 0.7661, + "mean_token_accuracy": 0.7620005011558533, + "num_tokens": 572752702.0, + "step": 22139 + }, + { + "epoch": 2.431363935866462, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.343402624130249, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7419858574867249, + "num_tokens": 572778951.0, + "step": 22140 + }, + { + "epoch": 2.431473753569075, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.797886371612549, + "learning_rate": 1e-06, + "loss": 0.7231, + "mean_token_accuracy": 0.7701760530471802, + "num_tokens": 572797376.0, + "step": 22141 + }, + { + "epoch": 2.431583571271689, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4128658771514893, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7252302169799805, + "num_tokens": 572821225.0, + "step": 22142 + }, + { + "epoch": 2.4316933889743026, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.1962478160858154, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7441772818565369, + "num_tokens": 572851640.0, + "step": 22143 + }, + { + "epoch": 2.4318032066769164, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3930721282958984, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7209503650665283, + "num_tokens": 572876784.0, + "step": 22144 + }, + { + "epoch": 2.43191302437953, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.6193621158599854, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7364550232887268, + "num_tokens": 572901633.0, + "step": 22145 + }, + { + "epoch": 2.4320228420821435, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4135541915893555, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7228782773017883, + "num_tokens": 572925554.0, + "step": 22146 + }, + { + "epoch": 2.4321326597847572, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.746720790863037, + "learning_rate": 1e-06, + "loss": 0.776, + "mean_token_accuracy": 0.7498953342437744, + "num_tokens": 572945689.0, + "step": 22147 + }, + { + "epoch": 2.432242477487371, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.422008752822876, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7339795827865601, + "num_tokens": 572969929.0, + "step": 22148 + }, + { + "epoch": 2.4323522951899847, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.4144160747528076, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7472872138023376, + "num_tokens": 572996094.0, + "step": 22149 + }, + { + "epoch": 2.4324621128925985, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.9919228553771973, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7451310753822327, + "num_tokens": 573013678.0, + "step": 22150 + }, + { + "epoch": 2.432571930595212, + "ewc_loss": 0.0001926422119140625, + "grad_norm": 2.3148765563964844, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7212569117546082, + "num_tokens": 573044076.0, + "step": 22151 + }, + { + "epoch": 2.4326817482978256, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2743520736694336, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.717126190662384, + "num_tokens": 573075911.0, + "step": 22152 + }, + { + "epoch": 2.4327915660004393, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2851855754852295, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7198242545127869, + "num_tokens": 573104044.0, + "step": 22153 + }, + { + "epoch": 2.432901383703053, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.507361650466919, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7244468927383423, + "num_tokens": 573128751.0, + "step": 22154 + }, + { + "epoch": 2.4330112014056664, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0462324619293213, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7518348693847656, + "num_tokens": 573161004.0, + "step": 22155 + }, + { + "epoch": 2.43312101910828, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.512378454208374, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7464231252670288, + "num_tokens": 573183890.0, + "step": 22156 + }, + { + "epoch": 2.433230836810894, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1316745281219482, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7073886394500732, + "num_tokens": 573214899.0, + "step": 22157 + }, + { + "epoch": 2.4333406545135077, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.220682382583618, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7251724600791931, + "num_tokens": 573244349.0, + "step": 22158 + }, + { + "epoch": 2.4334504722161214, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.564708709716797, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7344986200332642, + "num_tokens": 573266804.0, + "step": 22159 + }, + { + "epoch": 2.4335602899187347, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.8184962272644043, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7356624603271484, + "num_tokens": 573285840.0, + "step": 22160 + }, + { + "epoch": 2.4336701076213485, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1865170001983643, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7245733141899109, + "num_tokens": 573313675.0, + "step": 22161 + }, + { + "epoch": 2.4337799253239623, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.269989490509033, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7304773926734924, + "num_tokens": 573341634.0, + "step": 22162 + }, + { + "epoch": 2.433889743026576, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.122873306274414, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7121844291687012, + "num_tokens": 573372386.0, + "step": 22163 + }, + { + "epoch": 2.4339995607291893, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5921244621276855, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.746867835521698, + "num_tokens": 573393531.0, + "step": 22164 + }, + { + "epoch": 2.434109378431803, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.348818778991699, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7386505603790283, + "num_tokens": 573421354.0, + "step": 22165 + }, + { + "epoch": 2.434219196134417, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.140507459640503, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7104283571243286, + "num_tokens": 573450837.0, + "step": 22166 + }, + { + "epoch": 2.4343290138370306, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.64518141746521, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7418257594108582, + "num_tokens": 573472620.0, + "step": 22167 + }, + { + "epoch": 2.4344388315396444, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3339180946350098, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7145737409591675, + "num_tokens": 573499618.0, + "step": 22168 + }, + { + "epoch": 2.4345486492422577, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.423757553100586, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7400493621826172, + "num_tokens": 573525270.0, + "step": 22169 + }, + { + "epoch": 2.4346584669448714, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.18953800201416, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6943027973175049, + "num_tokens": 573558375.0, + "step": 22170 + }, + { + "epoch": 2.434768284647485, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5628974437713623, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7304360866546631, + "num_tokens": 573580819.0, + "step": 22171 + }, + { + "epoch": 2.434878102350099, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.609760046005249, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7194003462791443, + "num_tokens": 573603032.0, + "step": 22172 + }, + { + "epoch": 2.4349879200527127, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.558955669403076, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7515224814414978, + "num_tokens": 573624413.0, + "step": 22173 + }, + { + "epoch": 2.435097737755326, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4898788928985596, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7364557981491089, + "num_tokens": 573647899.0, + "step": 22174 + }, + { + "epoch": 2.4352075554579398, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6543214321136475, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7062368392944336, + "num_tokens": 573672654.0, + "step": 22175 + }, + { + "epoch": 2.4353173731605535, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.678100824356079, + "learning_rate": 1e-06, + "loss": 0.7521, + "mean_token_accuracy": 0.763934850692749, + "num_tokens": 573692131.0, + "step": 22176 + }, + { + "epoch": 2.4354271908631673, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.9063351154327393, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7278528809547424, + "num_tokens": 573712666.0, + "step": 22177 + }, + { + "epoch": 2.435537008565781, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.7147507667541504, + "learning_rate": 1e-06, + "loss": 0.7661, + "mean_token_accuracy": 0.7537707090377808, + "num_tokens": 573732616.0, + "step": 22178 + }, + { + "epoch": 2.4356468262683943, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4472875595092773, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7307654023170471, + "num_tokens": 573755985.0, + "step": 22179 + }, + { + "epoch": 2.435756643971008, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.38651442527771, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7124214172363281, + "num_tokens": 573781635.0, + "step": 22180 + }, + { + "epoch": 2.435866461673622, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5799779891967773, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7562322616577148, + "num_tokens": 573803827.0, + "step": 22181 + }, + { + "epoch": 2.4359762793762356, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4957902431488037, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7482734322547913, + "num_tokens": 573828707.0, + "step": 22182 + }, + { + "epoch": 2.436086097078849, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.069948434829712, + "learning_rate": 1e-06, + "loss": 0.7887, + "mean_token_accuracy": 0.7494608163833618, + "num_tokens": 573858685.0, + "step": 22183 + }, + { + "epoch": 2.4361959147814627, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.388866662979126, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7285570502281189, + "num_tokens": 573885476.0, + "step": 22184 + }, + { + "epoch": 2.4363057324840764, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.592162609100342, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7295446395874023, + "num_tokens": 573910411.0, + "step": 22185 + }, + { + "epoch": 2.43641555018669, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1584994792938232, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7391144037246704, + "num_tokens": 573940680.0, + "step": 22186 + }, + { + "epoch": 2.4365253678893035, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0121424198150635, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7436032295227051, + "num_tokens": 573974782.0, + "step": 22187 + }, + { + "epoch": 2.4366351855919173, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.327564239501953, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7006334066390991, + "num_tokens": 574001647.0, + "step": 22188 + }, + { + "epoch": 2.436745003294531, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3896615505218506, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7247096300125122, + "num_tokens": 574027374.0, + "step": 22189 + }, + { + "epoch": 2.436854820997145, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.8099637031555176, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7370444536209106, + "num_tokens": 574047848.0, + "step": 22190 + }, + { + "epoch": 2.4369646386997585, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4521493911743164, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7427040934562683, + "num_tokens": 574071840.0, + "step": 22191 + }, + { + "epoch": 2.437074456402372, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4989402294158936, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7217286229133606, + "num_tokens": 574096632.0, + "step": 22192 + }, + { + "epoch": 2.4371842741049856, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.440101385116577, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7246837615966797, + "num_tokens": 574121632.0, + "step": 22193 + }, + { + "epoch": 2.4372940918075994, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.58184552192688, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7398955821990967, + "num_tokens": 574143569.0, + "step": 22194 + }, + { + "epoch": 2.437403909510213, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6582143306732178, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7433087825775146, + "num_tokens": 574165403.0, + "step": 22195 + }, + { + "epoch": 2.437513727212827, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.273698091506958, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7215595245361328, + "num_tokens": 574192232.0, + "step": 22196 + }, + { + "epoch": 2.43762354491544, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3116238117218018, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7384915351867676, + "num_tokens": 574218583.0, + "step": 22197 + }, + { + "epoch": 2.437733362618054, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2442119121551514, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7318214774131775, + "num_tokens": 574247284.0, + "step": 22198 + }, + { + "epoch": 2.4378431803206677, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3999571800231934, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7308889627456665, + "num_tokens": 574272439.0, + "step": 22199 + }, + { + "epoch": 2.4379529980232815, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3293697834014893, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7374143600463867, + "num_tokens": 574300708.0, + "step": 22200 + }, + { + "epoch": 2.4380628157258952, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1903798580169678, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.726895272731781, + "num_tokens": 574329878.0, + "step": 22201 + }, + { + "epoch": 2.4381726334285085, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0853493213653564, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.726209282875061, + "num_tokens": 574362905.0, + "step": 22202 + }, + { + "epoch": 2.4382824511311223, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.329421043395996, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6983127593994141, + "num_tokens": 574389659.0, + "step": 22203 + }, + { + "epoch": 2.438392268833736, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.942556858062744, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7233390808105469, + "num_tokens": 574415611.0, + "step": 22204 + }, + { + "epoch": 2.43850208653635, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.188066005706787, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7148827314376831, + "num_tokens": 574444974.0, + "step": 22205 + }, + { + "epoch": 2.4386119042389636, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4909098148345947, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.745953381061554, + "num_tokens": 574467654.0, + "step": 22206 + }, + { + "epoch": 2.438721721941577, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.8873069286346436, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7498260736465454, + "num_tokens": 574487929.0, + "step": 22207 + }, + { + "epoch": 2.4388315396441906, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6596360206604004, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7339674234390259, + "num_tokens": 574510732.0, + "step": 22208 + }, + { + "epoch": 2.4389413573468044, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4198200702667236, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7257548570632935, + "num_tokens": 574536138.0, + "step": 22209 + }, + { + "epoch": 2.439051175049418, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.214203119277954, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7406216859817505, + "num_tokens": 574563499.0, + "step": 22210 + }, + { + "epoch": 2.4391609927520315, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.340636968612671, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7176522016525269, + "num_tokens": 574589829.0, + "step": 22211 + }, + { + "epoch": 2.439270810454645, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6741576194763184, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7364522814750671, + "num_tokens": 574611065.0, + "step": 22212 + }, + { + "epoch": 2.439380628157259, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.324995756149292, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7130982279777527, + "num_tokens": 574637669.0, + "step": 22213 + }, + { + "epoch": 2.4394904458598727, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.863626718521118, + "learning_rate": 1e-06, + "loss": 0.7793, + "mean_token_accuracy": 0.7552469968795776, + "num_tokens": 574656869.0, + "step": 22214 + }, + { + "epoch": 2.439600263562486, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.309112071990967, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7310038805007935, + "num_tokens": 574684604.0, + "step": 22215 + }, + { + "epoch": 2.4397100812651, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5141594409942627, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7133417725563049, + "num_tokens": 574708139.0, + "step": 22216 + }, + { + "epoch": 2.4398198989677136, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4265973567962646, + "learning_rate": 1e-06, + "loss": 0.7359, + "mean_token_accuracy": 0.7629467844963074, + "num_tokens": 574732374.0, + "step": 22217 + }, + { + "epoch": 2.4399297166703273, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2985029220581055, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7066429853439331, + "num_tokens": 574760582.0, + "step": 22218 + }, + { + "epoch": 2.440039534372941, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.40351939201355, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7353827953338623, + "num_tokens": 574786011.0, + "step": 22219 + }, + { + "epoch": 2.4401493520755544, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3881263732910156, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7256163358688354, + "num_tokens": 574812478.0, + "step": 22220 + }, + { + "epoch": 2.440259169778168, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2738256454467773, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.737339973449707, + "num_tokens": 574838868.0, + "step": 22221 + }, + { + "epoch": 2.440368987480782, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2130444049835205, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7124805450439453, + "num_tokens": 574867871.0, + "step": 22222 + }, + { + "epoch": 2.4404788051833957, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.446969747543335, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7309278249740601, + "num_tokens": 574890716.0, + "step": 22223 + }, + { + "epoch": 2.4405886228860094, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3240959644317627, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7303053736686707, + "num_tokens": 574916637.0, + "step": 22224 + }, + { + "epoch": 2.4406984405886227, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3980119228363037, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7339963912963867, + "num_tokens": 574943381.0, + "step": 22225 + }, + { + "epoch": 2.4408082582912365, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.211132526397705, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7263363003730774, + "num_tokens": 574972351.0, + "step": 22226 + }, + { + "epoch": 2.4409180759938502, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.207494020462036, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7402365803718567, + "num_tokens": 575000134.0, + "step": 22227 + }, + { + "epoch": 2.441027893696464, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.00951886177063, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7294713258743286, + "num_tokens": 575035356.0, + "step": 22228 + }, + { + "epoch": 2.4411377113990778, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.675960063934326, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7332388162612915, + "num_tokens": 575057505.0, + "step": 22229 + }, + { + "epoch": 2.441247529101691, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.389145612716675, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7426334619522095, + "num_tokens": 575081014.0, + "step": 22230 + }, + { + "epoch": 2.441357346804305, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0077009201049805, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7001378536224365, + "num_tokens": 575115581.0, + "step": 22231 + }, + { + "epoch": 2.4414671645069186, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.488670825958252, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7215313911437988, + "num_tokens": 575139436.0, + "step": 22232 + }, + { + "epoch": 2.4415769822095323, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1489925384521484, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7103970050811768, + "num_tokens": 575171593.0, + "step": 22233 + }, + { + "epoch": 2.4416867999121457, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3848588466644287, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7419400215148926, + "num_tokens": 575197302.0, + "step": 22234 + }, + { + "epoch": 2.4417966176147594, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3108057975769043, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7242818474769592, + "num_tokens": 575225667.0, + "step": 22235 + }, + { + "epoch": 2.441906435317373, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.36458683013916, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7325047254562378, + "num_tokens": 575250392.0, + "step": 22236 + }, + { + "epoch": 2.442016253019987, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6725988388061523, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7298875451087952, + "num_tokens": 575270867.0, + "step": 22237 + }, + { + "epoch": 2.4421260707226002, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.035656452178955, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7133418321609497, + "num_tokens": 575303943.0, + "step": 22238 + }, + { + "epoch": 2.442235888425214, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.599374294281006, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7333264350891113, + "num_tokens": 575325586.0, + "step": 22239 + }, + { + "epoch": 2.4423457061278278, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.812270402908325, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7376817464828491, + "num_tokens": 575345010.0, + "step": 22240 + }, + { + "epoch": 2.4424555238304415, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.314724922180176, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7058343887329102, + "num_tokens": 575371827.0, + "step": 22241 + }, + { + "epoch": 2.4425653415330553, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4353973865509033, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7423697710037231, + "num_tokens": 575397851.0, + "step": 22242 + }, + { + "epoch": 2.4426751592356686, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.529618978500366, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7264436483383179, + "num_tokens": 575420803.0, + "step": 22243 + }, + { + "epoch": 2.4427849769382823, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6567869186401367, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7462621927261353, + "num_tokens": 575441226.0, + "step": 22244 + }, + { + "epoch": 2.442894794640896, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3520052433013916, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7294685244560242, + "num_tokens": 575467323.0, + "step": 22245 + }, + { + "epoch": 2.44300461234351, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3976385593414307, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7309492826461792, + "num_tokens": 575492186.0, + "step": 22246 + }, + { + "epoch": 2.4431144300461236, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.711738109588623, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7378892302513123, + "num_tokens": 575512859.0, + "step": 22247 + }, + { + "epoch": 2.443224247748737, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5433669090270996, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7420283555984497, + "num_tokens": 575534950.0, + "step": 22248 + }, + { + "epoch": 2.4433340654513507, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.418710708618164, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7451387047767639, + "num_tokens": 575560010.0, + "step": 22249 + }, + { + "epoch": 2.4434438831539644, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.370063066482544, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7132143974304199, + "num_tokens": 575587780.0, + "step": 22250 + }, + { + "epoch": 2.443553700856578, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.415274143218994, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7206589579582214, + "num_tokens": 575613729.0, + "step": 22251 + }, + { + "epoch": 2.443663518559192, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3946280479431152, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6949937343597412, + "num_tokens": 575641897.0, + "step": 22252 + }, + { + "epoch": 2.4437733362618053, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5158205032348633, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7378212213516235, + "num_tokens": 575663939.0, + "step": 22253 + }, + { + "epoch": 2.443883153964419, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3485381603240967, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7232308387756348, + "num_tokens": 575693398.0, + "step": 22254 + }, + { + "epoch": 2.4439929716670328, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4832892417907715, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7337723970413208, + "num_tokens": 575716003.0, + "step": 22255 + }, + { + "epoch": 2.4441027893696465, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.620626926422119, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7338088750839233, + "num_tokens": 575737645.0, + "step": 22256 + }, + { + "epoch": 2.4442126070722603, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3080227375030518, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7350411415100098, + "num_tokens": 575764482.0, + "step": 22257 + }, + { + "epoch": 2.4443224247748736, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6130218505859375, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7207942008972168, + "num_tokens": 575785339.0, + "step": 22258 + }, + { + "epoch": 2.4444322424774874, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.352381467819214, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.7597329020500183, + "num_tokens": 575810353.0, + "step": 22259 + }, + { + "epoch": 2.444542060180101, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4447717666625977, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7246097326278687, + "num_tokens": 575837686.0, + "step": 22260 + }, + { + "epoch": 2.444651877882715, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0759382247924805, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7059116363525391, + "num_tokens": 575873131.0, + "step": 22261 + }, + { + "epoch": 2.444761695585328, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1860601902008057, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7096830010414124, + "num_tokens": 575901282.0, + "step": 22262 + }, + { + "epoch": 2.444871513287942, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.609482526779175, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7343366146087646, + "num_tokens": 575923327.0, + "step": 22263 + }, + { + "epoch": 2.4449813309905557, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1225507259368896, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7184752225875854, + "num_tokens": 575955089.0, + "step": 22264 + }, + { + "epoch": 2.4450911486931695, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3652055263519287, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7093644142150879, + "num_tokens": 575982716.0, + "step": 22265 + }, + { + "epoch": 2.4452009663957828, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2565224170684814, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7055765986442566, + "num_tokens": 576012552.0, + "step": 22266 + }, + { + "epoch": 2.4453107840983965, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6502506732940674, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7407071590423584, + "num_tokens": 576033064.0, + "step": 22267 + }, + { + "epoch": 2.4454206018010103, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.13664174079895, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7126256227493286, + "num_tokens": 576063360.0, + "step": 22268 + }, + { + "epoch": 2.445530419503624, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3049581050872803, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7479585409164429, + "num_tokens": 576088138.0, + "step": 22269 + }, + { + "epoch": 2.445640237206238, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.04807710647583, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.706618070602417, + "num_tokens": 576120804.0, + "step": 22270 + }, + { + "epoch": 2.445750054908851, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.298511505126953, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7318612337112427, + "num_tokens": 576146306.0, + "step": 22271 + }, + { + "epoch": 2.445859872611465, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.533543825149536, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7232584357261658, + "num_tokens": 576171014.0, + "step": 22272 + }, + { + "epoch": 2.4459696903140786, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3570101261138916, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7207122445106506, + "num_tokens": 576197866.0, + "step": 22273 + }, + { + "epoch": 2.4460795080166924, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3973779678344727, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.6992882490158081, + "num_tokens": 576225642.0, + "step": 22274 + }, + { + "epoch": 2.446189325719306, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.242321252822876, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7333050966262817, + "num_tokens": 576253010.0, + "step": 22275 + }, + { + "epoch": 2.4462991434219195, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.311142683029175, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7229635715484619, + "num_tokens": 576279767.0, + "step": 22276 + }, + { + "epoch": 2.446408961124533, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2043890953063965, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7105857729911804, + "num_tokens": 576310196.0, + "step": 22277 + }, + { + "epoch": 2.446518778827147, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3973228931427, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7047218680381775, + "num_tokens": 576336773.0, + "step": 22278 + }, + { + "epoch": 2.4466285965297607, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.68001651763916, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7341219186782837, + "num_tokens": 576358569.0, + "step": 22279 + }, + { + "epoch": 2.4467384142323745, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3088643550872803, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7348329424858093, + "num_tokens": 576386669.0, + "step": 22280 + }, + { + "epoch": 2.446848231934988, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.493785858154297, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7402598261833191, + "num_tokens": 576409761.0, + "step": 22281 + }, + { + "epoch": 2.4469580496376016, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.6678833961486816, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7406100630760193, + "num_tokens": 576432019.0, + "step": 22282 + }, + { + "epoch": 2.4470678673402153, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3968112468719482, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7306482195854187, + "num_tokens": 576459040.0, + "step": 22283 + }, + { + "epoch": 2.447177685042829, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3290321826934814, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7488978505134583, + "num_tokens": 576485630.0, + "step": 22284 + }, + { + "epoch": 2.4472875027454424, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.404064178466797, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7435159683227539, + "num_tokens": 576510185.0, + "step": 22285 + }, + { + "epoch": 2.447397320448056, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.622265100479126, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7292057871818542, + "num_tokens": 576533097.0, + "step": 22286 + }, + { + "epoch": 2.44750713815067, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.88358736038208, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7343413829803467, + "num_tokens": 576552424.0, + "step": 22287 + }, + { + "epoch": 2.4476169558532836, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.593811273574829, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7501038908958435, + "num_tokens": 576572851.0, + "step": 22288 + }, + { + "epoch": 2.447726773555897, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2706141471862793, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7385585904121399, + "num_tokens": 576599655.0, + "step": 22289 + }, + { + "epoch": 2.4478365912585107, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.133212089538574, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7283265590667725, + "num_tokens": 576628544.0, + "step": 22290 + }, + { + "epoch": 2.4479464089611245, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1792874336242676, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.742389440536499, + "num_tokens": 576657072.0, + "step": 22291 + }, + { + "epoch": 2.4480562266637382, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.307466983795166, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7261334657669067, + "num_tokens": 576683784.0, + "step": 22292 + }, + { + "epoch": 2.448166044366352, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.517840623855591, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7426636219024658, + "num_tokens": 576706149.0, + "step": 22293 + }, + { + "epoch": 2.4482758620689653, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5698142051696777, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7275668382644653, + "num_tokens": 576727558.0, + "step": 22294 + }, + { + "epoch": 2.448385679771579, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.252530097961426, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7316162586212158, + "num_tokens": 576754030.0, + "step": 22295 + }, + { + "epoch": 2.448495497474193, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.296306610107422, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7307628393173218, + "num_tokens": 576782912.0, + "step": 22296 + }, + { + "epoch": 2.4486053151768066, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.352567672729492, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7314940690994263, + "num_tokens": 576807119.0, + "step": 22297 + }, + { + "epoch": 2.4487151328794203, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4884886741638184, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7276546955108643, + "num_tokens": 576830037.0, + "step": 22298 + }, + { + "epoch": 2.4488249505820336, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5170509815216064, + "learning_rate": 1e-06, + "loss": 0.7327, + "mean_token_accuracy": 0.7679427862167358, + "num_tokens": 576851476.0, + "step": 22299 + }, + { + "epoch": 2.4489347682846474, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5739121437072754, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7183098793029785, + "num_tokens": 576875256.0, + "step": 22300 + }, + { + "epoch": 2.449044585987261, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1959898471832275, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7272684574127197, + "num_tokens": 576905309.0, + "step": 22301 + }, + { + "epoch": 2.449154403689875, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5705485343933105, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7254825830459595, + "num_tokens": 576930590.0, + "step": 22302 + }, + { + "epoch": 2.4492642213924887, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.640812635421753, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7328366041183472, + "num_tokens": 576952543.0, + "step": 22303 + }, + { + "epoch": 2.449374039095102, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3876121044158936, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7198456525802612, + "num_tokens": 576978843.0, + "step": 22304 + }, + { + "epoch": 2.4494838567977157, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.438497543334961, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7237109541893005, + "num_tokens": 577004525.0, + "step": 22305 + }, + { + "epoch": 2.4495936745003295, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4353911876678467, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7323018312454224, + "num_tokens": 577028989.0, + "step": 22306 + }, + { + "epoch": 2.4497034922029433, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.659024477005005, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7497154474258423, + "num_tokens": 577049797.0, + "step": 22307 + }, + { + "epoch": 2.449813309905557, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.57027006149292, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7347057461738586, + "num_tokens": 577072989.0, + "step": 22308 + }, + { + "epoch": 2.4499231276081703, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3803677558898926, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7306253910064697, + "num_tokens": 577097721.0, + "step": 22309 + }, + { + "epoch": 2.450032945310784, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.208589553833008, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7096608877182007, + "num_tokens": 577125235.0, + "step": 22310 + }, + { + "epoch": 2.450142763013398, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.552551031112671, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7489116191864014, + "num_tokens": 577146621.0, + "step": 22311 + }, + { + "epoch": 2.4502525807160116, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.8177292346954346, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.7513231635093689, + "num_tokens": 577165943.0, + "step": 22312 + }, + { + "epoch": 2.450362398418625, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.629643678665161, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7201715707778931, + "num_tokens": 577187725.0, + "step": 22313 + }, + { + "epoch": 2.4504722161212387, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3377578258514404, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7402467131614685, + "num_tokens": 577214425.0, + "step": 22314 + }, + { + "epoch": 2.4505820338238524, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3157742023468018, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7358497381210327, + "num_tokens": 577241342.0, + "step": 22315 + }, + { + "epoch": 2.450691851526466, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.453338623046875, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.743563175201416, + "num_tokens": 577264693.0, + "step": 22316 + }, + { + "epoch": 2.4508016692290795, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2454631328582764, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.737690806388855, + "num_tokens": 577293109.0, + "step": 22317 + }, + { + "epoch": 2.4509114869316933, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.252854824066162, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7326594591140747, + "num_tokens": 577322748.0, + "step": 22318 + }, + { + "epoch": 2.451021304634307, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.453279733657837, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7301880121231079, + "num_tokens": 577346093.0, + "step": 22319 + }, + { + "epoch": 2.4511311223369208, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1656510829925537, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7318738102912903, + "num_tokens": 577376360.0, + "step": 22320 + }, + { + "epoch": 2.4512409400395345, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.19533109664917, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.723213255405426, + "num_tokens": 577406450.0, + "step": 22321 + }, + { + "epoch": 2.451350757742148, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.651360511779785, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7163591980934143, + "num_tokens": 577428068.0, + "step": 22322 + }, + { + "epoch": 2.4514605754447616, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.466672420501709, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7390464544296265, + "num_tokens": 577451745.0, + "step": 22323 + }, + { + "epoch": 2.4515703931473753, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.665907621383667, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7123427391052246, + "num_tokens": 577473353.0, + "step": 22324 + }, + { + "epoch": 2.451680210849989, + "ewc_loss": 0.000194549560546875, + "grad_norm": 1.947493076324463, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7243310213088989, + "num_tokens": 577511202.0, + "step": 22325 + }, + { + "epoch": 2.451790028552603, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.221748113632202, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.742500901222229, + "num_tokens": 577540374.0, + "step": 22326 + }, + { + "epoch": 2.451899846255216, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.155217409133911, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7064264416694641, + "num_tokens": 577571983.0, + "step": 22327 + }, + { + "epoch": 2.45200966395783, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.535364866256714, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7224093675613403, + "num_tokens": 577596579.0, + "step": 22328 + }, + { + "epoch": 2.4521194816604437, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1860358715057373, + "learning_rate": 1e-06, + "loss": 0.7557, + "mean_token_accuracy": 0.7500645518302917, + "num_tokens": 577626159.0, + "step": 22329 + }, + { + "epoch": 2.4522292993630574, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.272066593170166, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7195298671722412, + "num_tokens": 577653838.0, + "step": 22330 + }, + { + "epoch": 2.452339117065671, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2644107341766357, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7237615585327148, + "num_tokens": 577681468.0, + "step": 22331 + }, + { + "epoch": 2.4524489347682845, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.215341329574585, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7259271144866943, + "num_tokens": 577711263.0, + "step": 22332 + }, + { + "epoch": 2.4525587524708983, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4814035892486572, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.730948269367218, + "num_tokens": 577734729.0, + "step": 22333 + }, + { + "epoch": 2.452668570173512, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.683861017227173, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7404448390007019, + "num_tokens": 577756749.0, + "step": 22334 + }, + { + "epoch": 2.452778387876126, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1807773113250732, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7265235781669617, + "num_tokens": 577785506.0, + "step": 22335 + }, + { + "epoch": 2.452888205578739, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4783198833465576, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7317253351211548, + "num_tokens": 577808889.0, + "step": 22336 + }, + { + "epoch": 2.452998023281353, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.318786859512329, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7171887755393982, + "num_tokens": 577837719.0, + "step": 22337 + }, + { + "epoch": 2.4531078409839666, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3786652088165283, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7352198958396912, + "num_tokens": 577863447.0, + "step": 22338 + }, + { + "epoch": 2.4532176586865804, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.673825263977051, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7281183004379272, + "num_tokens": 577883916.0, + "step": 22339 + }, + { + "epoch": 2.453327476389194, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3917407989501953, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7103099822998047, + "num_tokens": 577910642.0, + "step": 22340 + }, + { + "epoch": 2.4534372940918074, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.42374849319458, + "learning_rate": 1e-06, + "loss": 0.7527, + "mean_token_accuracy": 0.7555322647094727, + "num_tokens": 577934895.0, + "step": 22341 + }, + { + "epoch": 2.453547111794421, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.434032917022705, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7270209789276123, + "num_tokens": 577960405.0, + "step": 22342 + }, + { + "epoch": 2.453656929497035, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1358325481414795, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6939247250556946, + "num_tokens": 577991669.0, + "step": 22343 + }, + { + "epoch": 2.4537667471996487, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5155975818634033, + "learning_rate": 1e-06, + "loss": 0.769, + "mean_token_accuracy": 0.7540823221206665, + "num_tokens": 578012901.0, + "step": 22344 + }, + { + "epoch": 2.453876564902262, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.66644024848938, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7485507726669312, + "num_tokens": 578033330.0, + "step": 22345 + }, + { + "epoch": 2.453986382604876, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.8378262519836426, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.7469967603683472, + "num_tokens": 578053690.0, + "step": 22346 + }, + { + "epoch": 2.4540962003074895, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.231515407562256, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7285913825035095, + "num_tokens": 578084004.0, + "step": 22347 + }, + { + "epoch": 2.4542060180101033, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.351811647415161, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7398190498352051, + "num_tokens": 578109929.0, + "step": 22348 + }, + { + "epoch": 2.454315835712717, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1220977306365967, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7217780351638794, + "num_tokens": 578143144.0, + "step": 22349 + }, + { + "epoch": 2.4544256534153304, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.1316440105438232, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7199114561080933, + "num_tokens": 578172629.0, + "step": 22350 + }, + { + "epoch": 2.454535471117944, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4072389602661133, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7331893444061279, + "num_tokens": 578196976.0, + "step": 22351 + }, + { + "epoch": 2.454645288820558, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.302785873413086, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7142784595489502, + "num_tokens": 578226395.0, + "step": 22352 + }, + { + "epoch": 2.4547551065231716, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.324671506881714, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7234129905700684, + "num_tokens": 578254367.0, + "step": 22353 + }, + { + "epoch": 2.4548649242257854, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3485281467437744, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7311029434204102, + "num_tokens": 578279572.0, + "step": 22354 + }, + { + "epoch": 2.4549747419283987, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4894726276397705, + "learning_rate": 1e-06, + "loss": 0.6949, + "mean_token_accuracy": 0.7744405269622803, + "num_tokens": 578302458.0, + "step": 22355 + }, + { + "epoch": 2.4550845596310125, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.128478527069092, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7136876583099365, + "num_tokens": 578333129.0, + "step": 22356 + }, + { + "epoch": 2.455194377333626, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4283034801483154, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7445806264877319, + "num_tokens": 578357560.0, + "step": 22357 + }, + { + "epoch": 2.45530419503624, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2176122665405273, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.741375207901001, + "num_tokens": 578384080.0, + "step": 22358 + }, + { + "epoch": 2.4554140127388537, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.246614456176758, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7148652672767639, + "num_tokens": 578413855.0, + "step": 22359 + }, + { + "epoch": 2.455523830441467, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4371581077575684, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7289168834686279, + "num_tokens": 578441063.0, + "step": 22360 + }, + { + "epoch": 2.455633648144081, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3031094074249268, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7152093648910522, + "num_tokens": 578469577.0, + "step": 22361 + }, + { + "epoch": 2.4557434658466946, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.444108247756958, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7468438148498535, + "num_tokens": 578492684.0, + "step": 22362 + }, + { + "epoch": 2.4558532835493083, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4689764976501465, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7280631065368652, + "num_tokens": 578516977.0, + "step": 22363 + }, + { + "epoch": 2.4559631012519216, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5006120204925537, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7156602740287781, + "num_tokens": 578542266.0, + "step": 22364 + }, + { + "epoch": 2.4560729189545354, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.511356830596924, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7348278760910034, + "num_tokens": 578564821.0, + "step": 22365 + }, + { + "epoch": 2.456182736657149, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.846442461013794, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7360846996307373, + "num_tokens": 578586215.0, + "step": 22366 + }, + { + "epoch": 2.456292554359763, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.401963949203491, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7232603430747986, + "num_tokens": 578612609.0, + "step": 22367 + }, + { + "epoch": 2.456402372062376, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.485055685043335, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7086180448532104, + "num_tokens": 578637678.0, + "step": 22368 + }, + { + "epoch": 2.45651218976499, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0921988487243652, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7142516374588013, + "num_tokens": 578669710.0, + "step": 22369 + }, + { + "epoch": 2.4566220074676037, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.4762260913848877, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7474578619003296, + "num_tokens": 578693789.0, + "step": 22370 + }, + { + "epoch": 2.4567318251702175, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.198333263397217, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7276531457901001, + "num_tokens": 578724447.0, + "step": 22371 + }, + { + "epoch": 2.4568416428728312, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.322319269180298, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7237579822540283, + "num_tokens": 578750989.0, + "step": 22372 + }, + { + "epoch": 2.4569514605754446, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.5843091011047363, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7446920871734619, + "num_tokens": 578772778.0, + "step": 22373 + }, + { + "epoch": 2.4570612782780583, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.327824354171753, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7208795547485352, + "num_tokens": 578800895.0, + "step": 22374 + }, + { + "epoch": 2.457171095980672, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.180138349533081, + "learning_rate": 1e-06, + "loss": 0.7583, + "mean_token_accuracy": 0.7539708018302917, + "num_tokens": 578829230.0, + "step": 22375 + }, + { + "epoch": 2.457280913683286, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.2224903106689453, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7150357365608215, + "num_tokens": 578860437.0, + "step": 22376 + }, + { + "epoch": 2.4573907313858996, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3313629627227783, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7072848081588745, + "num_tokens": 578887887.0, + "step": 22377 + }, + { + "epoch": 2.457500549088513, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.3314712047576904, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7187638282775879, + "num_tokens": 578916142.0, + "step": 22378 + }, + { + "epoch": 2.4576103667911267, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.305356025695801, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7320249080657959, + "num_tokens": 578943239.0, + "step": 22379 + }, + { + "epoch": 2.4577201844937404, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.238441228866577, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7420755624771118, + "num_tokens": 578970595.0, + "step": 22380 + }, + { + "epoch": 2.457830002196354, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2382802963256836, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7259829640388489, + "num_tokens": 578999841.0, + "step": 22381 + }, + { + "epoch": 2.457939819898968, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.24358868598938, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.6950007081031799, + "num_tokens": 579028000.0, + "step": 22382 + }, + { + "epoch": 2.4580496376015812, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.504286527633667, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7256359457969666, + "num_tokens": 579051268.0, + "step": 22383 + }, + { + "epoch": 2.458159455304195, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3518295288085938, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7081660032272339, + "num_tokens": 579079968.0, + "step": 22384 + }, + { + "epoch": 2.4582692730068088, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.747626304626465, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7302400469779968, + "num_tokens": 579100722.0, + "step": 22385 + }, + { + "epoch": 2.4583790907094225, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2521719932556152, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7293386459350586, + "num_tokens": 579129843.0, + "step": 22386 + }, + { + "epoch": 2.4584889084120363, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.101001739501953, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7214133739471436, + "num_tokens": 579164875.0, + "step": 22387 + }, + { + "epoch": 2.4585987261146496, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.430260181427002, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7143807411193848, + "num_tokens": 579190306.0, + "step": 22388 + }, + { + "epoch": 2.4587085438172633, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.739091157913208, + "learning_rate": 1e-06, + "loss": 0.7425, + "mean_token_accuracy": 0.7673944234848022, + "num_tokens": 579209165.0, + "step": 22389 + }, + { + "epoch": 2.458818361519877, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3967716693878174, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7093796730041504, + "num_tokens": 579236012.0, + "step": 22390 + }, + { + "epoch": 2.458928179222491, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.195812463760376, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7113874554634094, + "num_tokens": 579266172.0, + "step": 22391 + }, + { + "epoch": 2.459037996925104, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3161513805389404, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7310144305229187, + "num_tokens": 579293066.0, + "step": 22392 + }, + { + "epoch": 2.459147814627718, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.383878231048584, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7221236228942871, + "num_tokens": 579318838.0, + "step": 22393 + }, + { + "epoch": 2.4592576323303317, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.0819685459136963, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7163287997245789, + "num_tokens": 579352649.0, + "step": 22394 + }, + { + "epoch": 2.4593674500329454, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3436827659606934, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7400462031364441, + "num_tokens": 579376862.0, + "step": 22395 + }, + { + "epoch": 2.4594772677355587, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2256155014038086, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7453267574310303, + "num_tokens": 579404419.0, + "step": 22396 + }, + { + "epoch": 2.4595870854381725, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.389171600341797, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7292874455451965, + "num_tokens": 579431207.0, + "step": 22397 + }, + { + "epoch": 2.4596969031407863, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 1.9763656854629517, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7308696508407593, + "num_tokens": 579463351.0, + "step": 22398 + }, + { + "epoch": 2.4598067208434, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.359609842300415, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7412068843841553, + "num_tokens": 579490210.0, + "step": 22399 + }, + { + "epoch": 2.4599165385460138, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.407278060913086, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7606291770935059, + "num_tokens": 579514549.0, + "step": 22400 + }, + { + "epoch": 2.460026356248627, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4637515544891357, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7344226837158203, + "num_tokens": 579538738.0, + "step": 22401 + }, + { + "epoch": 2.460136173951241, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.1904194355010986, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7192668318748474, + "num_tokens": 579568753.0, + "step": 22402 + }, + { + "epoch": 2.4602459916538546, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4297878742218018, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7455559372901917, + "num_tokens": 579593085.0, + "step": 22403 + }, + { + "epoch": 2.4603558093564684, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.5970678329467773, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7154669761657715, + "num_tokens": 579615851.0, + "step": 22404 + }, + { + "epoch": 2.460465627059082, + "ewc_loss": 0.000194549560546875, + "grad_norm": 2.246387004852295, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7502583861351013, + "num_tokens": 579641659.0, + "step": 22405 + }, + { + "epoch": 2.4605754447616954, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.5478153228759766, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7380455732345581, + "num_tokens": 579665632.0, + "step": 22406 + }, + { + "epoch": 2.460685262464309, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.37252140045166, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7539831399917603, + "num_tokens": 579689755.0, + "step": 22407 + }, + { + "epoch": 2.460795080166923, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4489240646362305, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7376961708068848, + "num_tokens": 579713267.0, + "step": 22408 + }, + { + "epoch": 2.4609048978695367, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.368062734603882, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7086194753646851, + "num_tokens": 579739361.0, + "step": 22409 + }, + { + "epoch": 2.4610147155721505, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.561692476272583, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7138826847076416, + "num_tokens": 579761589.0, + "step": 22410 + }, + { + "epoch": 2.4611245332747638, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4771316051483154, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7173364162445068, + "num_tokens": 579788253.0, + "step": 22411 + }, + { + "epoch": 2.4612343509773775, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.245396852493286, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7046299576759338, + "num_tokens": 579817323.0, + "step": 22412 + }, + { + "epoch": 2.4613441686799913, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4704630374908447, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7306855916976929, + "num_tokens": 579840291.0, + "step": 22413 + }, + { + "epoch": 2.461453986382605, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2100203037261963, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7028257846832275, + "num_tokens": 579872806.0, + "step": 22414 + }, + { + "epoch": 2.4615638040852184, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.0696139335632324, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7219458818435669, + "num_tokens": 579905609.0, + "step": 22415 + }, + { + "epoch": 2.461673621787832, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2404332160949707, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7256181240081787, + "num_tokens": 579934455.0, + "step": 22416 + }, + { + "epoch": 2.461783439490446, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.6279165744781494, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7517314553260803, + "num_tokens": 579956473.0, + "step": 22417 + }, + { + "epoch": 2.4618932571930596, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3612051010131836, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7129379510879517, + "num_tokens": 579983569.0, + "step": 22418 + }, + { + "epoch": 2.462003074895673, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.255182981491089, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7285633087158203, + "num_tokens": 580010210.0, + "step": 22419 + }, + { + "epoch": 2.4621128925982867, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.459019184112549, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6992002725601196, + "num_tokens": 580035267.0, + "step": 22420 + }, + { + "epoch": 2.4622227103009005, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.5353550910949707, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7380977869033813, + "num_tokens": 580059049.0, + "step": 22421 + }, + { + "epoch": 2.462332528003514, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3597424030303955, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7462561130523682, + "num_tokens": 580083726.0, + "step": 22422 + }, + { + "epoch": 2.462442345706128, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.450517177581787, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7185176610946655, + "num_tokens": 580109057.0, + "step": 22423 + }, + { + "epoch": 2.4625521634087413, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.518669366836548, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7307391166687012, + "num_tokens": 580131527.0, + "step": 22424 + }, + { + "epoch": 2.462661981111355, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3695051670074463, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7485742568969727, + "num_tokens": 580158296.0, + "step": 22425 + }, + { + "epoch": 2.462771798813969, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.29561710357666, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7199050784111023, + "num_tokens": 580185587.0, + "step": 22426 + }, + { + "epoch": 2.4628816165165826, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3080759048461914, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7318356037139893, + "num_tokens": 580211264.0, + "step": 22427 + }, + { + "epoch": 2.4629914342191963, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.290194511413574, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.744295060634613, + "num_tokens": 580238128.0, + "step": 22428 + }, + { + "epoch": 2.4631012519218096, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.338284730911255, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7295955419540405, + "num_tokens": 580263380.0, + "step": 22429 + }, + { + "epoch": 2.4632110696244234, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.309049367904663, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6961996555328369, + "num_tokens": 580291588.0, + "step": 22430 + }, + { + "epoch": 2.463320887327037, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3020784854888916, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7208313941955566, + "num_tokens": 580321155.0, + "step": 22431 + }, + { + "epoch": 2.463430705029651, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.349057912826538, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7303393483161926, + "num_tokens": 580346786.0, + "step": 22432 + }, + { + "epoch": 2.4635405227322646, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.5239052772521973, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7415278553962708, + "num_tokens": 580368009.0, + "step": 22433 + }, + { + "epoch": 2.463650340434878, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.4977035522460938, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7450677156448364, + "num_tokens": 580391179.0, + "step": 22434 + }, + { + "epoch": 2.4637601581374917, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2462666034698486, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7247618436813354, + "num_tokens": 580417400.0, + "step": 22435 + }, + { + "epoch": 2.4638699758401055, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.565288782119751, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7344895005226135, + "num_tokens": 580440859.0, + "step": 22436 + }, + { + "epoch": 2.4639797935427192, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.383324146270752, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7255470156669617, + "num_tokens": 580465329.0, + "step": 22437 + }, + { + "epoch": 2.464089611245333, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.175236940383911, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7463421821594238, + "num_tokens": 580494802.0, + "step": 22438 + }, + { + "epoch": 2.4641994289479463, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2854747772216797, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7420943975448608, + "num_tokens": 580520357.0, + "step": 22439 + }, + { + "epoch": 2.46430924665056, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.304860830307007, + "learning_rate": 1e-06, + "loss": 0.7129, + "mean_token_accuracy": 0.7702423334121704, + "num_tokens": 580544172.0, + "step": 22440 + }, + { + "epoch": 2.464419064353174, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.6413466930389404, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7473746538162231, + "num_tokens": 580565362.0, + "step": 22441 + }, + { + "epoch": 2.4645288820557876, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.2735719680786133, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7331987619400024, + "num_tokens": 580592144.0, + "step": 22442 + }, + { + "epoch": 2.464638699758401, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.3300304412841797, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7235944271087646, + "num_tokens": 580619409.0, + "step": 22443 + }, + { + "epoch": 2.4647485174610146, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.222522735595703, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7301517724990845, + "num_tokens": 580647254.0, + "step": 22444 + }, + { + "epoch": 2.4648583351636284, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.539761543273926, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7350727915763855, + "num_tokens": 580669445.0, + "step": 22445 + }, + { + "epoch": 2.464968152866242, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.231337070465088, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.741642415523529, + "num_tokens": 580698773.0, + "step": 22446 + }, + { + "epoch": 2.4650779705688555, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.130627155303955, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7074733376502991, + "num_tokens": 580728787.0, + "step": 22447 + }, + { + "epoch": 2.4651877882714692, + "ewc_loss": 0.00019550323486328125, + "grad_norm": 2.1284830570220947, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7287309169769287, + "num_tokens": 580758320.0, + "step": 22448 + }, + { + "epoch": 2.465297605974083, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.171356678009033, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7288131713867188, + "num_tokens": 580787451.0, + "step": 22449 + }, + { + "epoch": 2.4654074236766967, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2125182151794434, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7346259951591492, + "num_tokens": 580813681.0, + "step": 22450 + }, + { + "epoch": 2.4655172413793105, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3657426834106445, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.719016432762146, + "num_tokens": 580838606.0, + "step": 22451 + }, + { + "epoch": 2.465627059081924, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.278701066970825, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.734745979309082, + "num_tokens": 580866549.0, + "step": 22452 + }, + { + "epoch": 2.4657368767845376, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3133585453033447, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7170305252075195, + "num_tokens": 580893306.0, + "step": 22453 + }, + { + "epoch": 2.4658466944871513, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5964202880859375, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7317416667938232, + "num_tokens": 580916349.0, + "step": 22454 + }, + { + "epoch": 2.465956512189765, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6905150413513184, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7514338493347168, + "num_tokens": 580936238.0, + "step": 22455 + }, + { + "epoch": 2.466066329892379, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.0658669471740723, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7038629055023193, + "num_tokens": 580970218.0, + "step": 22456 + }, + { + "epoch": 2.466176147594992, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.597390651702881, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7424449920654297, + "num_tokens": 580993275.0, + "step": 22457 + }, + { + "epoch": 2.466285965297606, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.000999689102173, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7432133555412292, + "num_tokens": 581023208.0, + "step": 22458 + }, + { + "epoch": 2.4663957830002197, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.393848180770874, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7597835063934326, + "num_tokens": 581047242.0, + "step": 22459 + }, + { + "epoch": 2.4665056007028334, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.578195095062256, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.747923731803894, + "num_tokens": 581069604.0, + "step": 22460 + }, + { + "epoch": 2.466615418405447, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5218255519866943, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.739033043384552, + "num_tokens": 581093665.0, + "step": 22461 + }, + { + "epoch": 2.4667252361080605, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5065789222717285, + "learning_rate": 1e-06, + "loss": 0.7445, + "mean_token_accuracy": 0.7582312822341919, + "num_tokens": 581117503.0, + "step": 22462 + }, + { + "epoch": 2.4668350538106742, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4496304988861084, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.743069052696228, + "num_tokens": 581141075.0, + "step": 22463 + }, + { + "epoch": 2.466944871513288, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2786355018615723, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7190076112747192, + "num_tokens": 581168850.0, + "step": 22464 + }, + { + "epoch": 2.4670546892159018, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.380382537841797, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7334020137786865, + "num_tokens": 581194901.0, + "step": 22465 + }, + { + "epoch": 2.467164506918515, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2455861568450928, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.724661648273468, + "num_tokens": 581223253.0, + "step": 22466 + }, + { + "epoch": 2.467274324621129, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.085411548614502, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7084410190582275, + "num_tokens": 581257657.0, + "step": 22467 + }, + { + "epoch": 2.4673841423237426, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.679555892944336, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7311376333236694, + "num_tokens": 581280242.0, + "step": 22468 + }, + { + "epoch": 2.4674939600263563, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.288961410522461, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7251285910606384, + "num_tokens": 581309213.0, + "step": 22469 + }, + { + "epoch": 2.4676037777289697, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.815511465072632, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.734373927116394, + "num_tokens": 581328185.0, + "step": 22470 + }, + { + "epoch": 2.4677135954315834, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6283018589019775, + "learning_rate": 1e-06, + "loss": 0.7501, + "mean_token_accuracy": 0.7639052867889404, + "num_tokens": 581348910.0, + "step": 22471 + }, + { + "epoch": 2.467823413134197, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3879024982452393, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7324165105819702, + "num_tokens": 581375189.0, + "step": 22472 + }, + { + "epoch": 2.467933230836811, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5519073009490967, + "learning_rate": 1e-06, + "loss": 0.776, + "mean_token_accuracy": 0.7495497465133667, + "num_tokens": 581396012.0, + "step": 22473 + }, + { + "epoch": 2.4680430485394247, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.28022837638855, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7336323261260986, + "num_tokens": 581423460.0, + "step": 22474 + }, + { + "epoch": 2.468152866242038, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4637908935546875, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7461069226264954, + "num_tokens": 581449750.0, + "step": 22475 + }, + { + "epoch": 2.4682626839446518, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.572686195373535, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7420053482055664, + "num_tokens": 581470226.0, + "step": 22476 + }, + { + "epoch": 2.4683725016472655, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3617115020751953, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7259308099746704, + "num_tokens": 581496730.0, + "step": 22477 + }, + { + "epoch": 2.4684823193498793, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.346590995788574, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7394635677337646, + "num_tokens": 581520234.0, + "step": 22478 + }, + { + "epoch": 2.468592137052493, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.280438184738159, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.734025776386261, + "num_tokens": 581548145.0, + "step": 22479 + }, + { + "epoch": 2.4687019547551063, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.664324998855591, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7394219040870667, + "num_tokens": 581570008.0, + "step": 22480 + }, + { + "epoch": 2.46881177245772, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.383660078048706, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.743227481842041, + "num_tokens": 581594675.0, + "step": 22481 + }, + { + "epoch": 2.468921590160334, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.243264675140381, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7302831411361694, + "num_tokens": 581621903.0, + "step": 22482 + }, + { + "epoch": 2.4690314078629476, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.194549798965454, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7252283096313477, + "num_tokens": 581653424.0, + "step": 22483 + }, + { + "epoch": 2.4691412255655614, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3881759643554688, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7300890684127808, + "num_tokens": 581680659.0, + "step": 22484 + }, + { + "epoch": 2.4692510432681747, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.326476812362671, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.725633978843689, + "num_tokens": 581706034.0, + "step": 22485 + }, + { + "epoch": 2.4693608609707884, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.098007917404175, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7207635641098022, + "num_tokens": 581736537.0, + "step": 22486 + }, + { + "epoch": 2.469470678673402, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.689047336578369, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7520665526390076, + "num_tokens": 581758212.0, + "step": 22487 + }, + { + "epoch": 2.469580496376016, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.7816684246063232, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7397243976593018, + "num_tokens": 581777728.0, + "step": 22488 + }, + { + "epoch": 2.4696903140786297, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3914613723754883, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.7601889967918396, + "num_tokens": 581802384.0, + "step": 22489 + }, + { + "epoch": 2.469800131781243, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.231677770614624, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7187952995300293, + "num_tokens": 581831897.0, + "step": 22490 + }, + { + "epoch": 2.469909949483857, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.7823007106781006, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.748525857925415, + "num_tokens": 581852383.0, + "step": 22491 + }, + { + "epoch": 2.4700197671864705, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4515674114227295, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7314134836196899, + "num_tokens": 581876337.0, + "step": 22492 + }, + { + "epoch": 2.4701295848890843, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.231326103210449, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7124295234680176, + "num_tokens": 581904782.0, + "step": 22493 + }, + { + "epoch": 2.4702394025916976, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.069544792175293, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.6917815208435059, + "num_tokens": 581937627.0, + "step": 22494 + }, + { + "epoch": 2.4703492202943114, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.436367988586426, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7284685969352722, + "num_tokens": 581961556.0, + "step": 22495 + }, + { + "epoch": 2.470459037996925, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.558213710784912, + "learning_rate": 1e-06, + "loss": 0.7638, + "mean_token_accuracy": 0.7553141117095947, + "num_tokens": 581982982.0, + "step": 22496 + }, + { + "epoch": 2.470568855699539, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.521898031234741, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7465896010398865, + "num_tokens": 582007050.0, + "step": 22497 + }, + { + "epoch": 2.470678673402152, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.184286117553711, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7370090484619141, + "num_tokens": 582037396.0, + "step": 22498 + }, + { + "epoch": 2.470788491104766, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5555522441864014, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.710323691368103, + "num_tokens": 582060656.0, + "step": 22499 + }, + { + "epoch": 2.4708983088073797, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.0076584815979004, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.6994792819023132, + "num_tokens": 582098147.0, + "step": 22500 + }, + { + "epoch": 2.4710081265099935, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.561500072479248, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7129288911819458, + "num_tokens": 582120838.0, + "step": 22501 + }, + { + "epoch": 2.471117944212607, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.600933790206909, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7312383651733398, + "num_tokens": 582142587.0, + "step": 22502 + }, + { + "epoch": 2.4712277619152205, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6854238510131836, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7234553694725037, + "num_tokens": 582164094.0, + "step": 22503 + }, + { + "epoch": 2.4713375796178343, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1621100902557373, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.721587061882019, + "num_tokens": 582191897.0, + "step": 22504 + }, + { + "epoch": 2.471447397320448, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4767444133758545, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7321984767913818, + "num_tokens": 582215152.0, + "step": 22505 + }, + { + "epoch": 2.471557215023062, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.7621212005615234, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7352206707000732, + "num_tokens": 582235363.0, + "step": 22506 + }, + { + "epoch": 2.4716670327256756, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.0539183616638184, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7267476320266724, + "num_tokens": 582268099.0, + "step": 22507 + }, + { + "epoch": 2.471776850428289, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6838486194610596, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7326469421386719, + "num_tokens": 582287577.0, + "step": 22508 + }, + { + "epoch": 2.4718866681309026, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2665891647338867, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.729953408241272, + "num_tokens": 582316791.0, + "step": 22509 + }, + { + "epoch": 2.4719964858335164, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.355023145675659, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7367838025093079, + "num_tokens": 582343571.0, + "step": 22510 + }, + { + "epoch": 2.47210630353613, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.365687370300293, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7275534868240356, + "num_tokens": 582370286.0, + "step": 22511 + }, + { + "epoch": 2.472216121238744, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1643264293670654, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7494317889213562, + "num_tokens": 582396706.0, + "step": 22512 + }, + { + "epoch": 2.472325938941357, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6697158813476562, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7228214740753174, + "num_tokens": 582419388.0, + "step": 22513 + }, + { + "epoch": 2.472435756643971, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1770410537719727, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7154431343078613, + "num_tokens": 582449128.0, + "step": 22514 + }, + { + "epoch": 2.4725455743465847, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1476335525512695, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7300500869750977, + "num_tokens": 582481366.0, + "step": 22515 + }, + { + "epoch": 2.4726553920491985, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.270390272140503, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.715753972530365, + "num_tokens": 582509599.0, + "step": 22516 + }, + { + "epoch": 2.472765209751812, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.43579363822937, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7250522375106812, + "num_tokens": 582537510.0, + "step": 22517 + }, + { + "epoch": 2.4728750274544256, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1696693897247314, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7289162278175354, + "num_tokens": 582565833.0, + "step": 22518 + }, + { + "epoch": 2.4729848451570393, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.1654136180877686, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7285416126251221, + "num_tokens": 582595069.0, + "step": 22519 + }, + { + "epoch": 2.473094662859653, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.196086883544922, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7217283248901367, + "num_tokens": 582626140.0, + "step": 22520 + }, + { + "epoch": 2.473204480562267, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.096965789794922, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7377249002456665, + "num_tokens": 582656988.0, + "step": 22521 + }, + { + "epoch": 2.47331429826488, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.104969024658203, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.7017790079116821, + "num_tokens": 582690171.0, + "step": 22522 + }, + { + "epoch": 2.473424115967494, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3806824684143066, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7469428777694702, + "num_tokens": 582714233.0, + "step": 22523 + }, + { + "epoch": 2.4735339336701077, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.362401008605957, + "learning_rate": 1e-06, + "loss": 0.7848, + "mean_token_accuracy": 0.7503584623336792, + "num_tokens": 582738901.0, + "step": 22524 + }, + { + "epoch": 2.4736437513727214, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.421372890472412, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7533460855484009, + "num_tokens": 582761168.0, + "step": 22525 + }, + { + "epoch": 2.4737535690753347, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3520872592926025, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7274494171142578, + "num_tokens": 582786137.0, + "step": 22526 + }, + { + "epoch": 2.4738633867779485, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.291639566421509, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7177356481552124, + "num_tokens": 582812937.0, + "step": 22527 + }, + { + "epoch": 2.4739732044805622, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.40277361869812, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7210594415664673, + "num_tokens": 582840990.0, + "step": 22528 + }, + { + "epoch": 2.474083022183176, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.645516872406006, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7441318035125732, + "num_tokens": 582862034.0, + "step": 22529 + }, + { + "epoch": 2.4741928398857898, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.462766647338867, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7273665070533752, + "num_tokens": 582887359.0, + "step": 22530 + }, + { + "epoch": 2.474302657588403, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3560352325439453, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7536508440971375, + "num_tokens": 582911724.0, + "step": 22531 + }, + { + "epoch": 2.474412475291017, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.194467067718506, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7327620983123779, + "num_tokens": 582940079.0, + "step": 22532 + }, + { + "epoch": 2.4745222929936306, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4269301891326904, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7280391454696655, + "num_tokens": 582966645.0, + "step": 22533 + }, + { + "epoch": 2.4746321106962443, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5242598056793213, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7298791408538818, + "num_tokens": 582990337.0, + "step": 22534 + }, + { + "epoch": 2.474741928398858, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2718722820281982, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7270671725273132, + "num_tokens": 583019007.0, + "step": 22535 + }, + { + "epoch": 2.4748517461014714, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.319286346435547, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7337044477462769, + "num_tokens": 583046525.0, + "step": 22536 + }, + { + "epoch": 2.474961563804085, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3059422969818115, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7119463086128235, + "num_tokens": 583075457.0, + "step": 22537 + }, + { + "epoch": 2.475071381506699, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5796828269958496, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7320990562438965, + "num_tokens": 583097935.0, + "step": 22538 + }, + { + "epoch": 2.4751811992093127, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.836653232574463, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7454128265380859, + "num_tokens": 583116913.0, + "step": 22539 + }, + { + "epoch": 2.4752910169119264, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 1.9688448905944824, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7435047626495361, + "num_tokens": 583152270.0, + "step": 22540 + }, + { + "epoch": 2.4754008346145397, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2318766117095947, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7201086282730103, + "num_tokens": 583181986.0, + "step": 22541 + }, + { + "epoch": 2.4755106523171535, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.0709497928619385, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7221996784210205, + "num_tokens": 583212767.0, + "step": 22542 + }, + { + "epoch": 2.4756204700197673, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.8622279167175293, + "learning_rate": 1e-06, + "loss": 0.7722, + "mean_token_accuracy": 0.7537611722946167, + "num_tokens": 583230748.0, + "step": 22543 + }, + { + "epoch": 2.475730287722381, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.677452802658081, + "learning_rate": 1e-06, + "loss": 0.789, + "mean_token_accuracy": 0.7558306455612183, + "num_tokens": 583250752.0, + "step": 22544 + }, + { + "epoch": 2.4758401054249943, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3834950923919678, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7058501243591309, + "num_tokens": 583278437.0, + "step": 22545 + }, + { + "epoch": 2.475949923127608, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2306578159332275, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7106833457946777, + "num_tokens": 583305175.0, + "step": 22546 + }, + { + "epoch": 2.476059740830222, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5776374340057373, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7283333539962769, + "num_tokens": 583327026.0, + "step": 22547 + }, + { + "epoch": 2.4761695585328356, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.197341203689575, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7212300300598145, + "num_tokens": 583356237.0, + "step": 22548 + }, + { + "epoch": 2.476279376235449, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.185218095779419, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7279684543609619, + "num_tokens": 583384805.0, + "step": 22549 + }, + { + "epoch": 2.4763891939380627, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.3794069290161133, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7391402721405029, + "num_tokens": 583409634.0, + "step": 22550 + }, + { + "epoch": 2.4764990116406764, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.147810697555542, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7100161910057068, + "num_tokens": 583439142.0, + "step": 22551 + }, + { + "epoch": 2.47660882934329, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5264880657196045, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7179564237594604, + "num_tokens": 583462208.0, + "step": 22552 + }, + { + "epoch": 2.476718647045904, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4133691787719727, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7380324602127075, + "num_tokens": 583485821.0, + "step": 22553 + }, + { + "epoch": 2.4768284647485173, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.227790355682373, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7095922231674194, + "num_tokens": 583516572.0, + "step": 22554 + }, + { + "epoch": 2.476938282451131, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.297262668609619, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7268794775009155, + "num_tokens": 583542215.0, + "step": 22555 + }, + { + "epoch": 2.4770481001537448, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.451988697052002, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.713204562664032, + "num_tokens": 583568246.0, + "step": 22556 + }, + { + "epoch": 2.4771579178563585, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4686670303344727, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7416418790817261, + "num_tokens": 583593046.0, + "step": 22557 + }, + { + "epoch": 2.4772677355589723, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4652044773101807, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7463972568511963, + "num_tokens": 583617182.0, + "step": 22558 + }, + { + "epoch": 2.4773775532615856, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5181167125701904, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7265490889549255, + "num_tokens": 583640409.0, + "step": 22559 + }, + { + "epoch": 2.4774873709641994, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.276538133621216, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7301385998725891, + "num_tokens": 583667501.0, + "step": 22560 + }, + { + "epoch": 2.477597188666813, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4911327362060547, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7337497472763062, + "num_tokens": 583691981.0, + "step": 22561 + }, + { + "epoch": 2.477707006369427, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.0965709686279297, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7267730236053467, + "num_tokens": 583723544.0, + "step": 22562 + }, + { + "epoch": 2.4778168240720406, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6491236686706543, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7351565361022949, + "num_tokens": 583745249.0, + "step": 22563 + }, + { + "epoch": 2.477926641774654, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.5021719932556152, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7450046539306641, + "num_tokens": 583767713.0, + "step": 22564 + }, + { + "epoch": 2.4780364594772677, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.2523391246795654, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7328774929046631, + "num_tokens": 583796351.0, + "step": 22565 + }, + { + "epoch": 2.4781462771798815, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6709070205688477, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7444342374801636, + "num_tokens": 583818522.0, + "step": 22566 + }, + { + "epoch": 2.478256094882495, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.706975221633911, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7468566298484802, + "num_tokens": 583841306.0, + "step": 22567 + }, + { + "epoch": 2.478365912585109, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.6278913021087646, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7572181224822998, + "num_tokens": 583862720.0, + "step": 22568 + }, + { + "epoch": 2.4784757302877223, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.287269115447998, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7002992630004883, + "num_tokens": 583891557.0, + "step": 22569 + }, + { + "epoch": 2.478585547990336, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.4289355278015137, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7308651208877563, + "num_tokens": 583915524.0, + "step": 22570 + }, + { + "epoch": 2.47869536569295, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.111083984375, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7374905943870544, + "num_tokens": 583947311.0, + "step": 22571 + }, + { + "epoch": 2.4788051833955635, + "ewc_loss": 0.0001964569091796875, + "grad_norm": 2.382317304611206, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7244985103607178, + "num_tokens": 583971538.0, + "step": 22572 + }, + { + "epoch": 2.478915001098177, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.795757532119751, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7474134564399719, + "num_tokens": 583990662.0, + "step": 22573 + }, + { + "epoch": 2.4790248188007906, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2493956089019775, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7242026925086975, + "num_tokens": 584020069.0, + "step": 22574 + }, + { + "epoch": 2.4791346365034044, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5184173583984375, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7486885786056519, + "num_tokens": 584043942.0, + "step": 22575 + }, + { + "epoch": 2.479244454206018, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.6933505535125732, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7366088628768921, + "num_tokens": 584064359.0, + "step": 22576 + }, + { + "epoch": 2.4793542719086314, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.413118600845337, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7148240804672241, + "num_tokens": 584089446.0, + "step": 22577 + }, + { + "epoch": 2.479464089611245, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.629025936126709, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7570881843566895, + "num_tokens": 584112276.0, + "step": 22578 + }, + { + "epoch": 2.479573907313859, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.903890609741211, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7185900211334229, + "num_tokens": 584131282.0, + "step": 22579 + }, + { + "epoch": 2.4796837250164727, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.449563503265381, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.723379909992218, + "num_tokens": 584157027.0, + "step": 22580 + }, + { + "epoch": 2.4797935427190865, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.218088388442993, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.700166642665863, + "num_tokens": 584187271.0, + "step": 22581 + }, + { + "epoch": 2.4799033604217, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.361485242843628, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7291497588157654, + "num_tokens": 584214126.0, + "step": 22582 + }, + { + "epoch": 2.4800131781243135, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.457434892654419, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7445626854896545, + "num_tokens": 584238308.0, + "step": 22583 + }, + { + "epoch": 2.4801229958269273, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.6113078594207764, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.70718914270401, + "num_tokens": 584260562.0, + "step": 22584 + }, + { + "epoch": 2.480232813529541, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.547494649887085, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7242785692214966, + "num_tokens": 584285600.0, + "step": 22585 + }, + { + "epoch": 2.480342631232155, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.442148447036743, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.735856294631958, + "num_tokens": 584309999.0, + "step": 22586 + }, + { + "epoch": 2.480452448934768, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.7304837703704834, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7424145340919495, + "num_tokens": 584331339.0, + "step": 22587 + }, + { + "epoch": 2.480562266637382, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.6399428844451904, + "learning_rate": 1e-06, + "loss": 0.7218, + "mean_token_accuracy": 0.7688853144645691, + "num_tokens": 584352018.0, + "step": 22588 + }, + { + "epoch": 2.4806720843399956, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.521066188812256, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.766690731048584, + "num_tokens": 584374314.0, + "step": 22589 + }, + { + "epoch": 2.4807819020426094, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.55965256690979, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7043857574462891, + "num_tokens": 584396902.0, + "step": 22590 + }, + { + "epoch": 2.480891719745223, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4979629516601562, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7440106272697449, + "num_tokens": 584419211.0, + "step": 22591 + }, + { + "epoch": 2.4810015374478365, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.643758535385132, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7291229367256165, + "num_tokens": 584442109.0, + "step": 22592 + }, + { + "epoch": 2.4811113551504502, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.562669038772583, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7381324768066406, + "num_tokens": 584466087.0, + "step": 22593 + }, + { + "epoch": 2.481221172853064, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4541397094726562, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7211794853210449, + "num_tokens": 584490736.0, + "step": 22594 + }, + { + "epoch": 2.4813309905556777, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.029827117919922, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7248291969299316, + "num_tokens": 584525952.0, + "step": 22595 + }, + { + "epoch": 2.481440808258291, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2527472972869873, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7251582145690918, + "num_tokens": 584554175.0, + "step": 22596 + }, + { + "epoch": 2.481550625960905, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.368476152420044, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7356405258178711, + "num_tokens": 584581818.0, + "step": 22597 + }, + { + "epoch": 2.4816604436635186, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4272782802581787, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.732678234577179, + "num_tokens": 584608837.0, + "step": 22598 + }, + { + "epoch": 2.4817702613661323, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.309019088745117, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7159990668296814, + "num_tokens": 584637585.0, + "step": 22599 + }, + { + "epoch": 2.4818800790687456, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.443422794342041, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7325795292854309, + "num_tokens": 584661031.0, + "step": 22600 + }, + { + "epoch": 2.4819898967713594, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3949966430664062, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.742357075214386, + "num_tokens": 584684536.0, + "step": 22601 + }, + { + "epoch": 2.482099714473973, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4357657432556152, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.744604766368866, + "num_tokens": 584708199.0, + "step": 22602 + }, + { + "epoch": 2.482209532176587, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.391704797744751, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7105510830879211, + "num_tokens": 584733161.0, + "step": 22603 + }, + { + "epoch": 2.4823193498792007, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.471968412399292, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7399543523788452, + "num_tokens": 584757816.0, + "step": 22604 + }, + { + "epoch": 2.482429167581814, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.1917810440063477, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7495667934417725, + "num_tokens": 584787031.0, + "step": 22605 + }, + { + "epoch": 2.4825389852844277, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.574613571166992, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7460277080535889, + "num_tokens": 584808485.0, + "step": 22606 + }, + { + "epoch": 2.4826488029870415, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.747934341430664, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7329254150390625, + "num_tokens": 584830143.0, + "step": 22607 + }, + { + "epoch": 2.4827586206896552, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4350225925445557, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7305768132209778, + "num_tokens": 584854495.0, + "step": 22608 + }, + { + "epoch": 2.482868438392269, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5440144538879395, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7300081253051758, + "num_tokens": 584877030.0, + "step": 22609 + }, + { + "epoch": 2.4829782560948823, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5785887241363525, + "learning_rate": 1e-06, + "loss": 0.756, + "mean_token_accuracy": 0.7584966421127319, + "num_tokens": 584899034.0, + "step": 22610 + }, + { + "epoch": 2.483088073797496, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.53149151802063, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.725918173789978, + "num_tokens": 584924205.0, + "step": 22611 + }, + { + "epoch": 2.48319789150011, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.416700839996338, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7106019258499146, + "num_tokens": 584949529.0, + "step": 22612 + }, + { + "epoch": 2.4833077092027236, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.651190757751465, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7362197637557983, + "num_tokens": 584972765.0, + "step": 22613 + }, + { + "epoch": 2.4834175269053373, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.1101951599121094, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7101421356201172, + "num_tokens": 585004051.0, + "step": 22614 + }, + { + "epoch": 2.4835273446079507, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.121741533279419, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.6984708309173584, + "num_tokens": 585035043.0, + "step": 22615 + }, + { + "epoch": 2.4836371623105644, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2780911922454834, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7350996732711792, + "num_tokens": 585063833.0, + "step": 22616 + }, + { + "epoch": 2.483746980013178, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4399375915527344, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7163793444633484, + "num_tokens": 585089772.0, + "step": 22617 + }, + { + "epoch": 2.483856797715792, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.348667860031128, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7183343768119812, + "num_tokens": 585113862.0, + "step": 22618 + }, + { + "epoch": 2.4839666154184057, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.238049268722534, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7247783541679382, + "num_tokens": 585141848.0, + "step": 22619 + }, + { + "epoch": 2.484076433121019, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.493084669113159, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7261984348297119, + "num_tokens": 585165372.0, + "step": 22620 + }, + { + "epoch": 2.4841862508236328, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.406111478805542, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7245460152626038, + "num_tokens": 585190186.0, + "step": 22621 + }, + { + "epoch": 2.4842960685262465, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4576525688171387, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7423646450042725, + "num_tokens": 585215818.0, + "step": 22622 + }, + { + "epoch": 2.4844058862288603, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.400603771209717, + "learning_rate": 1e-06, + "loss": 0.7812, + "mean_token_accuracy": 0.752106785774231, + "num_tokens": 585238265.0, + "step": 22623 + }, + { + "epoch": 2.4845157039314736, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.1717100143432617, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7425050139427185, + "num_tokens": 585266467.0, + "step": 22624 + }, + { + "epoch": 2.4846255216340873, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.242511510848999, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.743614912033081, + "num_tokens": 585295652.0, + "step": 22625 + }, + { + "epoch": 2.484735339336701, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2408573627471924, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7403393983840942, + "num_tokens": 585323200.0, + "step": 22626 + }, + { + "epoch": 2.484845157039315, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.418670892715454, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7342813014984131, + "num_tokens": 585347754.0, + "step": 22627 + }, + { + "epoch": 2.484954974741928, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.366579294204712, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7321640253067017, + "num_tokens": 585373003.0, + "step": 22628 + }, + { + "epoch": 2.485064792444542, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2827699184417725, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7227452397346497, + "num_tokens": 585399580.0, + "step": 22629 + }, + { + "epoch": 2.4851746101471557, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5405538082122803, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7401192784309387, + "num_tokens": 585423217.0, + "step": 22630 + }, + { + "epoch": 2.4852844278497694, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2543532848358154, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7270168662071228, + "num_tokens": 585451117.0, + "step": 22631 + }, + { + "epoch": 2.485394245552383, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.145002603530884, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7275339365005493, + "num_tokens": 585480424.0, + "step": 22632 + }, + { + "epoch": 2.4855040632549965, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2534501552581787, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7435128688812256, + "num_tokens": 585508401.0, + "step": 22633 + }, + { + "epoch": 2.4856138809576103, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.249417304992676, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7161539793014526, + "num_tokens": 585537066.0, + "step": 22634 + }, + { + "epoch": 2.485723698660224, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3388476371765137, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7221312522888184, + "num_tokens": 585564235.0, + "step": 22635 + }, + { + "epoch": 2.485833516362838, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.6935880184173584, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7386291027069092, + "num_tokens": 585585897.0, + "step": 22636 + }, + { + "epoch": 2.4859433340654515, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.175116777420044, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7275878190994263, + "num_tokens": 585613902.0, + "step": 22637 + }, + { + "epoch": 2.486053151768065, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.382143259048462, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7301771640777588, + "num_tokens": 585640191.0, + "step": 22638 + }, + { + "epoch": 2.4861629694706786, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3741610050201416, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7287947535514832, + "num_tokens": 585668203.0, + "step": 22639 + }, + { + "epoch": 2.4862727871732924, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5467143058776855, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7370020747184753, + "num_tokens": 585692082.0, + "step": 22640 + }, + { + "epoch": 2.486382604875906, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5120911598205566, + "learning_rate": 1e-06, + "loss": 0.7497, + "mean_token_accuracy": 0.7560620307922363, + "num_tokens": 585713766.0, + "step": 22641 + }, + { + "epoch": 2.48649242257852, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.369657278060913, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7474150657653809, + "num_tokens": 585739024.0, + "step": 22642 + }, + { + "epoch": 2.486602240281133, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2029643058776855, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7226874232292175, + "num_tokens": 585768397.0, + "step": 22643 + }, + { + "epoch": 2.486712057983747, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.567056655883789, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7299593687057495, + "num_tokens": 585790832.0, + "step": 22644 + }, + { + "epoch": 2.4868218756863607, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.381240129470825, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7193830013275146, + "num_tokens": 585816244.0, + "step": 22645 + }, + { + "epoch": 2.4869316933889745, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3521993160247803, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7442486882209778, + "num_tokens": 585841796.0, + "step": 22646 + }, + { + "epoch": 2.4870415110915878, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.322216033935547, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7071583271026611, + "num_tokens": 585868691.0, + "step": 22647 + }, + { + "epoch": 2.4871513287942015, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.239443063735962, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7246636152267456, + "num_tokens": 585896525.0, + "step": 22648 + }, + { + "epoch": 2.4872611464968153, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.384371280670166, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7292249798774719, + "num_tokens": 585922581.0, + "step": 22649 + }, + { + "epoch": 2.487370964199429, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2684123516082764, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7402738928794861, + "num_tokens": 585952214.0, + "step": 22650 + }, + { + "epoch": 2.487480781902043, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.154841899871826, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7314446568489075, + "num_tokens": 585981483.0, + "step": 22651 + }, + { + "epoch": 2.487590599604656, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.331467866897583, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7350627779960632, + "num_tokens": 586006401.0, + "step": 22652 + }, + { + "epoch": 2.48770041730727, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.165600299835205, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7168979644775391, + "num_tokens": 586035868.0, + "step": 22653 + }, + { + "epoch": 2.4878102350098836, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3360795974731445, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7196527123451233, + "num_tokens": 586063210.0, + "step": 22654 + }, + { + "epoch": 2.4879200527124974, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5047898292541504, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7455503940582275, + "num_tokens": 586088749.0, + "step": 22655 + }, + { + "epoch": 2.4880298704151107, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.300032615661621, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7365105748176575, + "num_tokens": 586114721.0, + "step": 22656 + }, + { + "epoch": 2.4881396881177245, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.402740001678467, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7418825030326843, + "num_tokens": 586138740.0, + "step": 22657 + }, + { + "epoch": 2.488249505820338, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4267306327819824, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7213733196258545, + "num_tokens": 586163806.0, + "step": 22658 + }, + { + "epoch": 2.488359323522952, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3843135833740234, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.713437020778656, + "num_tokens": 586189894.0, + "step": 22659 + }, + { + "epoch": 2.4884691412255657, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.037374258041382, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7238461971282959, + "num_tokens": 586218294.0, + "step": 22660 + }, + { + "epoch": 2.488578958928179, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.5065953731536865, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7363384366035461, + "num_tokens": 586241428.0, + "step": 22661 + }, + { + "epoch": 2.488688776630793, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.6898365020751953, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7475364208221436, + "num_tokens": 586259610.0, + "step": 22662 + }, + { + "epoch": 2.4887985943334066, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3757009506225586, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.716070294380188, + "num_tokens": 586287264.0, + "step": 22663 + }, + { + "epoch": 2.4889084120360203, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3487589359283447, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7430881261825562, + "num_tokens": 586315879.0, + "step": 22664 + }, + { + "epoch": 2.489018229738634, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.724822521209717, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7154643535614014, + "num_tokens": 586336976.0, + "step": 22665 + }, + { + "epoch": 2.4891280474412474, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.094348669052124, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7125518321990967, + "num_tokens": 586369199.0, + "step": 22666 + }, + { + "epoch": 2.489237865143861, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.9727401733398438, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7378499507904053, + "num_tokens": 586387568.0, + "step": 22667 + }, + { + "epoch": 2.489347682846475, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2556405067443848, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7210872173309326, + "num_tokens": 586415242.0, + "step": 22668 + }, + { + "epoch": 2.4894575005490887, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.312285900115967, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7332795858383179, + "num_tokens": 586441756.0, + "step": 22669 + }, + { + "epoch": 2.4895673182517024, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2027363777160645, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7378890514373779, + "num_tokens": 586468933.0, + "step": 22670 + }, + { + "epoch": 2.4896771359543157, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.433891773223877, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7179911136627197, + "num_tokens": 586495660.0, + "step": 22671 + }, + { + "epoch": 2.4897869536569295, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3436808586120605, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7196602821350098, + "num_tokens": 586522449.0, + "step": 22672 + }, + { + "epoch": 2.4898967713595432, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.209259510040283, + "learning_rate": 1e-06, + "loss": 0.78, + "mean_token_accuracy": 0.7622210383415222, + "num_tokens": 586549179.0, + "step": 22673 + }, + { + "epoch": 2.490006589062157, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3867902755737305, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7323204874992371, + "num_tokens": 586574623.0, + "step": 22674 + }, + { + "epoch": 2.4901164067647703, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.492635488510132, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7197960615158081, + "num_tokens": 586600188.0, + "step": 22675 + }, + { + "epoch": 2.490226224467384, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.526566743850708, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7295949459075928, + "num_tokens": 586623978.0, + "step": 22676 + }, + { + "epoch": 2.490336042169998, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.544189214706421, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7205703854560852, + "num_tokens": 586647401.0, + "step": 22677 + }, + { + "epoch": 2.4904458598726116, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.769399404525757, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7286290526390076, + "num_tokens": 586668775.0, + "step": 22678 + }, + { + "epoch": 2.490555677575225, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.4363765716552734, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7360191941261292, + "num_tokens": 586694421.0, + "step": 22679 + }, + { + "epoch": 2.4906654952778386, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2831079959869385, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7234059572219849, + "num_tokens": 586721979.0, + "step": 22680 + }, + { + "epoch": 2.4907753129804524, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3334901332855225, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6905018091201782, + "num_tokens": 586752442.0, + "step": 22681 + }, + { + "epoch": 2.490885130683066, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3331382274627686, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7070258855819702, + "num_tokens": 586780628.0, + "step": 22682 + }, + { + "epoch": 2.49099494838568, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.134681463241577, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7464360594749451, + "num_tokens": 586809136.0, + "step": 22683 + }, + { + "epoch": 2.4911047660882932, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.151353359222412, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7275351285934448, + "num_tokens": 586838550.0, + "step": 22684 + }, + { + "epoch": 2.491214583790907, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.0785391330718994, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6937546133995056, + "num_tokens": 586872032.0, + "step": 22685 + }, + { + "epoch": 2.4913244014935207, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.1058859825134277, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7193760275840759, + "num_tokens": 586906818.0, + "step": 22686 + }, + { + "epoch": 2.4914342191961345, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.737349033355713, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.739824652671814, + "num_tokens": 586928153.0, + "step": 22687 + }, + { + "epoch": 2.4915440368987483, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.3025364875793457, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7202270030975342, + "num_tokens": 586954338.0, + "step": 22688 + }, + { + "epoch": 2.4916538546013616, + "ewc_loss": 0.0001983642578125, + "grad_norm": 2.2786452770233154, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.72389155626297, + "num_tokens": 586981709.0, + "step": 22689 + }, + { + "epoch": 2.4917636723039753, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2881832122802734, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7452484369277954, + "num_tokens": 587009605.0, + "step": 22690 + }, + { + "epoch": 2.491873490006589, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.483505964279175, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7463050484657288, + "num_tokens": 587032841.0, + "step": 22691 + }, + { + "epoch": 2.491983307709203, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.5985426902770996, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7308942079544067, + "num_tokens": 587056958.0, + "step": 22692 + }, + { + "epoch": 2.4920931254118166, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3117239475250244, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7335758209228516, + "num_tokens": 587083782.0, + "step": 22693 + }, + { + "epoch": 2.49220294311443, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.403352737426758, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7092682719230652, + "num_tokens": 587110823.0, + "step": 22694 + }, + { + "epoch": 2.4923127608170437, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.141885757446289, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7128667235374451, + "num_tokens": 587140248.0, + "step": 22695 + }, + { + "epoch": 2.4924225785196574, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.309978723526001, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7059870958328247, + "num_tokens": 587169688.0, + "step": 22696 + }, + { + "epoch": 2.492532396222271, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.544600248336792, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7417200207710266, + "num_tokens": 587192285.0, + "step": 22697 + }, + { + "epoch": 2.4926422139248845, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.524958848953247, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7426764965057373, + "num_tokens": 587213982.0, + "step": 22698 + }, + { + "epoch": 2.4927520316274983, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.252432107925415, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7321799993515015, + "num_tokens": 587241666.0, + "step": 22699 + }, + { + "epoch": 2.492861849330112, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.405858278274536, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7390475869178772, + "num_tokens": 587265561.0, + "step": 22700 + }, + { + "epoch": 2.4929716670327258, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2769594192504883, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7107736468315125, + "num_tokens": 587295396.0, + "step": 22701 + }, + { + "epoch": 2.4930814847353395, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.356048822402954, + "learning_rate": 1e-06, + "loss": 0.781, + "mean_token_accuracy": 0.7523127794265747, + "num_tokens": 587321031.0, + "step": 22702 + }, + { + "epoch": 2.493191302437953, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.9009811878204346, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.742268443107605, + "num_tokens": 587339850.0, + "step": 22703 + }, + { + "epoch": 2.4933011201405666, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.114985704421997, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7016521692276001, + "num_tokens": 587369105.0, + "step": 22704 + }, + { + "epoch": 2.4934109378431804, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3635318279266357, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7123672366142273, + "num_tokens": 587397579.0, + "step": 22705 + }, + { + "epoch": 2.493520755545794, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.4085307121276855, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7323823571205139, + "num_tokens": 587423378.0, + "step": 22706 + }, + { + "epoch": 2.4936305732484074, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.5530636310577393, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7328191995620728, + "num_tokens": 587446342.0, + "step": 22707 + }, + { + "epoch": 2.493740390951021, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.110403299331665, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7231909036636353, + "num_tokens": 587476983.0, + "step": 22708 + }, + { + "epoch": 2.493850208653635, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.813462972640991, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7330218553543091, + "num_tokens": 587501236.0, + "step": 22709 + }, + { + "epoch": 2.4939600263562487, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.7604424953460693, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7345953583717346, + "num_tokens": 587521020.0, + "step": 22710 + }, + { + "epoch": 2.4940698440588625, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.277470350265503, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7300096750259399, + "num_tokens": 587547275.0, + "step": 22711 + }, + { + "epoch": 2.4941796617614758, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.43178391456604, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.740354597568512, + "num_tokens": 587573084.0, + "step": 22712 + }, + { + "epoch": 2.4942894794640895, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.193953037261963, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7207769155502319, + "num_tokens": 587603257.0, + "step": 22713 + }, + { + "epoch": 2.4943992971667033, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3052866458892822, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7234610319137573, + "num_tokens": 587630187.0, + "step": 22714 + }, + { + "epoch": 2.494509114869317, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.8141117095947266, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7343208193778992, + "num_tokens": 587649842.0, + "step": 22715 + }, + { + "epoch": 2.494618932571931, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3563361167907715, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7262368202209473, + "num_tokens": 587675550.0, + "step": 22716 + }, + { + "epoch": 2.494728750274544, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.469625949859619, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7358759045600891, + "num_tokens": 587698137.0, + "step": 22717 + }, + { + "epoch": 2.494838567977158, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.325368642807007, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7341529130935669, + "num_tokens": 587726673.0, + "step": 22718 + }, + { + "epoch": 2.4949483856797716, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.132211208343506, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7320061326026917, + "num_tokens": 587756002.0, + "step": 22719 + }, + { + "epoch": 2.4950582033823854, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3248531818389893, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7166948914527893, + "num_tokens": 587784302.0, + "step": 22720 + }, + { + "epoch": 2.495168021084999, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2197136878967285, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7265125513076782, + "num_tokens": 587813162.0, + "step": 22721 + }, + { + "epoch": 2.4952778387876124, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.1307101249694824, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7422297596931458, + "num_tokens": 587844242.0, + "step": 22722 + }, + { + "epoch": 2.495387656490226, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.22129225730896, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.716443657875061, + "num_tokens": 587872383.0, + "step": 22723 + }, + { + "epoch": 2.49549747419284, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.480414390563965, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7277029156684875, + "num_tokens": 587895424.0, + "step": 22724 + }, + { + "epoch": 2.4956072918954537, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.0716259479522705, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7278927564620972, + "num_tokens": 587929377.0, + "step": 22725 + }, + { + "epoch": 2.495717109598067, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.551976442337036, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.73744797706604, + "num_tokens": 587952057.0, + "step": 22726 + }, + { + "epoch": 2.495826927300681, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.496589422225952, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7531812787055969, + "num_tokens": 587973851.0, + "step": 22727 + }, + { + "epoch": 2.4959367450032945, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.4001119136810303, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7339529991149902, + "num_tokens": 588000156.0, + "step": 22728 + }, + { + "epoch": 2.4960465627059083, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.208796739578247, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7313181161880493, + "num_tokens": 588027864.0, + "step": 22729 + }, + { + "epoch": 2.4961563804085216, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.450676441192627, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7050251960754395, + "num_tokens": 588053574.0, + "step": 22730 + }, + { + "epoch": 2.4962661981111354, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.252413034439087, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7412509918212891, + "num_tokens": 588083717.0, + "step": 22731 + }, + { + "epoch": 2.496376015813749, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.6871893405914307, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7334672808647156, + "num_tokens": 588106376.0, + "step": 22732 + }, + { + "epoch": 2.496485833516363, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3431649208068848, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7326671481132507, + "num_tokens": 588134677.0, + "step": 22733 + }, + { + "epoch": 2.4965956512189766, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.1487109661102295, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.719611644744873, + "num_tokens": 588163054.0, + "step": 22734 + }, + { + "epoch": 2.49670546892159, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.1729302406311035, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7243283987045288, + "num_tokens": 588192624.0, + "step": 22735 + }, + { + "epoch": 2.4968152866242037, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.4356794357299805, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7275488376617432, + "num_tokens": 588217520.0, + "step": 22736 + }, + { + "epoch": 2.4969251043268175, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.2829363346099854, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7172618508338928, + "num_tokens": 588246847.0, + "step": 22737 + }, + { + "epoch": 2.4970349220294312, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.555692672729492, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7259700298309326, + "num_tokens": 588271089.0, + "step": 22738 + }, + { + "epoch": 2.497144739732045, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.3714890480041504, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7275082468986511, + "num_tokens": 588296509.0, + "step": 22739 + }, + { + "epoch": 2.4972545574346583, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.580232858657837, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7271605134010315, + "num_tokens": 588320644.0, + "step": 22740 + }, + { + "epoch": 2.497364375137272, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.116353988647461, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7362965941429138, + "num_tokens": 588351365.0, + "step": 22741 + }, + { + "epoch": 2.497474192839886, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.434457778930664, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7366213202476501, + "num_tokens": 588374301.0, + "step": 22742 + }, + { + "epoch": 2.4975840105424996, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.6318283081054688, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7352228164672852, + "num_tokens": 588394979.0, + "step": 22743 + }, + { + "epoch": 2.4976938282451133, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.5653345584869385, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7267978191375732, + "num_tokens": 588418678.0, + "step": 22744 + }, + { + "epoch": 2.4978036459477266, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.43082332611084, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7238240838050842, + "num_tokens": 588442045.0, + "step": 22745 + }, + { + "epoch": 2.4979134636503404, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.608717918395996, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.736312210559845, + "num_tokens": 588464154.0, + "step": 22746 + }, + { + "epoch": 2.498023281352954, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.328589677810669, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7438828349113464, + "num_tokens": 588490177.0, + "step": 22747 + }, + { + "epoch": 2.498133099055568, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.4428625106811523, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7206762433052063, + "num_tokens": 588514830.0, + "step": 22748 + }, + { + "epoch": 2.4982429167581817, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.8847644329071045, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.745279848575592, + "num_tokens": 588533952.0, + "step": 22749 + }, + { + "epoch": 2.498352734460795, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.808044910430908, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7305822372436523, + "num_tokens": 588553512.0, + "step": 22750 + }, + { + "epoch": 2.4984625521634087, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.9812259674072266, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.7627835273742676, + "num_tokens": 588570616.0, + "step": 22751 + }, + { + "epoch": 2.4985723698660225, + "ewc_loss": 0.00019931793212890625, + "grad_norm": 2.4460525512695312, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7207756042480469, + "num_tokens": 588595651.0, + "step": 22752 + }, + { + "epoch": 2.4986821875686362, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.634854793548584, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7473372220993042, + "num_tokens": 588617392.0, + "step": 22753 + }, + { + "epoch": 2.4987920052712496, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.114928722381592, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6982500553131104, + "num_tokens": 588649690.0, + "step": 22754 + }, + { + "epoch": 2.4989018229738633, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2553369998931885, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7200151681900024, + "num_tokens": 588680395.0, + "step": 22755 + }, + { + "epoch": 2.499011640676477, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 1.9807766675949097, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7130601406097412, + "num_tokens": 588715486.0, + "step": 22756 + }, + { + "epoch": 2.499121458379091, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4807755947113037, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7213582992553711, + "num_tokens": 588741555.0, + "step": 22757 + }, + { + "epoch": 2.499231276081704, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.127269983291626, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7321470975875854, + "num_tokens": 588771413.0, + "step": 22758 + }, + { + "epoch": 2.499341093784318, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.610807418823242, + "learning_rate": 1e-06, + "loss": 0.7729, + "mean_token_accuracy": 0.7550597190856934, + "num_tokens": 588791492.0, + "step": 22759 + }, + { + "epoch": 2.4994509114869317, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.9779305458068848, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7382901906967163, + "num_tokens": 588811461.0, + "step": 22760 + }, + { + "epoch": 2.4995607291895454, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3774547576904297, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7185086607933044, + "num_tokens": 588840279.0, + "step": 22761 + }, + { + "epoch": 2.499670546892159, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.348133087158203, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7329712510108948, + "num_tokens": 588866667.0, + "step": 22762 + }, + { + "epoch": 2.4997803645947725, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.253783702850342, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7235251665115356, + "num_tokens": 588894255.0, + "step": 22763 + }, + { + "epoch": 2.4998901822973862, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3964285850524902, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7285119295120239, + "num_tokens": 588921334.0, + "step": 22764 + }, + { + "epoch": 2.5, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.656325101852417, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7248872518539429, + "num_tokens": 588942704.0, + "step": 22765 + }, + { + "epoch": 2.5001098177026138, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.210134983062744, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7186334729194641, + "num_tokens": 588971823.0, + "step": 22766 + }, + { + "epoch": 2.5002196354052275, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7021994590759277, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7443854808807373, + "num_tokens": 588993280.0, + "step": 22767 + }, + { + "epoch": 2.500329453107841, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.548738718032837, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7160770893096924, + "num_tokens": 589016288.0, + "step": 22768 + }, + { + "epoch": 2.5004392708104546, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.343994379043579, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7334509491920471, + "num_tokens": 589042750.0, + "step": 22769 + }, + { + "epoch": 2.5005490885130683, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.437535524368286, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7159531712532043, + "num_tokens": 589068047.0, + "step": 22770 + }, + { + "epoch": 2.500658906215682, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.491434097290039, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7379226684570312, + "num_tokens": 589090217.0, + "step": 22771 + }, + { + "epoch": 2.500768723918296, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2496097087860107, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.726938784122467, + "num_tokens": 589120073.0, + "step": 22772 + }, + { + "epoch": 2.500878541620909, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.397814989089966, + "learning_rate": 1e-06, + "loss": 0.7681, + "mean_token_accuracy": 0.7533888816833496, + "num_tokens": 589144251.0, + "step": 22773 + }, + { + "epoch": 2.500988359323523, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4243619441986084, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7298744916915894, + "num_tokens": 589168790.0, + "step": 22774 + }, + { + "epoch": 2.5010981770261367, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3586716651916504, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7392557263374329, + "num_tokens": 589193331.0, + "step": 22775 + }, + { + "epoch": 2.5012079947287504, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.282749891281128, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7277576923370361, + "num_tokens": 589221114.0, + "step": 22776 + }, + { + "epoch": 2.501317812431364, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2921390533447266, + "learning_rate": 1e-06, + "loss": 0.743, + "mean_token_accuracy": 0.7609174847602844, + "num_tokens": 589246004.0, + "step": 22777 + }, + { + "epoch": 2.5014276301339775, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.938849687576294, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7333629131317139, + "num_tokens": 589265631.0, + "step": 22778 + }, + { + "epoch": 2.5015374478365913, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5019302368164062, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7545636892318726, + "num_tokens": 589287728.0, + "step": 22779 + }, + { + "epoch": 2.501647265539205, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.318976640701294, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7298969030380249, + "num_tokens": 589313655.0, + "step": 22780 + }, + { + "epoch": 2.5017570832418183, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4743549823760986, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7319265007972717, + "num_tokens": 589337141.0, + "step": 22781 + }, + { + "epoch": 2.501866900944432, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.473398208618164, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7318878173828125, + "num_tokens": 589359967.0, + "step": 22782 + }, + { + "epoch": 2.501976718647046, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.184115409851074, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7559373378753662, + "num_tokens": 589387337.0, + "step": 22783 + }, + { + "epoch": 2.5020865363496596, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.351146936416626, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7274989485740662, + "num_tokens": 589414927.0, + "step": 22784 + }, + { + "epoch": 2.5021963540522734, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7357804775238037, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7413138151168823, + "num_tokens": 589436169.0, + "step": 22785 + }, + { + "epoch": 2.5023061717548867, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1870577335357666, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7247556447982788, + "num_tokens": 589464166.0, + "step": 22786 + }, + { + "epoch": 2.5024159894575004, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1740856170654297, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7124002575874329, + "num_tokens": 589495499.0, + "step": 22787 + }, + { + "epoch": 2.502525807160114, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.324131727218628, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7241922616958618, + "num_tokens": 589522298.0, + "step": 22788 + }, + { + "epoch": 2.502635624862728, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7927868366241455, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7401776313781738, + "num_tokens": 589542211.0, + "step": 22789 + }, + { + "epoch": 2.5027454425653417, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.850550889968872, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7132164239883423, + "num_tokens": 589563880.0, + "step": 22790 + }, + { + "epoch": 2.502855260267955, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2720706462860107, + "learning_rate": 1e-06, + "loss": 0.7084, + "mean_token_accuracy": 0.7718561887741089, + "num_tokens": 589590574.0, + "step": 22791 + }, + { + "epoch": 2.5029650779705688, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2738664150238037, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7369323968887329, + "num_tokens": 589619617.0, + "step": 22792 + }, + { + "epoch": 2.5030748956731825, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.721116065979004, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7214816808700562, + "num_tokens": 589640033.0, + "step": 22793 + }, + { + "epoch": 2.5031847133757963, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.293400764465332, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7439870834350586, + "num_tokens": 589665609.0, + "step": 22794 + }, + { + "epoch": 2.50329453107841, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5791878700256348, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7389878630638123, + "num_tokens": 589687734.0, + "step": 22795 + }, + { + "epoch": 2.5034043487810234, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6167497634887695, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7595911622047424, + "num_tokens": 589707580.0, + "step": 22796 + }, + { + "epoch": 2.503514166483637, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3034770488739014, + "learning_rate": 1e-06, + "loss": 0.7544, + "mean_token_accuracy": 0.7630774974822998, + "num_tokens": 589734307.0, + "step": 22797 + }, + { + "epoch": 2.503623984186251, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6129679679870605, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7364116907119751, + "num_tokens": 589757196.0, + "step": 22798 + }, + { + "epoch": 2.5037338018888646, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3821215629577637, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7418181896209717, + "num_tokens": 589780928.0, + "step": 22799 + }, + { + "epoch": 2.5038436195914784, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.245260000228882, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7371737957000732, + "num_tokens": 589810612.0, + "step": 22800 + }, + { + "epoch": 2.5039534372940917, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.699963331222534, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7274605631828308, + "num_tokens": 589831931.0, + "step": 22801 + }, + { + "epoch": 2.5040632549967055, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.198730945587158, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7238521575927734, + "num_tokens": 589861413.0, + "step": 22802 + }, + { + "epoch": 2.504173072699319, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1591877937316895, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7247937917709351, + "num_tokens": 589891007.0, + "step": 22803 + }, + { + "epoch": 2.5042828904019325, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.242426633834839, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7233175039291382, + "num_tokens": 589918620.0, + "step": 22804 + }, + { + "epoch": 2.5043927081045467, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.313722848892212, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7302293181419373, + "num_tokens": 589946586.0, + "step": 22805 + }, + { + "epoch": 2.50450252580716, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2340643405914307, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7389981746673584, + "num_tokens": 589974246.0, + "step": 22806 + }, + { + "epoch": 2.504612343509774, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2716867923736572, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7245141267776489, + "num_tokens": 590002770.0, + "step": 22807 + }, + { + "epoch": 2.5047221612123876, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4706904888153076, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7277938723564148, + "num_tokens": 590027131.0, + "step": 22808 + }, + { + "epoch": 2.504831978915001, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.8527629375457764, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7337095141410828, + "num_tokens": 590046283.0, + "step": 22809 + }, + { + "epoch": 2.5049417966176146, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.0619871616363525, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7054842114448547, + "num_tokens": 590076302.0, + "step": 22810 + }, + { + "epoch": 2.5050516143202284, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.290682077407837, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.715053915977478, + "num_tokens": 590102808.0, + "step": 22811 + }, + { + "epoch": 2.505161432022842, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.337679624557495, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.726425290107727, + "num_tokens": 590128825.0, + "step": 22812 + }, + { + "epoch": 2.505271249725456, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3151073455810547, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7442254424095154, + "num_tokens": 590153176.0, + "step": 22813 + }, + { + "epoch": 2.505381067428069, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.365957498550415, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7302742004394531, + "num_tokens": 590179251.0, + "step": 22814 + }, + { + "epoch": 2.505490885130683, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5735421180725098, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.738728404045105, + "num_tokens": 590199334.0, + "step": 22815 + }, + { + "epoch": 2.5056007028332967, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.120488405227661, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7275658249855042, + "num_tokens": 590230521.0, + "step": 22816 + }, + { + "epoch": 2.5057105205359105, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3885886669158936, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.717893123626709, + "num_tokens": 590256396.0, + "step": 22817 + }, + { + "epoch": 2.5058203382385242, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.461301326751709, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.731039822101593, + "num_tokens": 590280641.0, + "step": 22818 + }, + { + "epoch": 2.5059301559411375, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3860857486724854, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7168500423431396, + "num_tokens": 590307933.0, + "step": 22819 + }, + { + "epoch": 2.5060399736437513, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5894875526428223, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7197266221046448, + "num_tokens": 590328960.0, + "step": 22820 + }, + { + "epoch": 2.506149791346365, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4117536544799805, + "learning_rate": 1e-06, + "loss": 0.7826, + "mean_token_accuracy": 0.7482383251190186, + "num_tokens": 590350891.0, + "step": 22821 + }, + { + "epoch": 2.506259609048979, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.668858766555786, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7451242208480835, + "num_tokens": 590371449.0, + "step": 22822 + }, + { + "epoch": 2.5063694267515926, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.605642557144165, + "learning_rate": 1e-06, + "loss": 0.7332, + "mean_token_accuracy": 0.7642351984977722, + "num_tokens": 590393247.0, + "step": 22823 + }, + { + "epoch": 2.506479244454206, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.561685085296631, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7326786518096924, + "num_tokens": 590416572.0, + "step": 22824 + }, + { + "epoch": 2.5065890621568196, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4743595123291016, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7434104681015015, + "num_tokens": 590442330.0, + "step": 22825 + }, + { + "epoch": 2.5066988798594334, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2321994304656982, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7098439335823059, + "num_tokens": 590474719.0, + "step": 22826 + }, + { + "epoch": 2.506808697562047, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.110820770263672, + "learning_rate": 1e-06, + "loss": 0.7325, + "mean_token_accuracy": 0.7681128978729248, + "num_tokens": 590502613.0, + "step": 22827 + }, + { + "epoch": 2.506918515264661, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3348121643066406, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7341417670249939, + "num_tokens": 590528490.0, + "step": 22828 + }, + { + "epoch": 2.5070283329672742, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.321187973022461, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7084128260612488, + "num_tokens": 590556084.0, + "step": 22829 + }, + { + "epoch": 2.507138150669888, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6054654121398926, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.728331446647644, + "num_tokens": 590579798.0, + "step": 22830 + }, + { + "epoch": 2.5072479683725017, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6066107749938965, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7352961301803589, + "num_tokens": 590601789.0, + "step": 22831 + }, + { + "epoch": 2.507357786075115, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.607330560684204, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7317507266998291, + "num_tokens": 590626173.0, + "step": 22832 + }, + { + "epoch": 2.507467603777729, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4000837802886963, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7233409881591797, + "num_tokens": 590653382.0, + "step": 22833 + }, + { + "epoch": 2.5075774214803426, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.264721393585205, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7523081302642822, + "num_tokens": 590681835.0, + "step": 22834 + }, + { + "epoch": 2.5076872391829563, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.585253953933716, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.742516815662384, + "num_tokens": 590703115.0, + "step": 22835 + }, + { + "epoch": 2.50779705688557, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.287912607192993, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7414804100990295, + "num_tokens": 590729143.0, + "step": 22836 + }, + { + "epoch": 2.5079068745881834, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.263526678085327, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7146756649017334, + "num_tokens": 590757784.0, + "step": 22837 + }, + { + "epoch": 2.508016692290797, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.9393436908721924, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.7460930943489075, + "num_tokens": 590777166.0, + "step": 22838 + }, + { + "epoch": 2.508126509993411, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.46824312210083, + "learning_rate": 1e-06, + "loss": 0.7615, + "mean_token_accuracy": 0.7593833804130554, + "num_tokens": 590801084.0, + "step": 22839 + }, + { + "epoch": 2.5082363276960247, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.608624219894409, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7357993721961975, + "num_tokens": 590823667.0, + "step": 22840 + }, + { + "epoch": 2.5083461453986384, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.24294376373291, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7075666785240173, + "num_tokens": 590853423.0, + "step": 22841 + }, + { + "epoch": 2.5084559631012517, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6432785987854004, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7335958480834961, + "num_tokens": 590876373.0, + "step": 22842 + }, + { + "epoch": 2.5085657808038655, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4413177967071533, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7294358015060425, + "num_tokens": 590900751.0, + "step": 22843 + }, + { + "epoch": 2.5086755985064793, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2281713485717773, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7213759422302246, + "num_tokens": 590929624.0, + "step": 22844 + }, + { + "epoch": 2.508785416209093, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2587108612060547, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7267049551010132, + "num_tokens": 590957559.0, + "step": 22845 + }, + { + "epoch": 2.5088952339117068, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7856130599975586, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7371623516082764, + "num_tokens": 590977189.0, + "step": 22846 + }, + { + "epoch": 2.50900505161432, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2628672122955322, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.736582338809967, + "num_tokens": 591006782.0, + "step": 22847 + }, + { + "epoch": 2.509114869316934, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4094901084899902, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7409043312072754, + "num_tokens": 591032821.0, + "step": 22848 + }, + { + "epoch": 2.5092246870195476, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.332103729248047, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.739377498626709, + "num_tokens": 591063078.0, + "step": 22849 + }, + { + "epoch": 2.5093345047221614, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2321648597717285, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7244633436203003, + "num_tokens": 591093057.0, + "step": 22850 + }, + { + "epoch": 2.509444322424775, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3660595417022705, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7488876581192017, + "num_tokens": 591117707.0, + "step": 22851 + }, + { + "epoch": 2.5095541401273884, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7412655353546143, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7382901906967163, + "num_tokens": 591139261.0, + "step": 22852 + }, + { + "epoch": 2.509663957830002, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5248780250549316, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7342053651809692, + "num_tokens": 591163001.0, + "step": 22853 + }, + { + "epoch": 2.509773775532616, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.212498903274536, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7371631860733032, + "num_tokens": 591191088.0, + "step": 22854 + }, + { + "epoch": 2.5098835932352292, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.390134572982788, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7062836289405823, + "num_tokens": 591217394.0, + "step": 22855 + }, + { + "epoch": 2.5099934109378434, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.7776801586151123, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.744037389755249, + "num_tokens": 591238614.0, + "step": 22856 + }, + { + "epoch": 2.5101032286404568, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.471073627471924, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7251855731010437, + "num_tokens": 591264604.0, + "step": 22857 + }, + { + "epoch": 2.5102130463430705, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5821824073791504, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7521213293075562, + "num_tokens": 591287228.0, + "step": 22858 + }, + { + "epoch": 2.5103228640456843, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3538882732391357, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7223904132843018, + "num_tokens": 591312753.0, + "step": 22859 + }, + { + "epoch": 2.5104326817482976, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4522767066955566, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7293097972869873, + "num_tokens": 591338562.0, + "step": 22860 + }, + { + "epoch": 2.5105424994509113, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2490689754486084, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7029356956481934, + "num_tokens": 591366615.0, + "step": 22861 + }, + { + "epoch": 2.510652317153525, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.481797933578491, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.723833441734314, + "num_tokens": 591390433.0, + "step": 22862 + }, + { + "epoch": 2.510762134856139, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5000643730163574, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7190588712692261, + "num_tokens": 591412696.0, + "step": 22863 + }, + { + "epoch": 2.5108719525587526, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2762441635131836, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7337685823440552, + "num_tokens": 591439559.0, + "step": 22864 + }, + { + "epoch": 2.510981770261366, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5978994369506836, + "learning_rate": 1e-06, + "loss": 0.7322, + "mean_token_accuracy": 0.765331506729126, + "num_tokens": 591459784.0, + "step": 22865 + }, + { + "epoch": 2.5110915879639797, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.459045886993408, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7351651191711426, + "num_tokens": 591483730.0, + "step": 22866 + }, + { + "epoch": 2.5112014056665934, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.472641706466675, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7333885431289673, + "num_tokens": 591509318.0, + "step": 22867 + }, + { + "epoch": 2.511311223369207, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5939855575561523, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7340055108070374, + "num_tokens": 591532774.0, + "step": 22868 + }, + { + "epoch": 2.511421041071821, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6678805351257324, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7246288061141968, + "num_tokens": 591553811.0, + "step": 22869 + }, + { + "epoch": 2.5115308587744343, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.023388385772705, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7146987318992615, + "num_tokens": 591587246.0, + "step": 22870 + }, + { + "epoch": 2.511640676477048, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3075051307678223, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7051839828491211, + "num_tokens": 591616083.0, + "step": 22871 + }, + { + "epoch": 2.511750494179662, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2207396030426025, + "learning_rate": 1e-06, + "loss": 0.7728, + "mean_token_accuracy": 0.7555088996887207, + "num_tokens": 591642943.0, + "step": 22872 + }, + { + "epoch": 2.5118603118822755, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.547783136367798, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7498312592506409, + "num_tokens": 591668235.0, + "step": 22873 + }, + { + "epoch": 2.5119701295848893, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.224133253097534, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7261019349098206, + "num_tokens": 591696854.0, + "step": 22874 + }, + { + "epoch": 2.5120799472875026, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.400764226913452, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7229742407798767, + "num_tokens": 591722220.0, + "step": 22875 + }, + { + "epoch": 2.5121897649901164, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3186511993408203, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7163056135177612, + "num_tokens": 591749674.0, + "step": 22876 + }, + { + "epoch": 2.51229958269273, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3210690021514893, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7551518082618713, + "num_tokens": 591777655.0, + "step": 22877 + }, + { + "epoch": 2.512409400395344, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.0921642780303955, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7272924780845642, + "num_tokens": 591810654.0, + "step": 22878 + }, + { + "epoch": 2.5125192180979576, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.29384446144104, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7312745451927185, + "num_tokens": 591838576.0, + "step": 22879 + }, + { + "epoch": 2.512629035800571, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4323818683624268, + "learning_rate": 1e-06, + "loss": 0.7551, + "mean_token_accuracy": 0.7581812143325806, + "num_tokens": 591860586.0, + "step": 22880 + }, + { + "epoch": 2.5127388535031847, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5169267654418945, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7371777892112732, + "num_tokens": 591885342.0, + "step": 22881 + }, + { + "epoch": 2.5128486712057985, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6161210536956787, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7430286407470703, + "num_tokens": 591907658.0, + "step": 22882 + }, + { + "epoch": 2.512958488908412, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4038503170013428, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7297220230102539, + "num_tokens": 591933944.0, + "step": 22883 + }, + { + "epoch": 2.5130683066110255, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1329355239868164, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7218778729438782, + "num_tokens": 591965979.0, + "step": 22884 + }, + { + "epoch": 2.5131781243136393, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.778533697128296, + "learning_rate": 1e-06, + "loss": 0.7711, + "mean_token_accuracy": 0.7504599094390869, + "num_tokens": 591985122.0, + "step": 22885 + }, + { + "epoch": 2.513287942016253, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5564181804656982, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7309547662734985, + "num_tokens": 592009719.0, + "step": 22886 + }, + { + "epoch": 2.513397759718867, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3589606285095215, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7302426099777222, + "num_tokens": 592036278.0, + "step": 22887 + }, + { + "epoch": 2.51350757742148, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2520627975463867, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7213895320892334, + "num_tokens": 592062284.0, + "step": 22888 + }, + { + "epoch": 2.513617395124094, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 1.9027029275894165, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7165789604187012, + "num_tokens": 592098411.0, + "step": 22889 + }, + { + "epoch": 2.5137272128267076, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.463491678237915, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7412494421005249, + "num_tokens": 592124820.0, + "step": 22890 + }, + { + "epoch": 2.5138370305293214, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.720759868621826, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7452142834663391, + "num_tokens": 592143372.0, + "step": 22891 + }, + { + "epoch": 2.513946848231935, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2493233680725098, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7249386310577393, + "num_tokens": 592171269.0, + "step": 22892 + }, + { + "epoch": 2.5140566659345485, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.407982349395752, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7246546745300293, + "num_tokens": 592196793.0, + "step": 22893 + }, + { + "epoch": 2.514166483637162, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3301453590393066, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7377121448516846, + "num_tokens": 592221733.0, + "step": 22894 + }, + { + "epoch": 2.514276301339776, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.379307746887207, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7194219827651978, + "num_tokens": 592247370.0, + "step": 22895 + }, + { + "epoch": 2.5143861190423897, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1767418384552, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7037264108657837, + "num_tokens": 592276498.0, + "step": 22896 + }, + { + "epoch": 2.5144959367450035, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.401911497116089, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7329044342041016, + "num_tokens": 592300544.0, + "step": 22897 + }, + { + "epoch": 2.514605754447617, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2414889335632324, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7111876010894775, + "num_tokens": 592330526.0, + "step": 22898 + }, + { + "epoch": 2.5147155721502306, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2509169578552246, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7384488582611084, + "num_tokens": 592358300.0, + "step": 22899 + }, + { + "epoch": 2.5148253898528443, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2008652687072754, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7339768409729004, + "num_tokens": 592385102.0, + "step": 22900 + }, + { + "epoch": 2.514935207555458, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2046866416931152, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7373762726783752, + "num_tokens": 592415240.0, + "step": 22901 + }, + { + "epoch": 2.515045025258072, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.902979612350464, + "learning_rate": 1e-06, + "loss": 0.762, + "mean_token_accuracy": 0.7498868107795715, + "num_tokens": 592432834.0, + "step": 22902 + }, + { + "epoch": 2.515154842960685, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.408604621887207, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.747215211391449, + "num_tokens": 592457682.0, + "step": 22903 + }, + { + "epoch": 2.515264660663299, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.393609046936035, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.7583639025688171, + "num_tokens": 592481379.0, + "step": 22904 + }, + { + "epoch": 2.5153744783659127, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.0946695804595947, + "learning_rate": 1e-06, + "loss": 0.7809, + "mean_token_accuracy": 0.7519875764846802, + "num_tokens": 592511498.0, + "step": 22905 + }, + { + "epoch": 2.515484296068526, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5135443210601807, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7287700176239014, + "num_tokens": 592535201.0, + "step": 22906 + }, + { + "epoch": 2.51559411377114, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 1.947180151939392, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7250216007232666, + "num_tokens": 592569210.0, + "step": 22907 + }, + { + "epoch": 2.5157039314737535, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 1.9010182619094849, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7126939296722412, + "num_tokens": 592606097.0, + "step": 22908 + }, + { + "epoch": 2.5158137491763672, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.9032790660858154, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7316274642944336, + "num_tokens": 592625761.0, + "step": 22909 + }, + { + "epoch": 2.515923566878981, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4020166397094727, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7516037225723267, + "num_tokens": 592649864.0, + "step": 22910 + }, + { + "epoch": 2.5160333845815943, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4334278106689453, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7258089780807495, + "num_tokens": 592674280.0, + "step": 22911 + }, + { + "epoch": 2.516143202284208, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5917234420776367, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7157591581344604, + "num_tokens": 592696943.0, + "step": 22912 + }, + { + "epoch": 2.516253019986822, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.368410587310791, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7348539233207703, + "num_tokens": 592722978.0, + "step": 22913 + }, + { + "epoch": 2.5163628376894356, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4951868057250977, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7485926747322083, + "num_tokens": 592746688.0, + "step": 22914 + }, + { + "epoch": 2.5164726553920493, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.22615647315979, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.730191171169281, + "num_tokens": 592776387.0, + "step": 22915 + }, + { + "epoch": 2.5165824730946627, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.428652286529541, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.720521867275238, + "num_tokens": 592800075.0, + "step": 22916 + }, + { + "epoch": 2.5166922907972764, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2081596851348877, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7478806376457214, + "num_tokens": 592826613.0, + "step": 22917 + }, + { + "epoch": 2.51680210849989, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.345245361328125, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7351019978523254, + "num_tokens": 592851609.0, + "step": 22918 + }, + { + "epoch": 2.516911926202504, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.315077304840088, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7266614437103271, + "num_tokens": 592879143.0, + "step": 22919 + }, + { + "epoch": 2.5170217439051177, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.333439350128174, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7346065044403076, + "num_tokens": 592905359.0, + "step": 22920 + }, + { + "epoch": 2.517131561607731, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.1983304023742676, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7187352180480957, + "num_tokens": 592934634.0, + "step": 22921 + }, + { + "epoch": 2.5172413793103448, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.346935987472534, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7425758242607117, + "num_tokens": 592961039.0, + "step": 22922 + }, + { + "epoch": 2.5173511970129585, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3498904705047607, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7204896211624146, + "num_tokens": 592988378.0, + "step": 22923 + }, + { + "epoch": 2.5174610147155723, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2535383701324463, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7192234992980957, + "num_tokens": 593017157.0, + "step": 22924 + }, + { + "epoch": 2.517570832418186, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.19600248336792, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7035549283027649, + "num_tokens": 593052808.0, + "step": 22925 + }, + { + "epoch": 2.5176806501207993, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.501342535018921, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7521277666091919, + "num_tokens": 593077480.0, + "step": 22926 + }, + { + "epoch": 2.517790467823413, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.057001829147339, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7346868515014648, + "num_tokens": 593111996.0, + "step": 22927 + }, + { + "epoch": 2.517900285526027, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2632014751434326, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7067395448684692, + "num_tokens": 593139114.0, + "step": 22928 + }, + { + "epoch": 2.5180101032286406, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.248983860015869, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7284945249557495, + "num_tokens": 593167042.0, + "step": 22929 + }, + { + "epoch": 2.5181199209312544, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3909361362457275, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7238695025444031, + "num_tokens": 593191849.0, + "step": 22930 + }, + { + "epoch": 2.5182297386338677, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.427700996398926, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7340644598007202, + "num_tokens": 593218714.0, + "step": 22931 + }, + { + "epoch": 2.5183395563364814, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.27223801612854, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7261838912963867, + "num_tokens": 593247208.0, + "step": 22932 + }, + { + "epoch": 2.518449374039095, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4340527057647705, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7172405123710632, + "num_tokens": 593272635.0, + "step": 22933 + }, + { + "epoch": 2.5185591917417085, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4603912830352783, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7271511554718018, + "num_tokens": 593297594.0, + "step": 22934 + }, + { + "epoch": 2.5186690094443223, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.224233627319336, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7230399250984192, + "num_tokens": 593329123.0, + "step": 22935 + }, + { + "epoch": 2.518778827146936, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2767345905303955, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7270691394805908, + "num_tokens": 593358256.0, + "step": 22936 + }, + { + "epoch": 2.5188886448495498, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.682861089706421, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.7564084529876709, + "num_tokens": 593379323.0, + "step": 22937 + }, + { + "epoch": 2.5189984625521635, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.086928606033325, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7392610311508179, + "num_tokens": 593412384.0, + "step": 22938 + }, + { + "epoch": 2.519108280254777, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3903350830078125, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7234625816345215, + "num_tokens": 593439200.0, + "step": 22939 + }, + { + "epoch": 2.5192180979573906, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2541146278381348, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7095794677734375, + "num_tokens": 593467494.0, + "step": 22940 + }, + { + "epoch": 2.5193279156600044, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.100609302520752, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7304167747497559, + "num_tokens": 593497676.0, + "step": 22941 + }, + { + "epoch": 2.519437733362618, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6312174797058105, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7515572905540466, + "num_tokens": 593519999.0, + "step": 22942 + }, + { + "epoch": 2.519547551065232, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4148762226104736, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.709291398525238, + "num_tokens": 593546497.0, + "step": 22943 + }, + { + "epoch": 2.519657368767845, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.2921035289764404, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7323360443115234, + "num_tokens": 593571869.0, + "step": 22944 + }, + { + "epoch": 2.519767186470459, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.4491710662841797, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7350314855575562, + "num_tokens": 593595902.0, + "step": 22945 + }, + { + "epoch": 2.5198770041730727, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.828054428100586, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7525351047515869, + "num_tokens": 593616021.0, + "step": 22946 + }, + { + "epoch": 2.5199868218756865, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.382627010345459, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7161466479301453, + "num_tokens": 593640677.0, + "step": 22947 + }, + { + "epoch": 2.5200966395783, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.007359743118286, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7113099098205566, + "num_tokens": 593675273.0, + "step": 22948 + }, + { + "epoch": 2.5202064572809135, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.6174521446228027, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7225680351257324, + "num_tokens": 593698371.0, + "step": 22949 + }, + { + "epoch": 2.5203162749835273, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.61244535446167, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7102545499801636, + "num_tokens": 593719487.0, + "step": 22950 + }, + { + "epoch": 2.520426092686141, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.225212574005127, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7321051955223083, + "num_tokens": 593748209.0, + "step": 22951 + }, + { + "epoch": 2.520535910388755, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.078906774520874, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7197616100311279, + "num_tokens": 593779992.0, + "step": 22952 + }, + { + "epoch": 2.5206457280913686, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.3262810707092285, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7362511157989502, + "num_tokens": 593805162.0, + "step": 22953 + }, + { + "epoch": 2.520755545793982, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.5222177505493164, + "learning_rate": 1e-06, + "loss": 0.7858, + "mean_token_accuracy": 0.7507510185241699, + "num_tokens": 593826809.0, + "step": 22954 + }, + { + "epoch": 2.5208653634965956, + "ewc_loss": 0.00020122528076171875, + "grad_norm": 2.355290412902832, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.730107307434082, + "num_tokens": 593854502.0, + "step": 22955 + }, + { + "epoch": 2.5209751811992094, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.3468830585479736, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7085968852043152, + "num_tokens": 593882364.0, + "step": 22956 + }, + { + "epoch": 2.521084998901823, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.09582257270813, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7218785285949707, + "num_tokens": 593914002.0, + "step": 22957 + }, + { + "epoch": 2.521194816604437, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.285529136657715, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7486532926559448, + "num_tokens": 593942545.0, + "step": 22958 + }, + { + "epoch": 2.52130463430705, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.2735278606414795, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7287565469741821, + "num_tokens": 593969663.0, + "step": 22959 + }, + { + "epoch": 2.521414452009664, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.5178682804107666, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.718241810798645, + "num_tokens": 593993735.0, + "step": 22960 + }, + { + "epoch": 2.5215242697122777, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.228290319442749, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.731117844581604, + "num_tokens": 594022149.0, + "step": 22961 + }, + { + "epoch": 2.521634087414891, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.359959363937378, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7374751567840576, + "num_tokens": 594048254.0, + "step": 22962 + }, + { + "epoch": 2.521743905117505, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.667633533477783, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7438833713531494, + "num_tokens": 594069964.0, + "step": 22963 + }, + { + "epoch": 2.5218537228201185, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.3544933795928955, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7433795928955078, + "num_tokens": 594094639.0, + "step": 22964 + }, + { + "epoch": 2.5219635405227323, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.7079551219940186, + "learning_rate": 1e-06, + "loss": 0.7976, + "mean_token_accuracy": 0.7478886842727661, + "num_tokens": 594115174.0, + "step": 22965 + }, + { + "epoch": 2.522073358225346, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.5798418521881104, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7352700233459473, + "num_tokens": 594140095.0, + "step": 22966 + }, + { + "epoch": 2.5221831759279594, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.6328935623168945, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7247246503829956, + "num_tokens": 594162787.0, + "step": 22967 + }, + { + "epoch": 2.522292993630573, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.5199759006500244, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7313889265060425, + "num_tokens": 594188094.0, + "step": 22968 + }, + { + "epoch": 2.522402811333187, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.408364772796631, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.729912281036377, + "num_tokens": 594213783.0, + "step": 22969 + }, + { + "epoch": 2.5225126290358006, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.223120927810669, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7207481861114502, + "num_tokens": 594240306.0, + "step": 22970 + }, + { + "epoch": 2.5226224467384144, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.141683340072632, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7293720841407776, + "num_tokens": 594269709.0, + "step": 22971 + }, + { + "epoch": 2.5227322644410277, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.6706485748291016, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7511922717094421, + "num_tokens": 594289564.0, + "step": 22972 + }, + { + "epoch": 2.5228420821436415, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.2802555561065674, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7280229330062866, + "num_tokens": 594317962.0, + "step": 22973 + }, + { + "epoch": 2.5229518998462552, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.5303187370300293, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7470673322677612, + "num_tokens": 594340334.0, + "step": 22974 + }, + { + "epoch": 2.523061717548869, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.520462989807129, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7168634533882141, + "num_tokens": 594364305.0, + "step": 22975 + }, + { + "epoch": 2.5231715352514827, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.669438600540161, + "learning_rate": 1e-06, + "loss": 0.7612, + "mean_token_accuracy": 0.7546865940093994, + "num_tokens": 594383972.0, + "step": 22976 + }, + { + "epoch": 2.523281352954096, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.113334894180298, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7364988923072815, + "num_tokens": 594413752.0, + "step": 22977 + }, + { + "epoch": 2.52339117065671, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.296604633331299, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7152190208435059, + "num_tokens": 594442079.0, + "step": 22978 + }, + { + "epoch": 2.5235009883593236, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.594982862472534, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7077713012695312, + "num_tokens": 594465901.0, + "step": 22979 + }, + { + "epoch": 2.5236108060619373, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.2235772609710693, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.718498945236206, + "num_tokens": 594496278.0, + "step": 22980 + }, + { + "epoch": 2.523720623764551, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.3868846893310547, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7368101477622986, + "num_tokens": 594520948.0, + "step": 22981 + }, + { + "epoch": 2.5238304414671644, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.21767258644104, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7397074699401855, + "num_tokens": 594548810.0, + "step": 22982 + }, + { + "epoch": 2.523940259169778, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.471539258956909, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7039874792098999, + "num_tokens": 594574861.0, + "step": 22983 + }, + { + "epoch": 2.524050076872392, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.2222535610198975, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7117452621459961, + "num_tokens": 594604307.0, + "step": 22984 + }, + { + "epoch": 2.5241598945750052, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.3740615844726562, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7153294086456299, + "num_tokens": 594629850.0, + "step": 22985 + }, + { + "epoch": 2.5242697122776194, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.900480270385742, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7385530471801758, + "num_tokens": 594649519.0, + "step": 22986 + }, + { + "epoch": 2.5243795299802327, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.7751083374023438, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7342886924743652, + "num_tokens": 594669664.0, + "step": 22987 + }, + { + "epoch": 2.5244893476828465, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.245551586151123, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7199492454528809, + "num_tokens": 594698318.0, + "step": 22988 + }, + { + "epoch": 2.5245991653854603, + "ewc_loss": 0.000202178955078125, + "grad_norm": 2.459883213043213, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.722333550453186, + "num_tokens": 594724229.0, + "step": 22989 + }, + { + "epoch": 2.5247089830880736, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4601285457611084, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7256875038146973, + "num_tokens": 594750082.0, + "step": 22990 + }, + { + "epoch": 2.5248188007906873, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.8273818492889404, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7441304922103882, + "num_tokens": 594769454.0, + "step": 22991 + }, + { + "epoch": 2.524928618493301, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6492652893066406, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7454143762588501, + "num_tokens": 594790967.0, + "step": 22992 + }, + { + "epoch": 2.525038436195915, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.709458827972412, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7167847156524658, + "num_tokens": 594813218.0, + "step": 22993 + }, + { + "epoch": 2.5251482538985286, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.055008888244629, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7186459302902222, + "num_tokens": 594844603.0, + "step": 22994 + }, + { + "epoch": 2.525258071601142, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.310760974884033, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7273497581481934, + "num_tokens": 594871719.0, + "step": 22995 + }, + { + "epoch": 2.5253678893037557, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4852492809295654, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7255587577819824, + "num_tokens": 594895488.0, + "step": 22996 + }, + { + "epoch": 2.5254777070063694, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5742440223693848, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.737613320350647, + "num_tokens": 594919364.0, + "step": 22997 + }, + { + "epoch": 2.525587524708983, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4830007553100586, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7296134233474731, + "num_tokens": 594946118.0, + "step": 22998 + }, + { + "epoch": 2.525697342411597, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.470437526702881, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7176581025123596, + "num_tokens": 594969450.0, + "step": 22999 + }, + { + "epoch": 2.5258071601142102, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.799870252609253, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7374628782272339, + "num_tokens": 594988795.0, + "step": 23000 + }, + { + "epoch": 2.525916977816824, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3284242153167725, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7456519603729248, + "num_tokens": 595016239.0, + "step": 23001 + }, + { + "epoch": 2.5260267955194378, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.413905620574951, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.745837390422821, + "num_tokens": 595041564.0, + "step": 23002 + }, + { + "epoch": 2.5261366132220515, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.484421968460083, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.74061119556427, + "num_tokens": 595064933.0, + "step": 23003 + }, + { + "epoch": 2.5262464309246653, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2308313846588135, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7187820672988892, + "num_tokens": 595094820.0, + "step": 23004 + }, + { + "epoch": 2.5263562486272786, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4625160694122314, + "learning_rate": 1e-06, + "loss": 0.7843, + "mean_token_accuracy": 0.754961371421814, + "num_tokens": 595120478.0, + "step": 23005 + }, + { + "epoch": 2.5264660663298923, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.396045446395874, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7436642646789551, + "num_tokens": 595144689.0, + "step": 23006 + }, + { + "epoch": 2.526575884032506, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2670631408691406, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7301604747772217, + "num_tokens": 595171859.0, + "step": 23007 + }, + { + "epoch": 2.52668570173512, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4136857986450195, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7376623749732971, + "num_tokens": 595196206.0, + "step": 23008 + }, + { + "epoch": 2.5267955194377336, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5212202072143555, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7143110036849976, + "num_tokens": 595219237.0, + "step": 23009 + }, + { + "epoch": 2.526905337140347, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2238335609436035, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6966192722320557, + "num_tokens": 595249726.0, + "step": 23010 + }, + { + "epoch": 2.5270151548429607, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3878486156463623, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.73110032081604, + "num_tokens": 595274027.0, + "step": 23011 + }, + { + "epoch": 2.5271249725455744, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.560140371322632, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7165756225585938, + "num_tokens": 595296661.0, + "step": 23012 + }, + { + "epoch": 2.5272347902481878, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.571864604949951, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7400346994400024, + "num_tokens": 595319437.0, + "step": 23013 + }, + { + "epoch": 2.5273446079508015, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0921576023101807, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7092225551605225, + "num_tokens": 595351096.0, + "step": 23014 + }, + { + "epoch": 2.5274544256534153, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.469060182571411, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7273609638214111, + "num_tokens": 595376735.0, + "step": 23015 + }, + { + "epoch": 2.527564243356029, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.292450189590454, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.723298192024231, + "num_tokens": 595404045.0, + "step": 23016 + }, + { + "epoch": 2.527674061058643, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4595465660095215, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7246783971786499, + "num_tokens": 595429491.0, + "step": 23017 + }, + { + "epoch": 2.527783878761256, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3232944011688232, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7099795937538147, + "num_tokens": 595457524.0, + "step": 23018 + }, + { + "epoch": 2.52789369646387, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.026827335357666, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7343469262123108, + "num_tokens": 595490527.0, + "step": 23019 + }, + { + "epoch": 2.5280035141664836, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4945592880249023, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7250030636787415, + "num_tokens": 595515326.0, + "step": 23020 + }, + { + "epoch": 2.5281133318690974, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0659074783325195, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.71958327293396, + "num_tokens": 595547889.0, + "step": 23021 + }, + { + "epoch": 2.528223149571711, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.259507894515991, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7181438207626343, + "num_tokens": 595576016.0, + "step": 23022 + }, + { + "epoch": 2.5283329672743244, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.594942569732666, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7191118001937866, + "num_tokens": 595599972.0, + "step": 23023 + }, + { + "epoch": 2.528442784976938, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.200700521469116, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7385005950927734, + "num_tokens": 595629268.0, + "step": 23024 + }, + { + "epoch": 2.528552602679552, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.463397979736328, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7312344312667847, + "num_tokens": 595653835.0, + "step": 23025 + }, + { + "epoch": 2.5286624203821657, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.389366388320923, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7249378561973572, + "num_tokens": 595679525.0, + "step": 23026 + }, + { + "epoch": 2.5287722380847795, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.535346746444702, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7157379388809204, + "num_tokens": 595704461.0, + "step": 23027 + }, + { + "epoch": 2.528882055787393, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.586855888366699, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7340208292007446, + "num_tokens": 595726467.0, + "step": 23028 + }, + { + "epoch": 2.5289918734900065, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.527855396270752, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7406824827194214, + "num_tokens": 595749917.0, + "step": 23029 + }, + { + "epoch": 2.5291016911926203, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1606333255767822, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.714388370513916, + "num_tokens": 595778799.0, + "step": 23030 + }, + { + "epoch": 2.529211508895234, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.640977144241333, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7330141067504883, + "num_tokens": 595798731.0, + "step": 23031 + }, + { + "epoch": 2.529321326597848, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.499727487564087, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.719944417476654, + "num_tokens": 595824073.0, + "step": 23032 + }, + { + "epoch": 2.529431144300461, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2884299755096436, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7234491109848022, + "num_tokens": 595849564.0, + "step": 23033 + }, + { + "epoch": 2.529540962003075, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.309079170227051, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7280834913253784, + "num_tokens": 595878108.0, + "step": 23034 + }, + { + "epoch": 2.5296507797056886, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.186593770980835, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7173734903335571, + "num_tokens": 595906890.0, + "step": 23035 + }, + { + "epoch": 2.529760597408302, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.783445119857788, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7358081340789795, + "num_tokens": 595926522.0, + "step": 23036 + }, + { + "epoch": 2.529870415110916, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.415780544281006, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.724465548992157, + "num_tokens": 595949645.0, + "step": 23037 + }, + { + "epoch": 2.5299802328135295, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.234851598739624, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7223060131072998, + "num_tokens": 595977614.0, + "step": 23038 + }, + { + "epoch": 2.530090050516143, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1197121143341064, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7192527055740356, + "num_tokens": 596007884.0, + "step": 23039 + }, + { + "epoch": 2.530199868218757, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4587855339050293, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7506317496299744, + "num_tokens": 596031739.0, + "step": 23040 + }, + { + "epoch": 2.5303096859213703, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.827453136444092, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.738815188407898, + "num_tokens": 596050106.0, + "step": 23041 + }, + { + "epoch": 2.530419503623984, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.7008824348449707, + "learning_rate": 1e-06, + "loss": 0.7045, + "mean_token_accuracy": 0.7712540626525879, + "num_tokens": 596069132.0, + "step": 23042 + }, + { + "epoch": 2.530529321326598, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1934993267059326, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7356131076812744, + "num_tokens": 596097303.0, + "step": 23043 + }, + { + "epoch": 2.5306391390292116, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4486677646636963, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.747197687625885, + "num_tokens": 596121083.0, + "step": 23044 + }, + { + "epoch": 2.5307489567318253, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3234710693359375, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7221402525901794, + "num_tokens": 596148564.0, + "step": 23045 + }, + { + "epoch": 2.5308587744344386, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2971343994140625, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7427139282226562, + "num_tokens": 596174298.0, + "step": 23046 + }, + { + "epoch": 2.5309685921370524, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.392340660095215, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7088870406150818, + "num_tokens": 596198003.0, + "step": 23047 + }, + { + "epoch": 2.531078409839666, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.60585880279541, + "learning_rate": 1e-06, + "loss": 0.7607, + "mean_token_accuracy": 0.7545084953308105, + "num_tokens": 596220404.0, + "step": 23048 + }, + { + "epoch": 2.53118822754228, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.186647415161133, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7134990096092224, + "num_tokens": 596253105.0, + "step": 23049 + }, + { + "epoch": 2.5312980452448937, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4100615978240967, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7283742427825928, + "num_tokens": 596277084.0, + "step": 23050 + }, + { + "epoch": 2.531407862947507, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.554849147796631, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7118309736251831, + "num_tokens": 596300341.0, + "step": 23051 + }, + { + "epoch": 2.5315176806501207, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.555250644683838, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7469378709793091, + "num_tokens": 596322781.0, + "step": 23052 + }, + { + "epoch": 2.5316274983527345, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5149943828582764, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7326663732528687, + "num_tokens": 596345388.0, + "step": 23053 + }, + { + "epoch": 2.5317373160553482, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2256851196289062, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7255879044532776, + "num_tokens": 596373827.0, + "step": 23054 + }, + { + "epoch": 2.531847133757962, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4867045879364014, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7468183040618896, + "num_tokens": 596396903.0, + "step": 23055 + }, + { + "epoch": 2.5319569514605753, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3053863048553467, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7152183055877686, + "num_tokens": 596424384.0, + "step": 23056 + }, + { + "epoch": 2.532066769163189, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.435073137283325, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.728550910949707, + "num_tokens": 596448517.0, + "step": 23057 + }, + { + "epoch": 2.532176586865803, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.210789918899536, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7080867886543274, + "num_tokens": 596476470.0, + "step": 23058 + }, + { + "epoch": 2.5322864045684166, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2207469940185547, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7343794703483582, + "num_tokens": 596504634.0, + "step": 23059 + }, + { + "epoch": 2.5323962222710303, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1468324661254883, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.714846134185791, + "num_tokens": 596533937.0, + "step": 23060 + }, + { + "epoch": 2.5325060399736437, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.105597496032715, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7018587589263916, + "num_tokens": 596566767.0, + "step": 23061 + }, + { + "epoch": 2.5326158576762574, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4226176738739014, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7207515239715576, + "num_tokens": 596592004.0, + "step": 23062 + }, + { + "epoch": 2.532725675378871, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.383443832397461, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7358605265617371, + "num_tokens": 596617623.0, + "step": 23063 + }, + { + "epoch": 2.5328354930814845, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1574759483337402, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7336549758911133, + "num_tokens": 596647337.0, + "step": 23064 + }, + { + "epoch": 2.5329453107840982, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.626020908355713, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.734447717666626, + "num_tokens": 596668598.0, + "step": 23065 + }, + { + "epoch": 2.533055128486712, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.146672248840332, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7261643409729004, + "num_tokens": 596699776.0, + "step": 23066 + }, + { + "epoch": 2.5331649461893258, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4493613243103027, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7158096432685852, + "num_tokens": 596725281.0, + "step": 23067 + }, + { + "epoch": 2.5332747638919395, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.7012221813201904, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7471821904182434, + "num_tokens": 596747043.0, + "step": 23068 + }, + { + "epoch": 2.533384581594553, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6438612937927246, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7074043154716492, + "num_tokens": 596771535.0, + "step": 23069 + }, + { + "epoch": 2.5334943992971666, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3318166732788086, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7207975387573242, + "num_tokens": 596798443.0, + "step": 23070 + }, + { + "epoch": 2.5336042169997803, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.7000153064727783, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7463775277137756, + "num_tokens": 596819292.0, + "step": 23071 + }, + { + "epoch": 2.533714034702394, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3133795261383057, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7410035729408264, + "num_tokens": 596848840.0, + "step": 23072 + }, + { + "epoch": 2.533823852405008, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2879674434661865, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.73329097032547, + "num_tokens": 596876828.0, + "step": 23073 + }, + { + "epoch": 2.533933670107621, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.318894147872925, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7179458141326904, + "num_tokens": 596903668.0, + "step": 23074 + }, + { + "epoch": 2.534043487810235, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.033729076385498, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.715054988861084, + "num_tokens": 596939752.0, + "step": 23075 + }, + { + "epoch": 2.5341533055128487, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3974742889404297, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7197839021682739, + "num_tokens": 596965149.0, + "step": 23076 + }, + { + "epoch": 2.5342631232154624, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3969523906707764, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7348492741584778, + "num_tokens": 596988327.0, + "step": 23077 + }, + { + "epoch": 2.534372940918076, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5346598625183105, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7286365032196045, + "num_tokens": 597013645.0, + "step": 23078 + }, + { + "epoch": 2.5344827586206895, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4954915046691895, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7367337942123413, + "num_tokens": 597036964.0, + "step": 23079 + }, + { + "epoch": 2.5345925763233033, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.684614419937134, + "learning_rate": 1e-06, + "loss": 0.7758, + "mean_token_accuracy": 0.7504907846450806, + "num_tokens": 597055558.0, + "step": 23080 + }, + { + "epoch": 2.534702394025917, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3272459506988525, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7160095572471619, + "num_tokens": 597081249.0, + "step": 23081 + }, + { + "epoch": 2.5348122117285308, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3892834186553955, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7417508363723755, + "num_tokens": 597104520.0, + "step": 23082 + }, + { + "epoch": 2.5349220294311445, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0446858406066895, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7322240471839905, + "num_tokens": 597136510.0, + "step": 23083 + }, + { + "epoch": 2.535031847133758, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6875290870666504, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7261836528778076, + "num_tokens": 597157658.0, + "step": 23084 + }, + { + "epoch": 2.5351416648363716, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0683064460754395, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7272793650627136, + "num_tokens": 597189398.0, + "step": 23085 + }, + { + "epoch": 2.5352514825389854, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6512815952301025, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.748126208782196, + "num_tokens": 597210921.0, + "step": 23086 + }, + { + "epoch": 2.5353613002415987, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2754554748535156, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7442144751548767, + "num_tokens": 597237123.0, + "step": 23087 + }, + { + "epoch": 2.535471117944213, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4658591747283936, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7399264574050903, + "num_tokens": 597260578.0, + "step": 23088 + }, + { + "epoch": 2.535580935646826, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2761428356170654, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7412538528442383, + "num_tokens": 597289215.0, + "step": 23089 + }, + { + "epoch": 2.53569075334944, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.266615152359009, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7288592457771301, + "num_tokens": 597316786.0, + "step": 23090 + }, + { + "epoch": 2.5358005710520537, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3930628299713135, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7204995155334473, + "num_tokens": 597340930.0, + "step": 23091 + }, + { + "epoch": 2.535910388754667, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.538881778717041, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7260793447494507, + "num_tokens": 597364706.0, + "step": 23092 + }, + { + "epoch": 2.5360202064572808, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4975037574768066, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7248340845108032, + "num_tokens": 597386630.0, + "step": 23093 + }, + { + "epoch": 2.5361300241598945, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.460876941680908, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7339140176773071, + "num_tokens": 597410659.0, + "step": 23094 + }, + { + "epoch": 2.5362398418625083, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.395622968673706, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7279569506645203, + "num_tokens": 597437572.0, + "step": 23095 + }, + { + "epoch": 2.536349659565122, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4529895782470703, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7331662178039551, + "num_tokens": 597460736.0, + "step": 23096 + }, + { + "epoch": 2.5364594772677354, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4475276470184326, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7308591604232788, + "num_tokens": 597485885.0, + "step": 23097 + }, + { + "epoch": 2.536569294970349, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3565688133239746, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.733693540096283, + "num_tokens": 597511911.0, + "step": 23098 + }, + { + "epoch": 2.536679112672963, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1966021060943604, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7330398559570312, + "num_tokens": 597540549.0, + "step": 23099 + }, + { + "epoch": 2.5367889303755766, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.191549301147461, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7281184196472168, + "num_tokens": 597568988.0, + "step": 23100 + }, + { + "epoch": 2.5368987480781904, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1584670543670654, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7238812446594238, + "num_tokens": 597599355.0, + "step": 23101 + }, + { + "epoch": 2.5370085657808037, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2384026050567627, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7316969633102417, + "num_tokens": 597626892.0, + "step": 23102 + }, + { + "epoch": 2.5371183834834174, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.086437225341797, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7198941111564636, + "num_tokens": 597656696.0, + "step": 23103 + }, + { + "epoch": 2.537228201186031, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1684646606445312, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6928595304489136, + "num_tokens": 597686492.0, + "step": 23104 + }, + { + "epoch": 2.537338018888645, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0753138065338135, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7065076231956482, + "num_tokens": 597720272.0, + "step": 23105 + }, + { + "epoch": 2.5374478365912587, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5170185565948486, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7304271459579468, + "num_tokens": 597745245.0, + "step": 23106 + }, + { + "epoch": 2.537557654293872, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.512816905975342, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7321370244026184, + "num_tokens": 597769483.0, + "step": 23107 + }, + { + "epoch": 2.537667471996486, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3139400482177734, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7384518980979919, + "num_tokens": 597794789.0, + "step": 23108 + }, + { + "epoch": 2.5377772896990995, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.290116786956787, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7297673225402832, + "num_tokens": 597823774.0, + "step": 23109 + }, + { + "epoch": 2.5378871074017133, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1956934928894043, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7227784991264343, + "num_tokens": 597854226.0, + "step": 23110 + }, + { + "epoch": 2.537996925104327, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.426083564758301, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7436000108718872, + "num_tokens": 597879666.0, + "step": 23111 + }, + { + "epoch": 2.5381067428069404, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5935070514678955, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7408226132392883, + "num_tokens": 597903409.0, + "step": 23112 + }, + { + "epoch": 2.538216560509554, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2976160049438477, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7397125363349915, + "num_tokens": 597932638.0, + "step": 23113 + }, + { + "epoch": 2.538326378212168, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.560943603515625, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7436384558677673, + "num_tokens": 597954380.0, + "step": 23114 + }, + { + "epoch": 2.538436195914781, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4315948486328125, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7243309020996094, + "num_tokens": 597980225.0, + "step": 23115 + }, + { + "epoch": 2.538546013617395, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.591242551803589, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.736583948135376, + "num_tokens": 598001818.0, + "step": 23116 + }, + { + "epoch": 2.5386558313200087, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.167501211166382, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.739448070526123, + "num_tokens": 598031330.0, + "step": 23117 + }, + { + "epoch": 2.5387656490226225, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.588977813720703, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7401100397109985, + "num_tokens": 598052679.0, + "step": 23118 + }, + { + "epoch": 2.5388754667252362, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6172444820404053, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7389746308326721, + "num_tokens": 598073395.0, + "step": 23119 + }, + { + "epoch": 2.5389852844278495, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4142823219299316, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7315073013305664, + "num_tokens": 598099590.0, + "step": 23120 + }, + { + "epoch": 2.5390951021304633, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 3.012481927871704, + "learning_rate": 1e-06, + "loss": 0.75, + "mean_token_accuracy": 0.7557982206344604, + "num_tokens": 598117357.0, + "step": 23121 + }, + { + "epoch": 2.539204919833077, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1346657276153564, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7465742826461792, + "num_tokens": 598146407.0, + "step": 23122 + }, + { + "epoch": 2.539314737535691, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6353771686553955, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7296068668365479, + "num_tokens": 598168807.0, + "step": 23123 + }, + { + "epoch": 2.5394245552383046, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.484509229660034, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.733731210231781, + "num_tokens": 598193948.0, + "step": 23124 + }, + { + "epoch": 2.539534372940918, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3099963665008545, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7337021827697754, + "num_tokens": 598220707.0, + "step": 23125 + }, + { + "epoch": 2.5396441906435316, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3322653770446777, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7421928644180298, + "num_tokens": 598246641.0, + "step": 23126 + }, + { + "epoch": 2.5397540083461454, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.334352731704712, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7341806292533875, + "num_tokens": 598275816.0, + "step": 23127 + }, + { + "epoch": 2.539863826048759, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.293257236480713, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.72879558801651, + "num_tokens": 598304609.0, + "step": 23128 + }, + { + "epoch": 2.539973643751373, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1468780040740967, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7159372568130493, + "num_tokens": 598332952.0, + "step": 23129 + }, + { + "epoch": 2.5400834614539862, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1880571842193604, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7126569747924805, + "num_tokens": 598361065.0, + "step": 23130 + }, + { + "epoch": 2.5401932791566, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1707253456115723, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7069361209869385, + "num_tokens": 598394400.0, + "step": 23131 + }, + { + "epoch": 2.5403030968592137, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4064958095550537, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7361325621604919, + "num_tokens": 598421433.0, + "step": 23132 + }, + { + "epoch": 2.5404129145618275, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.676888942718506, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7309579253196716, + "num_tokens": 598441169.0, + "step": 23133 + }, + { + "epoch": 2.5405227322644413, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.552741765975952, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7335683107376099, + "num_tokens": 598463886.0, + "step": 23134 + }, + { + "epoch": 2.5406325499670546, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.733825445175171, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7246401309967041, + "num_tokens": 598485379.0, + "step": 23135 + }, + { + "epoch": 2.5407423676696683, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.489279270172119, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7332615852355957, + "num_tokens": 598511676.0, + "step": 23136 + }, + { + "epoch": 2.540852185372282, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6235973834991455, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7176405191421509, + "num_tokens": 598533470.0, + "step": 23137 + }, + { + "epoch": 2.540962003074896, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3222413063049316, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7188572883605957, + "num_tokens": 598560612.0, + "step": 23138 + }, + { + "epoch": 2.5410718207775096, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.556333541870117, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7448263168334961, + "num_tokens": 598583473.0, + "step": 23139 + }, + { + "epoch": 2.541181638480123, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2195942401885986, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7280137538909912, + "num_tokens": 598611489.0, + "step": 23140 + }, + { + "epoch": 2.5412914561827367, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4423060417175293, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7313027381896973, + "num_tokens": 598633506.0, + "step": 23141 + }, + { + "epoch": 2.5414012738853504, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5523221492767334, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7431909441947937, + "num_tokens": 598655695.0, + "step": 23142 + }, + { + "epoch": 2.5415110915879637, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4146103858947754, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.747451901435852, + "num_tokens": 598678323.0, + "step": 23143 + }, + { + "epoch": 2.5416209092905775, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.372046947479248, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.73963862657547, + "num_tokens": 598702544.0, + "step": 23144 + }, + { + "epoch": 2.5417307269931912, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.36960506439209, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7124322652816772, + "num_tokens": 598728721.0, + "step": 23145 + }, + { + "epoch": 2.541840544695805, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.7378058433532715, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7517746686935425, + "num_tokens": 598748373.0, + "step": 23146 + }, + { + "epoch": 2.5419503623984188, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4304730892181396, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7296543121337891, + "num_tokens": 598772238.0, + "step": 23147 + }, + { + "epoch": 2.542060180101032, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3827764987945557, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7322024703025818, + "num_tokens": 598798213.0, + "step": 23148 + }, + { + "epoch": 2.542169997803646, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3019492626190186, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7089843153953552, + "num_tokens": 598825979.0, + "step": 23149 + }, + { + "epoch": 2.5422798155062596, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.626429319381714, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7421078085899353, + "num_tokens": 598848689.0, + "step": 23150 + }, + { + "epoch": 2.5423896332088733, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1306381225585938, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7293370962142944, + "num_tokens": 598879821.0, + "step": 23151 + }, + { + "epoch": 2.542499450911487, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0770199298858643, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7127695679664612, + "num_tokens": 598912920.0, + "step": 23152 + }, + { + "epoch": 2.5426092686141004, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.183032512664795, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7312405109405518, + "num_tokens": 598941236.0, + "step": 23153 + }, + { + "epoch": 2.542719086316714, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.057194232940674, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7364857196807861, + "num_tokens": 598974363.0, + "step": 23154 + }, + { + "epoch": 2.542828904019328, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4670591354370117, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7209597826004028, + "num_tokens": 598998809.0, + "step": 23155 + }, + { + "epoch": 2.5429387217219417, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.0880849361419678, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7247980833053589, + "num_tokens": 599030383.0, + "step": 23156 + }, + { + "epoch": 2.5430485394245554, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1558761596679688, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7243101596832275, + "num_tokens": 599061889.0, + "step": 23157 + }, + { + "epoch": 2.5431583571271688, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3698816299438477, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.714891791343689, + "num_tokens": 599087343.0, + "step": 23158 + }, + { + "epoch": 2.5432681748297825, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.260152816772461, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7200319766998291, + "num_tokens": 599116440.0, + "step": 23159 + }, + { + "epoch": 2.5433779925323963, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.7579827308654785, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7506762742996216, + "num_tokens": 599137004.0, + "step": 23160 + }, + { + "epoch": 2.54348781023501, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3262202739715576, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7220282554626465, + "num_tokens": 599164010.0, + "step": 23161 + }, + { + "epoch": 2.543597627937624, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4828734397888184, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7490535974502563, + "num_tokens": 599186751.0, + "step": 23162 + }, + { + "epoch": 2.543707445640237, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3081653118133545, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7124003171920776, + "num_tokens": 599216747.0, + "step": 23163 + }, + { + "epoch": 2.543817263342851, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3559670448303223, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7324314713478088, + "num_tokens": 599242956.0, + "step": 23164 + }, + { + "epoch": 2.5439270810454646, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4402031898498535, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7309753894805908, + "num_tokens": 599268162.0, + "step": 23165 + }, + { + "epoch": 2.544036898748078, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.493777275085449, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7317896485328674, + "num_tokens": 599292277.0, + "step": 23166 + }, + { + "epoch": 2.544146716450692, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.73179030418396, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7429186701774597, + "num_tokens": 599314496.0, + "step": 23167 + }, + { + "epoch": 2.5442565341533054, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3666555881500244, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7089424133300781, + "num_tokens": 599341428.0, + "step": 23168 + }, + { + "epoch": 2.544366351855919, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.68757700920105, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7598971724510193, + "num_tokens": 599362376.0, + "step": 23169 + }, + { + "epoch": 2.544476169558533, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6186015605926514, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7524347305297852, + "num_tokens": 599382189.0, + "step": 23170 + }, + { + "epoch": 2.5445859872611463, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.631688356399536, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7425171136856079, + "num_tokens": 599404613.0, + "step": 23171 + }, + { + "epoch": 2.54469580496376, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.9386515617370605, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7340236306190491, + "num_tokens": 599422510.0, + "step": 23172 + }, + { + "epoch": 2.544805622666374, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.466470956802368, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7244151830673218, + "num_tokens": 599448293.0, + "step": 23173 + }, + { + "epoch": 2.5449154403689875, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2474818229675293, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7128801345825195, + "num_tokens": 599478848.0, + "step": 23174 + }, + { + "epoch": 2.5450252580716013, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4545750617980957, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7494465112686157, + "num_tokens": 599502001.0, + "step": 23175 + }, + { + "epoch": 2.5451350757742146, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3003602027893066, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7289154529571533, + "num_tokens": 599528639.0, + "step": 23176 + }, + { + "epoch": 2.5452448934768284, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3706743717193604, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7001842260360718, + "num_tokens": 599556815.0, + "step": 23177 + }, + { + "epoch": 2.545354711179442, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.433842658996582, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7486974596977234, + "num_tokens": 599580546.0, + "step": 23178 + }, + { + "epoch": 2.545464528882056, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.164673328399658, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7234078049659729, + "num_tokens": 599610573.0, + "step": 23179 + }, + { + "epoch": 2.5455743465846696, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.6048059463500977, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7185795307159424, + "num_tokens": 599632809.0, + "step": 23180 + }, + { + "epoch": 2.545684164287283, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.407054901123047, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7130551338195801, + "num_tokens": 599660754.0, + "step": 23181 + }, + { + "epoch": 2.5457939819898967, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.1317954063415527, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7228177785873413, + "num_tokens": 599692272.0, + "step": 23182 + }, + { + "epoch": 2.5459037996925105, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.273679733276367, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7165419459342957, + "num_tokens": 599722418.0, + "step": 23183 + }, + { + "epoch": 2.546013617395124, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.2533066272735596, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7349270582199097, + "num_tokens": 599750615.0, + "step": 23184 + }, + { + "epoch": 2.546123435097738, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3423044681549072, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7406710982322693, + "num_tokens": 599779424.0, + "step": 23185 + }, + { + "epoch": 2.5462332528003513, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3853673934936523, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7376874089241028, + "num_tokens": 599804042.0, + "step": 23186 + }, + { + "epoch": 2.546343070502965, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.410229444503784, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.6929531097412109, + "num_tokens": 599831383.0, + "step": 23187 + }, + { + "epoch": 2.546452888205579, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.4841866493225098, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7423136234283447, + "num_tokens": 599855286.0, + "step": 23188 + }, + { + "epoch": 2.5465627059081926, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.5595741271972656, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7403925657272339, + "num_tokens": 599878271.0, + "step": 23189 + }, + { + "epoch": 2.5466725236108063, + "ewc_loss": 0.00020313262939453125, + "grad_norm": 2.3221139907836914, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7320535182952881, + "num_tokens": 599905729.0, + "step": 23190 + }, + { + "epoch": 2.5467823413134196, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.609616279602051, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.758103609085083, + "num_tokens": 599928077.0, + "step": 23191 + }, + { + "epoch": 2.5468921590160334, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.52651047706604, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7454095482826233, + "num_tokens": 599951375.0, + "step": 23192 + }, + { + "epoch": 2.547001976718647, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2741141319274902, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7277886867523193, + "num_tokens": 599976493.0, + "step": 23193 + }, + { + "epoch": 2.5471117944212605, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5364904403686523, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7392419576644897, + "num_tokens": 600000689.0, + "step": 23194 + }, + { + "epoch": 2.547221612123874, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.188434362411499, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6951103210449219, + "num_tokens": 600031697.0, + "step": 23195 + }, + { + "epoch": 2.547331429826488, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.286290407180786, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7463585734367371, + "num_tokens": 600056997.0, + "step": 23196 + }, + { + "epoch": 2.5474412475291017, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.469775676727295, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7439688444137573, + "num_tokens": 600081186.0, + "step": 23197 + }, + { + "epoch": 2.5475510652317155, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.368407726287842, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7188916802406311, + "num_tokens": 600106506.0, + "step": 23198 + }, + { + "epoch": 2.547660882934329, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2046000957489014, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7125326991081238, + "num_tokens": 600137370.0, + "step": 23199 + }, + { + "epoch": 2.5477707006369426, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.296553611755371, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7373517155647278, + "num_tokens": 600164594.0, + "step": 23200 + }, + { + "epoch": 2.5478805183395563, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3793892860412598, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.727423906326294, + "num_tokens": 600189231.0, + "step": 23201 + }, + { + "epoch": 2.54799033604217, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3630383014678955, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7304229736328125, + "num_tokens": 600214379.0, + "step": 23202 + }, + { + "epoch": 2.548100153744784, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2105278968811035, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7463099360466003, + "num_tokens": 600241873.0, + "step": 23203 + }, + { + "epoch": 2.548209971447397, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3766050338745117, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7400622963905334, + "num_tokens": 600265780.0, + "step": 23204 + }, + { + "epoch": 2.548319789150011, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.291757822036743, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7353604435920715, + "num_tokens": 600294184.0, + "step": 23205 + }, + { + "epoch": 2.5484296068526247, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.237849473953247, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7429428100585938, + "num_tokens": 600322499.0, + "step": 23206 + }, + { + "epoch": 2.5485394245552384, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2874603271484375, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7235761880874634, + "num_tokens": 600351212.0, + "step": 23207 + }, + { + "epoch": 2.548649242257852, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3828985691070557, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7201575636863708, + "num_tokens": 600376114.0, + "step": 23208 + }, + { + "epoch": 2.5487590599604655, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.698056936264038, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7284582853317261, + "num_tokens": 600397781.0, + "step": 23209 + }, + { + "epoch": 2.5488688776630792, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2963905334472656, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7479627728462219, + "num_tokens": 600425370.0, + "step": 23210 + }, + { + "epoch": 2.548978695365693, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.389883279800415, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7137323021888733, + "num_tokens": 600451785.0, + "step": 23211 + }, + { + "epoch": 2.5490885130683067, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.100771427154541, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7330758571624756, + "num_tokens": 600485089.0, + "step": 23212 + }, + { + "epoch": 2.5491983307709205, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.7400455474853516, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7457588315010071, + "num_tokens": 600505764.0, + "step": 23213 + }, + { + "epoch": 2.549308148473534, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.568737506866455, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.738013744354248, + "num_tokens": 600529607.0, + "step": 23214 + }, + { + "epoch": 2.5494179661761476, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.738708019256592, + "learning_rate": 1e-06, + "loss": 0.7629, + "mean_token_accuracy": 0.7611759305000305, + "num_tokens": 600549122.0, + "step": 23215 + }, + { + "epoch": 2.5495277838787613, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.159982442855835, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7331246137619019, + "num_tokens": 600578980.0, + "step": 23216 + }, + { + "epoch": 2.5496376015813746, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4786980152130127, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.762527346611023, + "num_tokens": 600602436.0, + "step": 23217 + }, + { + "epoch": 2.549747419283989, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.10866379737854, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.744821310043335, + "num_tokens": 600631297.0, + "step": 23218 + }, + { + "epoch": 2.549857236986602, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.373962163925171, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7276414632797241, + "num_tokens": 600656923.0, + "step": 23219 + }, + { + "epoch": 2.549967054689216, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2506799697875977, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7280793190002441, + "num_tokens": 600683939.0, + "step": 23220 + }, + { + "epoch": 2.5500768723918297, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2111377716064453, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7234894037246704, + "num_tokens": 600712750.0, + "step": 23221 + }, + { + "epoch": 2.550186690094443, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.443350315093994, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7210109233856201, + "num_tokens": 600736385.0, + "step": 23222 + }, + { + "epoch": 2.5502965077970567, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.677809000015259, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.745448112487793, + "num_tokens": 600756943.0, + "step": 23223 + }, + { + "epoch": 2.5504063254996705, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.296407699584961, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7256699204444885, + "num_tokens": 600784891.0, + "step": 23224 + }, + { + "epoch": 2.5505161432022843, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.097243070602417, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.723344087600708, + "num_tokens": 600815435.0, + "step": 23225 + }, + { + "epoch": 2.550625960904898, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.338578701019287, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7096775770187378, + "num_tokens": 600842585.0, + "step": 23226 + }, + { + "epoch": 2.5507357786075113, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.230745553970337, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7532191872596741, + "num_tokens": 600870637.0, + "step": 23227 + }, + { + "epoch": 2.550845596310125, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5328242778778076, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7347657680511475, + "num_tokens": 600894183.0, + "step": 23228 + }, + { + "epoch": 2.550955414012739, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5391273498535156, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7343326807022095, + "num_tokens": 600917597.0, + "step": 23229 + }, + { + "epoch": 2.5510652317153526, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.0313735008239746, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7220713496208191, + "num_tokens": 600950127.0, + "step": 23230 + }, + { + "epoch": 2.5511750494179664, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.32478928565979, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7309165000915527, + "num_tokens": 600976942.0, + "step": 23231 + }, + { + "epoch": 2.5512848671205797, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.507219076156616, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7338247299194336, + "num_tokens": 601000981.0, + "step": 23232 + }, + { + "epoch": 2.5513946848231934, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3749771118164062, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7480744123458862, + "num_tokens": 601027115.0, + "step": 23233 + }, + { + "epoch": 2.551504502525807, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3925788402557373, + "learning_rate": 1e-06, + "loss": 0.7499, + "mean_token_accuracy": 0.7587887048721313, + "num_tokens": 601050975.0, + "step": 23234 + }, + { + "epoch": 2.551614320228421, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2335703372955322, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7396753430366516, + "num_tokens": 601079541.0, + "step": 23235 + }, + { + "epoch": 2.5517241379310347, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2896153926849365, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7193580865859985, + "num_tokens": 601109464.0, + "step": 23236 + }, + { + "epoch": 2.551833955633648, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2284653186798096, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7454367876052856, + "num_tokens": 601137656.0, + "step": 23237 + }, + { + "epoch": 2.5519437733362618, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.50648832321167, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7371277809143066, + "num_tokens": 601160857.0, + "step": 23238 + }, + { + "epoch": 2.5520535910388755, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3175082206726074, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7441473007202148, + "num_tokens": 601187777.0, + "step": 23239 + }, + { + "epoch": 2.5521634087414893, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.427703380584717, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7470017671585083, + "num_tokens": 601211846.0, + "step": 23240 + }, + { + "epoch": 2.552273226444103, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.439840793609619, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7252274751663208, + "num_tokens": 601236545.0, + "step": 23241 + }, + { + "epoch": 2.5523830441467164, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2280590534210205, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7300235033035278, + "num_tokens": 601264423.0, + "step": 23242 + }, + { + "epoch": 2.55249286184933, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.298661708831787, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7208192944526672, + "num_tokens": 601290790.0, + "step": 23243 + }, + { + "epoch": 2.552602679551944, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.564067840576172, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7407644987106323, + "num_tokens": 601311762.0, + "step": 23244 + }, + { + "epoch": 2.552712497254557, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.275742769241333, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7352107167243958, + "num_tokens": 601338352.0, + "step": 23245 + }, + { + "epoch": 2.552822314957171, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5789778232574463, + "learning_rate": 1e-06, + "loss": 0.7937, + "mean_token_accuracy": 0.7527331709861755, + "num_tokens": 601360971.0, + "step": 23246 + }, + { + "epoch": 2.5529321326597847, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.583270311355591, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7412724494934082, + "num_tokens": 601382171.0, + "step": 23247 + }, + { + "epoch": 2.5530419503623984, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4713568687438965, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7371062636375427, + "num_tokens": 601405621.0, + "step": 23248 + }, + { + "epoch": 2.553151768065012, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.275275468826294, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7251394987106323, + "num_tokens": 601433299.0, + "step": 23249 + }, + { + "epoch": 2.5532615857676255, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.181288957595825, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7363272309303284, + "num_tokens": 601461285.0, + "step": 23250 + }, + { + "epoch": 2.5533714034702393, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2412805557250977, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7105193138122559, + "num_tokens": 601491882.0, + "step": 23251 + }, + { + "epoch": 2.553481221172853, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.72033429145813, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7318191528320312, + "num_tokens": 601513290.0, + "step": 23252 + }, + { + "epoch": 2.553591038875467, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4108176231384277, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7070831656455994, + "num_tokens": 601538138.0, + "step": 23253 + }, + { + "epoch": 2.5537008565780805, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2527034282684326, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7087922096252441, + "num_tokens": 601568430.0, + "step": 23254 + }, + { + "epoch": 2.553810674280694, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3916563987731934, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.732097864151001, + "num_tokens": 601594573.0, + "step": 23255 + }, + { + "epoch": 2.5539204919833076, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.1886038780212402, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7433545589447021, + "num_tokens": 601622798.0, + "step": 23256 + }, + { + "epoch": 2.5540303096859214, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5286295413970947, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7272557020187378, + "num_tokens": 601645628.0, + "step": 23257 + }, + { + "epoch": 2.554140127388535, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.771418809890747, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7212402820587158, + "num_tokens": 601666359.0, + "step": 23258 + }, + { + "epoch": 2.554249945091149, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5918378829956055, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7252103090286255, + "num_tokens": 601687415.0, + "step": 23259 + }, + { + "epoch": 2.554359762793762, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4967761039733887, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.706972599029541, + "num_tokens": 601712944.0, + "step": 23260 + }, + { + "epoch": 2.554469580496376, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.274600028991699, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7279431819915771, + "num_tokens": 601740554.0, + "step": 23261 + }, + { + "epoch": 2.5545793981989897, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.7200875282287598, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7282251119613647, + "num_tokens": 601762863.0, + "step": 23262 + }, + { + "epoch": 2.5546892159016035, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3849756717681885, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7248026132583618, + "num_tokens": 601790171.0, + "step": 23263 + }, + { + "epoch": 2.5547990336042172, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.226412773132324, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7509647607803345, + "num_tokens": 601817728.0, + "step": 23264 + }, + { + "epoch": 2.5549088513068305, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5383875370025635, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7349268198013306, + "num_tokens": 601840399.0, + "step": 23265 + }, + { + "epoch": 2.5550186690094443, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4472694396972656, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7465226650238037, + "num_tokens": 601864120.0, + "step": 23266 + }, + { + "epoch": 2.555128486712058, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.534311294555664, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7268493175506592, + "num_tokens": 601887840.0, + "step": 23267 + }, + { + "epoch": 2.5552383044146714, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.573836326599121, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.716912567615509, + "num_tokens": 601911940.0, + "step": 23268 + }, + { + "epoch": 2.5553481221172856, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.756610631942749, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7277877926826477, + "num_tokens": 601932390.0, + "step": 23269 + }, + { + "epoch": 2.555457939819899, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.229550838470459, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7307226657867432, + "num_tokens": 601961872.0, + "step": 23270 + }, + { + "epoch": 2.5555677575225126, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2735393047332764, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7207459211349487, + "num_tokens": 601990854.0, + "step": 23271 + }, + { + "epoch": 2.5556775752251264, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.6989638805389404, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7376322150230408, + "num_tokens": 602010988.0, + "step": 23272 + }, + { + "epoch": 2.5557873929277397, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3886682987213135, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7409458756446838, + "num_tokens": 602035175.0, + "step": 23273 + }, + { + "epoch": 2.5558972106303535, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4259183406829834, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7184100151062012, + "num_tokens": 602057633.0, + "step": 23274 + }, + { + "epoch": 2.5560070283329672, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.420764446258545, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7450833320617676, + "num_tokens": 602082040.0, + "step": 23275 + }, + { + "epoch": 2.556116846035581, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.590705394744873, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7300872802734375, + "num_tokens": 602104266.0, + "step": 23276 + }, + { + "epoch": 2.5562266637381947, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2319862842559814, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7120149731636047, + "num_tokens": 602131165.0, + "step": 23277 + }, + { + "epoch": 2.556336481440808, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.369675636291504, + "learning_rate": 1e-06, + "loss": 0.7421, + "mean_token_accuracy": 0.7593129873275757, + "num_tokens": 602155657.0, + "step": 23278 + }, + { + "epoch": 2.556446299143422, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.357421875, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7315237522125244, + "num_tokens": 602181965.0, + "step": 23279 + }, + { + "epoch": 2.5565561168460356, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.24442195892334, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7201112508773804, + "num_tokens": 602211976.0, + "step": 23280 + }, + { + "epoch": 2.5566659345486493, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.1340579986572266, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7534940242767334, + "num_tokens": 602239196.0, + "step": 23281 + }, + { + "epoch": 2.556775752251263, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2326996326446533, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7343144416809082, + "num_tokens": 602267444.0, + "step": 23282 + }, + { + "epoch": 2.5568855699538764, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.0522372722625732, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7288972735404968, + "num_tokens": 602298803.0, + "step": 23283 + }, + { + "epoch": 2.55699538765649, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2272071838378906, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7321151494979858, + "num_tokens": 602328423.0, + "step": 23284 + }, + { + "epoch": 2.557105205359104, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5914039611816406, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7217872142791748, + "num_tokens": 602350943.0, + "step": 23285 + }, + { + "epoch": 2.5572150230617177, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.261471748352051, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7352895736694336, + "num_tokens": 602378886.0, + "step": 23286 + }, + { + "epoch": 2.5573248407643314, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.536588430404663, + "learning_rate": 1e-06, + "loss": 0.7677, + "mean_token_accuracy": 0.7547542452812195, + "num_tokens": 602400804.0, + "step": 23287 + }, + { + "epoch": 2.5574346584669447, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4329044818878174, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7201521992683411, + "num_tokens": 602426240.0, + "step": 23288 + }, + { + "epoch": 2.5575444761695585, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.717313289642334, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7326023578643799, + "num_tokens": 602447651.0, + "step": 23289 + }, + { + "epoch": 2.5576542938721722, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.844233274459839, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7262523174285889, + "num_tokens": 602469324.0, + "step": 23290 + }, + { + "epoch": 2.557764111574786, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3599045276641846, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7551131248474121, + "num_tokens": 602493436.0, + "step": 23291 + }, + { + "epoch": 2.5578739292773998, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2115068435668945, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7198337316513062, + "num_tokens": 602524451.0, + "step": 23292 + }, + { + "epoch": 2.557983746980013, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.264280319213867, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7363521456718445, + "num_tokens": 602552139.0, + "step": 23293 + }, + { + "epoch": 2.558093564682627, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.127298355102539, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7339121699333191, + "num_tokens": 602584609.0, + "step": 23294 + }, + { + "epoch": 2.5582033823852406, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2499070167541504, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7070316076278687, + "num_tokens": 602613567.0, + "step": 23295 + }, + { + "epoch": 2.558313200087854, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 3.0425612926483154, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7488038539886475, + "num_tokens": 602630565.0, + "step": 23296 + }, + { + "epoch": 2.5584230177904677, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4568521976470947, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7178021669387817, + "num_tokens": 602655516.0, + "step": 23297 + }, + { + "epoch": 2.5585328354930814, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2908427715301514, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.735214352607727, + "num_tokens": 602681620.0, + "step": 23298 + }, + { + "epoch": 2.558642653195695, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2489383220672607, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7464489340782166, + "num_tokens": 602707308.0, + "step": 23299 + }, + { + "epoch": 2.558752470898309, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.315505027770996, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7344279289245605, + "num_tokens": 602731336.0, + "step": 23300 + }, + { + "epoch": 2.5588622886009222, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.771674156188965, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7526065111160278, + "num_tokens": 602751478.0, + "step": 23301 + }, + { + "epoch": 2.558972106303536, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.433506488800049, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7119455933570862, + "num_tokens": 602776797.0, + "step": 23302 + }, + { + "epoch": 2.5590819240061498, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4057695865631104, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7441428899765015, + "num_tokens": 602799985.0, + "step": 23303 + }, + { + "epoch": 2.5591917417087635, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.453346014022827, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7284332513809204, + "num_tokens": 602825496.0, + "step": 23304 + }, + { + "epoch": 2.5593015594113773, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4905972480773926, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.720283567905426, + "num_tokens": 602849205.0, + "step": 23305 + }, + { + "epoch": 2.5594113771139906, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.1935086250305176, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7122499346733093, + "num_tokens": 602879621.0, + "step": 23306 + }, + { + "epoch": 2.5595211948166043, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.1632611751556396, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7397283315658569, + "num_tokens": 602908426.0, + "step": 23307 + }, + { + "epoch": 2.559631012519218, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3036444187164307, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7402999401092529, + "num_tokens": 602934702.0, + "step": 23308 + }, + { + "epoch": 2.559740830221832, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.6032919883728027, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7347640991210938, + "num_tokens": 602956640.0, + "step": 23309 + }, + { + "epoch": 2.5598506479244456, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3771376609802246, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7471789121627808, + "num_tokens": 602980767.0, + "step": 23310 + }, + { + "epoch": 2.559960465627059, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4359850883483887, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.736991286277771, + "num_tokens": 603004907.0, + "step": 23311 + }, + { + "epoch": 2.5600702833296727, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.246129274368286, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7376465797424316, + "num_tokens": 603031709.0, + "step": 23312 + }, + { + "epoch": 2.5601801010322864, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.224398612976074, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7154872417449951, + "num_tokens": 603061858.0, + "step": 23313 + }, + { + "epoch": 2.5602899187349, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4549665451049805, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7381537556648254, + "num_tokens": 603085881.0, + "step": 23314 + }, + { + "epoch": 2.560399736437514, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.9555115699768066, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7386824488639832, + "num_tokens": 603103768.0, + "step": 23315 + }, + { + "epoch": 2.5605095541401273, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.193646192550659, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7320186495780945, + "num_tokens": 603132650.0, + "step": 23316 + }, + { + "epoch": 2.560619371842741, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.475224733352661, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7373793125152588, + "num_tokens": 603157782.0, + "step": 23317 + }, + { + "epoch": 2.560729189545355, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.838059902191162, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7272215485572815, + "num_tokens": 603177854.0, + "step": 23318 + }, + { + "epoch": 2.5608390072479685, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.7244105339050293, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7474555969238281, + "num_tokens": 603199065.0, + "step": 23319 + }, + { + "epoch": 2.5609488249505823, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.352971076965332, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7321349382400513, + "num_tokens": 603225984.0, + "step": 23320 + }, + { + "epoch": 2.5610586426531956, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3322367668151855, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7330693006515503, + "num_tokens": 603253866.0, + "step": 23321 + }, + { + "epoch": 2.5611684603558094, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.5059096813201904, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7447600364685059, + "num_tokens": 603276823.0, + "step": 23322 + }, + { + "epoch": 2.561278278058423, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.4938812255859375, + "learning_rate": 1e-06, + "loss": 0.7935, + "mean_token_accuracy": 0.7534344792366028, + "num_tokens": 603298508.0, + "step": 23323 + }, + { + "epoch": 2.5613880957610364, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2468197345733643, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7191687822341919, + "num_tokens": 603329742.0, + "step": 23324 + }, + { + "epoch": 2.56149791346365, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2998926639556885, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.70838862657547, + "num_tokens": 603358198.0, + "step": 23325 + }, + { + "epoch": 2.561607731166264, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.441857099533081, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.739142894744873, + "num_tokens": 603383642.0, + "step": 23326 + }, + { + "epoch": 2.5617175488688777, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.33772873878479, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7460235953330994, + "num_tokens": 603409275.0, + "step": 23327 + }, + { + "epoch": 2.5618273665714915, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.141843795776367, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7134274840354919, + "num_tokens": 603440126.0, + "step": 23328 + }, + { + "epoch": 2.5619371842741048, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.777627944946289, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7392054796218872, + "num_tokens": 603459754.0, + "step": 23329 + }, + { + "epoch": 2.5620470019767185, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.040034532546997, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7127516269683838, + "num_tokens": 603491305.0, + "step": 23330 + }, + { + "epoch": 2.5621568196793323, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.3657426834106445, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7387239933013916, + "num_tokens": 603516967.0, + "step": 23331 + }, + { + "epoch": 2.562266637381946, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.2375428676605225, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7365909814834595, + "num_tokens": 603543598.0, + "step": 23332 + }, + { + "epoch": 2.56237645508456, + "ewc_loss": 0.0002040863037109375, + "grad_norm": 2.183915138244629, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7291643023490906, + "num_tokens": 603572869.0, + "step": 23333 + }, + { + "epoch": 2.562486272787173, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.601268768310547, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7195112705230713, + "num_tokens": 603595639.0, + "step": 23334 + }, + { + "epoch": 2.562596090489787, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6705102920532227, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7131050825119019, + "num_tokens": 603619042.0, + "step": 23335 + }, + { + "epoch": 2.5627059081924006, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.031836986541748, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7112321853637695, + "num_tokens": 603654904.0, + "step": 23336 + }, + { + "epoch": 2.5628157258950144, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3982901573181152, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7282926440238953, + "num_tokens": 603679865.0, + "step": 23337 + }, + { + "epoch": 2.562925543597628, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.478856086730957, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7335649132728577, + "num_tokens": 603705678.0, + "step": 23338 + }, + { + "epoch": 2.5630353613002415, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5493204593658447, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7088094353675842, + "num_tokens": 603729771.0, + "step": 23339 + }, + { + "epoch": 2.563145179002855, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.36368989944458, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7228080034255981, + "num_tokens": 603755805.0, + "step": 23340 + }, + { + "epoch": 2.563254996705469, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3544039726257324, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7161010503768921, + "num_tokens": 603784069.0, + "step": 23341 + }, + { + "epoch": 2.5633648144080827, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3896594047546387, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7403407096862793, + "num_tokens": 603809966.0, + "step": 23342 + }, + { + "epoch": 2.5634746321106965, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5648531913757324, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7138224840164185, + "num_tokens": 603833997.0, + "step": 23343 + }, + { + "epoch": 2.56358444981331, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2402522563934326, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7130860090255737, + "num_tokens": 603862599.0, + "step": 23344 + }, + { + "epoch": 2.5636942675159236, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.409815549850464, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7123652100563049, + "num_tokens": 603889002.0, + "step": 23345 + }, + { + "epoch": 2.5638040852185373, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3574345111846924, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7197641134262085, + "num_tokens": 603915766.0, + "step": 23346 + }, + { + "epoch": 2.5639139029211506, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.121483325958252, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7059564590454102, + "num_tokens": 603949935.0, + "step": 23347 + }, + { + "epoch": 2.564023720623765, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4103963375091553, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7503795027732849, + "num_tokens": 603976908.0, + "step": 23348 + }, + { + "epoch": 2.564133538326378, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5143401622772217, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7243325710296631, + "num_tokens": 604002638.0, + "step": 23349 + }, + { + "epoch": 2.564243356028992, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2370755672454834, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7148973941802979, + "num_tokens": 604032425.0, + "step": 23350 + }, + { + "epoch": 2.5643531737316057, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.207864284515381, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7279208898544312, + "num_tokens": 604062291.0, + "step": 23351 + }, + { + "epoch": 2.564462991434219, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4978435039520264, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7299178838729858, + "num_tokens": 604085436.0, + "step": 23352 + }, + { + "epoch": 2.5645728091368327, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.564537763595581, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7188712358474731, + "num_tokens": 604108637.0, + "step": 23353 + }, + { + "epoch": 2.5646826268394465, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.424171209335327, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.720375120639801, + "num_tokens": 604133226.0, + "step": 23354 + }, + { + "epoch": 2.5647924445420602, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4140918254852295, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7274415493011475, + "num_tokens": 604159178.0, + "step": 23355 + }, + { + "epoch": 2.564902262244674, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.630765914916992, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7405984997749329, + "num_tokens": 604181713.0, + "step": 23356 + }, + { + "epoch": 2.5650120799472873, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.137730121612549, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.723461389541626, + "num_tokens": 604211696.0, + "step": 23357 + }, + { + "epoch": 2.565121897649901, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2944159507751465, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7362939715385437, + "num_tokens": 604238132.0, + "step": 23358 + }, + { + "epoch": 2.565231715352515, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.197075605392456, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7304956912994385, + "num_tokens": 604267759.0, + "step": 23359 + }, + { + "epoch": 2.5653415330551286, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.342102527618408, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.727310299873352, + "num_tokens": 604295286.0, + "step": 23360 + }, + { + "epoch": 2.5654513507577423, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5362656116485596, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7366592288017273, + "num_tokens": 604318751.0, + "step": 23361 + }, + { + "epoch": 2.5655611684603556, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2995498180389404, + "learning_rate": 1e-06, + "loss": 0.7641, + "mean_token_accuracy": 0.755923867225647, + "num_tokens": 604343713.0, + "step": 23362 + }, + { + "epoch": 2.5656709861629694, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.077592134475708, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7157425880432129, + "num_tokens": 604376454.0, + "step": 23363 + }, + { + "epoch": 2.565780803865583, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6250064373016357, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7440032958984375, + "num_tokens": 604397121.0, + "step": 23364 + }, + { + "epoch": 2.565890621568197, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.58081316947937, + "learning_rate": 1e-06, + "loss": 0.7479, + "mean_token_accuracy": 0.7627408504486084, + "num_tokens": 604417671.0, + "step": 23365 + }, + { + "epoch": 2.5660004392708107, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5270869731903076, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7244633436203003, + "num_tokens": 604441561.0, + "step": 23366 + }, + { + "epoch": 2.566110256973424, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4443979263305664, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7235954999923706, + "num_tokens": 604467657.0, + "step": 23367 + }, + { + "epoch": 2.5662200746760377, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3698952198028564, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7150677442550659, + "num_tokens": 604497964.0, + "step": 23368 + }, + { + "epoch": 2.5663298923786515, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4477696418762207, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.729728102684021, + "num_tokens": 604523342.0, + "step": 23369 + }, + { + "epoch": 2.5664397100812653, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4224119186401367, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7210313081741333, + "num_tokens": 604548761.0, + "step": 23370 + }, + { + "epoch": 2.566549527783879, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.454710006713867, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7144026756286621, + "num_tokens": 604574531.0, + "step": 23371 + }, + { + "epoch": 2.5666593454864923, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.063356399536133, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7112447023391724, + "num_tokens": 604605288.0, + "step": 23372 + }, + { + "epoch": 2.566769163189106, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.862859010696411, + "learning_rate": 1e-06, + "loss": 0.782, + "mean_token_accuracy": 0.7555532455444336, + "num_tokens": 604624112.0, + "step": 23373 + }, + { + "epoch": 2.56687898089172, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.509209156036377, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7435556650161743, + "num_tokens": 604647904.0, + "step": 23374 + }, + { + "epoch": 2.566988798594333, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5254619121551514, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7294178009033203, + "num_tokens": 604674123.0, + "step": 23375 + }, + { + "epoch": 2.567098616296947, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5106987953186035, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.743202805519104, + "num_tokens": 604698545.0, + "step": 23376 + }, + { + "epoch": 2.5672084339995607, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2687864303588867, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7255116701126099, + "num_tokens": 604728830.0, + "step": 23377 + }, + { + "epoch": 2.5673182517021744, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4412965774536133, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7512651085853577, + "num_tokens": 604753387.0, + "step": 23378 + }, + { + "epoch": 2.567428069404788, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3267247676849365, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.746963381767273, + "num_tokens": 604778137.0, + "step": 23379 + }, + { + "epoch": 2.5675378871074015, + "ewc_loss": 0.00020599365234375, + "grad_norm": 1.8614901304244995, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7327774167060852, + "num_tokens": 604814087.0, + "step": 23380 + }, + { + "epoch": 2.5676477048100153, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.50154709815979, + "learning_rate": 1e-06, + "loss": 0.7593, + "mean_token_accuracy": 0.7585561275482178, + "num_tokens": 604836600.0, + "step": 23381 + }, + { + "epoch": 2.567757522512629, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.7181968688964844, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7538748979568481, + "num_tokens": 604855732.0, + "step": 23382 + }, + { + "epoch": 2.5678673402152428, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3318684101104736, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.751177191734314, + "num_tokens": 604883780.0, + "step": 23383 + }, + { + "epoch": 2.5679771579178565, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.524257183074951, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.7485480308532715, + "num_tokens": 604906772.0, + "step": 23384 + }, + { + "epoch": 2.56808697562047, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2160098552703857, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7003480195999146, + "num_tokens": 604937195.0, + "step": 23385 + }, + { + "epoch": 2.5681967933230836, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5204267501831055, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7269375920295715, + "num_tokens": 604964163.0, + "step": 23386 + }, + { + "epoch": 2.5683066110256974, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2261390686035156, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7182012796401978, + "num_tokens": 604992272.0, + "step": 23387 + }, + { + "epoch": 2.568416428728311, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3786158561706543, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7279535531997681, + "num_tokens": 605016750.0, + "step": 23388 + }, + { + "epoch": 2.568526246430925, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2875633239746094, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7543088793754578, + "num_tokens": 605041357.0, + "step": 23389 + }, + { + "epoch": 2.568636064133538, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.25111722946167, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7452239394187927, + "num_tokens": 605067758.0, + "step": 23390 + }, + { + "epoch": 2.568745881836152, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4223880767822266, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7382342219352722, + "num_tokens": 605092976.0, + "step": 23391 + }, + { + "epoch": 2.5688556995387657, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2073376178741455, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7248700261116028, + "num_tokens": 605121236.0, + "step": 23392 + }, + { + "epoch": 2.5689655172413794, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.113208293914795, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7256516218185425, + "num_tokens": 605152048.0, + "step": 23393 + }, + { + "epoch": 2.569075334943993, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.670292854309082, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7471939921379089, + "num_tokens": 605172717.0, + "step": 23394 + }, + { + "epoch": 2.5691851526466065, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.510399103164673, + "learning_rate": 1e-06, + "loss": 0.775, + "mean_token_accuracy": 0.7517102360725403, + "num_tokens": 605193782.0, + "step": 23395 + }, + { + "epoch": 2.5692949703492203, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.707476854324341, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7283381223678589, + "num_tokens": 605218931.0, + "step": 23396 + }, + { + "epoch": 2.569404788051834, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.359670877456665, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7241982221603394, + "num_tokens": 605246891.0, + "step": 23397 + }, + { + "epoch": 2.5695146057544473, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.323920726776123, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7398197054862976, + "num_tokens": 605272368.0, + "step": 23398 + }, + { + "epoch": 2.5696244234570615, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.715479612350464, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7461408376693726, + "num_tokens": 605292360.0, + "step": 23399 + }, + { + "epoch": 2.569734241159675, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.343667984008789, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7146826982498169, + "num_tokens": 605322031.0, + "step": 23400 + }, + { + "epoch": 2.5698440588622886, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.34237003326416, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7376708984375, + "num_tokens": 605347994.0, + "step": 23401 + }, + { + "epoch": 2.5699538765649024, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.51407527923584, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7256436347961426, + "num_tokens": 605372584.0, + "step": 23402 + }, + { + "epoch": 2.5700636942675157, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.375082492828369, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7351507544517517, + "num_tokens": 605398031.0, + "step": 23403 + }, + { + "epoch": 2.5701735119701294, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.7164857387542725, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7270775437355042, + "num_tokens": 605420220.0, + "step": 23404 + }, + { + "epoch": 2.570283329672743, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.276219606399536, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7126368284225464, + "num_tokens": 605447632.0, + "step": 23405 + }, + { + "epoch": 2.570393147375357, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5798096656799316, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7358967065811157, + "num_tokens": 605469430.0, + "step": 23406 + }, + { + "epoch": 2.5705029650779707, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1793599128723145, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7063931226730347, + "num_tokens": 605499226.0, + "step": 23407 + }, + { + "epoch": 2.570612782780584, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2519114017486572, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7202409505844116, + "num_tokens": 605527397.0, + "step": 23408 + }, + { + "epoch": 2.570722600483198, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2313826084136963, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7187658548355103, + "num_tokens": 605553009.0, + "step": 23409 + }, + { + "epoch": 2.5708324181858115, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.795790672302246, + "learning_rate": 1e-06, + "loss": 0.7716, + "mean_token_accuracy": 0.7520925998687744, + "num_tokens": 605572554.0, + "step": 23410 + }, + { + "epoch": 2.5709422358884253, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.329693555831909, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7249586582183838, + "num_tokens": 605600291.0, + "step": 23411 + }, + { + "epoch": 2.571052053591039, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4683444499969482, + "learning_rate": 1e-06, + "loss": 0.7182, + "mean_token_accuracy": 0.7713327407836914, + "num_tokens": 605624462.0, + "step": 23412 + }, + { + "epoch": 2.5711618712936524, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.182225227355957, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7371189594268799, + "num_tokens": 605654859.0, + "step": 23413 + }, + { + "epoch": 2.571271688996266, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5726706981658936, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7496680617332458, + "num_tokens": 605677494.0, + "step": 23414 + }, + { + "epoch": 2.57138150669888, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2848753929138184, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7236704230308533, + "num_tokens": 605703942.0, + "step": 23415 + }, + { + "epoch": 2.5714913244014936, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5002119541168213, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7073058485984802, + "num_tokens": 605728956.0, + "step": 23416 + }, + { + "epoch": 2.5716011421041074, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4848546981811523, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7498120069503784, + "num_tokens": 605752376.0, + "step": 23417 + }, + { + "epoch": 2.5717109598067207, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5149776935577393, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7245262265205383, + "num_tokens": 605776295.0, + "step": 23418 + }, + { + "epoch": 2.5718207775093345, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5219099521636963, + "learning_rate": 1e-06, + "loss": 0.7901, + "mean_token_accuracy": 0.7536218166351318, + "num_tokens": 605798727.0, + "step": 23419 + }, + { + "epoch": 2.5719305952119482, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.203943967819214, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7047827243804932, + "num_tokens": 605830369.0, + "step": 23420 + }, + { + "epoch": 2.572040412914562, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.489588499069214, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7290841341018677, + "num_tokens": 605853381.0, + "step": 23421 + }, + { + "epoch": 2.5721502306171757, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2766096591949463, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7222583889961243, + "num_tokens": 605881397.0, + "step": 23422 + }, + { + "epoch": 2.572260048319789, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.473620653152466, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7283256649971008, + "num_tokens": 605905674.0, + "step": 23423 + }, + { + "epoch": 2.572369866022403, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.56644606590271, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7271562814712524, + "num_tokens": 605931293.0, + "step": 23424 + }, + { + "epoch": 2.5724796837250166, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.902111530303955, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7304741144180298, + "num_tokens": 605949840.0, + "step": 23425 + }, + { + "epoch": 2.57258950142763, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.702716588973999, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7523852586746216, + "num_tokens": 605969918.0, + "step": 23426 + }, + { + "epoch": 2.5726993191302436, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.429171323776245, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7043664455413818, + "num_tokens": 605997400.0, + "step": 23427 + }, + { + "epoch": 2.5728091368328574, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.414593458175659, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7128937244415283, + "num_tokens": 606022978.0, + "step": 23428 + }, + { + "epoch": 2.572918954535471, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.720686912536621, + "learning_rate": 1e-06, + "loss": 0.7479, + "mean_token_accuracy": 0.7654229402542114, + "num_tokens": 606043132.0, + "step": 23429 + }, + { + "epoch": 2.573028772238085, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.442326784133911, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7121351361274719, + "num_tokens": 606070287.0, + "step": 23430 + }, + { + "epoch": 2.573138589940698, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1633996963500977, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7209728956222534, + "num_tokens": 606100701.0, + "step": 23431 + }, + { + "epoch": 2.573248407643312, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.286738872528076, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7328108549118042, + "num_tokens": 606128149.0, + "step": 23432 + }, + { + "epoch": 2.5733582253459257, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6222870349884033, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7219628095626831, + "num_tokens": 606150533.0, + "step": 23433 + }, + { + "epoch": 2.5734680430485395, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.7423903942108154, + "learning_rate": 1e-06, + "loss": 0.7309, + "mean_token_accuracy": 0.7686194181442261, + "num_tokens": 606169482.0, + "step": 23434 + }, + { + "epoch": 2.5735778607511532, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3280627727508545, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7371150255203247, + "num_tokens": 606195832.0, + "step": 23435 + }, + { + "epoch": 2.5736876784537666, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.178864002227783, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7300400733947754, + "num_tokens": 606225471.0, + "step": 23436 + }, + { + "epoch": 2.5737974961563803, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3419902324676514, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7143028974533081, + "num_tokens": 606251218.0, + "step": 23437 + }, + { + "epoch": 2.573907313858994, + "ewc_loss": 0.00020599365234375, + "grad_norm": 1.9939963817596436, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7095659375190735, + "num_tokens": 606287133.0, + "step": 23438 + }, + { + "epoch": 2.574017131561608, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.0535500049591064, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7135913968086243, + "num_tokens": 606319512.0, + "step": 23439 + }, + { + "epoch": 2.5741269492642216, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.086516857147217, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7304939031600952, + "num_tokens": 606351132.0, + "step": 23440 + }, + { + "epoch": 2.574236766966835, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.259129762649536, + "learning_rate": 1e-06, + "loss": 0.7591, + "mean_token_accuracy": 0.766891598701477, + "num_tokens": 606377231.0, + "step": 23441 + }, + { + "epoch": 2.5743465846694487, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.435079574584961, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7230648994445801, + "num_tokens": 606401545.0, + "step": 23442 + }, + { + "epoch": 2.5744564023720624, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.290093421936035, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.742469072341919, + "num_tokens": 606426396.0, + "step": 23443 + }, + { + "epoch": 2.574566220074676, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.336578369140625, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7371987104415894, + "num_tokens": 606455352.0, + "step": 23444 + }, + { + "epoch": 2.57467603777729, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.539341688156128, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7377581596374512, + "num_tokens": 606478255.0, + "step": 23445 + }, + { + "epoch": 2.5747858554799032, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5689139366149902, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.717631459236145, + "num_tokens": 606502966.0, + "step": 23446 + }, + { + "epoch": 2.574895673182517, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.8744378089904785, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7406135201454163, + "num_tokens": 606520752.0, + "step": 23447 + }, + { + "epoch": 2.5750054908851308, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.185336112976074, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7382569313049316, + "num_tokens": 606549079.0, + "step": 23448 + }, + { + "epoch": 2.575115308587744, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.521873712539673, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7340691089630127, + "num_tokens": 606571157.0, + "step": 23449 + }, + { + "epoch": 2.5752251262903583, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3688673973083496, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7215746641159058, + "num_tokens": 606596777.0, + "step": 23450 + }, + { + "epoch": 2.5753349439929716, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5367307662963867, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7335406541824341, + "num_tokens": 606619862.0, + "step": 23451 + }, + { + "epoch": 2.5754447616955853, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5769851207733154, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7327337861061096, + "num_tokens": 606643526.0, + "step": 23452 + }, + { + "epoch": 2.575554579398199, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6998910903930664, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7429178357124329, + "num_tokens": 606665192.0, + "step": 23453 + }, + { + "epoch": 2.5756643971008124, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2766027450561523, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7091020345687866, + "num_tokens": 606694666.0, + "step": 23454 + }, + { + "epoch": 2.575774214803426, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2876858711242676, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.715730607509613, + "num_tokens": 606721178.0, + "step": 23455 + }, + { + "epoch": 2.57588403250604, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1947872638702393, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7309216260910034, + "num_tokens": 606751290.0, + "step": 23456 + }, + { + "epoch": 2.5759938502086537, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3784019947052, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7206741571426392, + "num_tokens": 606777418.0, + "step": 23457 + }, + { + "epoch": 2.5761036679112674, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4304583072662354, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7337436676025391, + "num_tokens": 606802705.0, + "step": 23458 + }, + { + "epoch": 2.5762134856138807, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.384171962738037, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7406081557273865, + "num_tokens": 606826176.0, + "step": 23459 + }, + { + "epoch": 2.5763233033164945, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4757633209228516, + "learning_rate": 1e-06, + "loss": 0.7684, + "mean_token_accuracy": 0.7553884983062744, + "num_tokens": 606851494.0, + "step": 23460 + }, + { + "epoch": 2.5764331210191083, + "ewc_loss": 0.00020599365234375, + "grad_norm": 3.063120126724243, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7192996144294739, + "num_tokens": 606880711.0, + "step": 23461 + }, + { + "epoch": 2.576542938721722, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.260502338409424, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7118380069732666, + "num_tokens": 606909157.0, + "step": 23462 + }, + { + "epoch": 2.5766527564243358, + "ewc_loss": 0.00020694732666015625, + "grad_norm": 2.490539073944092, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7230772972106934, + "num_tokens": 606934181.0, + "step": 23463 + }, + { + "epoch": 2.576762574126949, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2091169357299805, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6994609832763672, + "num_tokens": 606967102.0, + "step": 23464 + }, + { + "epoch": 2.576872391829563, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1637306213378906, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7421544194221497, + "num_tokens": 606994772.0, + "step": 23465 + }, + { + "epoch": 2.5769822095321766, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2233173847198486, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7067272067070007, + "num_tokens": 607024344.0, + "step": 23466 + }, + { + "epoch": 2.5770920272347904, + "ewc_loss": 0.00020694732666015625, + "grad_norm": 2.6188788414001465, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7318033576011658, + "num_tokens": 607045444.0, + "step": 23467 + }, + { + "epoch": 2.577201844937404, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.274966239929199, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7443764209747314, + "num_tokens": 607072488.0, + "step": 23468 + }, + { + "epoch": 2.5773116626400174, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5558083057403564, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7169668078422546, + "num_tokens": 607095180.0, + "step": 23469 + }, + { + "epoch": 2.577421480342631, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.495574951171875, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.7457119822502136, + "num_tokens": 607118068.0, + "step": 23470 + }, + { + "epoch": 2.577531298045245, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1650795936584473, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7252932786941528, + "num_tokens": 607151100.0, + "step": 23471 + }, + { + "epoch": 2.5776411157478587, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4152512550354004, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.725513756275177, + "num_tokens": 607176109.0, + "step": 23472 + }, + { + "epoch": 2.5777509334504725, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5123512744903564, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7433431148529053, + "num_tokens": 607200271.0, + "step": 23473 + }, + { + "epoch": 2.5778607511530858, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.142732858657837, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7123384475708008, + "num_tokens": 607231627.0, + "step": 23474 + }, + { + "epoch": 2.5779705688556995, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.208768367767334, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7320882678031921, + "num_tokens": 607261567.0, + "step": 23475 + }, + { + "epoch": 2.5780803865583133, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.547375202178955, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7394427061080933, + "num_tokens": 607283953.0, + "step": 23476 + }, + { + "epoch": 2.5781902042609266, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.8396122455596924, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.6996309757232666, + "num_tokens": 607314055.0, + "step": 23477 + }, + { + "epoch": 2.578300021963541, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.229905366897583, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7378726005554199, + "num_tokens": 607341826.0, + "step": 23478 + }, + { + "epoch": 2.578409839666154, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5099093914031982, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7454095482826233, + "num_tokens": 607362328.0, + "step": 23479 + }, + { + "epoch": 2.578519657368768, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6941044330596924, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7329147458076477, + "num_tokens": 607385401.0, + "step": 23480 + }, + { + "epoch": 2.5786294750713816, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2777950763702393, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7316812872886658, + "num_tokens": 607411585.0, + "step": 23481 + }, + { + "epoch": 2.578739292773995, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.427469491958618, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7203222513198853, + "num_tokens": 607438405.0, + "step": 23482 + }, + { + "epoch": 2.5788491104766087, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.33089280128479, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7411276698112488, + "num_tokens": 607464513.0, + "step": 23483 + }, + { + "epoch": 2.5789589281792225, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4030532836914062, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.740371823310852, + "num_tokens": 607489036.0, + "step": 23484 + }, + { + "epoch": 2.579068745881836, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.129941701889038, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7232463359832764, + "num_tokens": 607518788.0, + "step": 23485 + }, + { + "epoch": 2.57917856358445, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2497739791870117, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7410603761672974, + "num_tokens": 607545199.0, + "step": 23486 + }, + { + "epoch": 2.5792883812870633, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3516831398010254, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7129206657409668, + "num_tokens": 607571705.0, + "step": 23487 + }, + { + "epoch": 2.579398198989677, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2014520168304443, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7432365417480469, + "num_tokens": 607600658.0, + "step": 23488 + }, + { + "epoch": 2.579508016692291, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5165765285491943, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7461689710617065, + "num_tokens": 607622523.0, + "step": 23489 + }, + { + "epoch": 2.5796178343949046, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2444822788238525, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.749916672706604, + "num_tokens": 607649235.0, + "step": 23490 + }, + { + "epoch": 2.5797276520975183, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.36130952835083, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7367359399795532, + "num_tokens": 607674183.0, + "step": 23491 + }, + { + "epoch": 2.5798374698001316, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6149191856384277, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.719418466091156, + "num_tokens": 607697653.0, + "step": 23492 + }, + { + "epoch": 2.5799472875027454, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.395662307739258, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7328858375549316, + "num_tokens": 607720883.0, + "step": 23493 + }, + { + "epoch": 2.580057105205359, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6200618743896484, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7424386143684387, + "num_tokens": 607742152.0, + "step": 23494 + }, + { + "epoch": 2.580166922907973, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3746514320373535, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7312283515930176, + "num_tokens": 607768335.0, + "step": 23495 + }, + { + "epoch": 2.5802767406105866, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.328094482421875, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.724366307258606, + "num_tokens": 607795566.0, + "step": 23496 + }, + { + "epoch": 2.5803865583132, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2690014839172363, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7129036784172058, + "num_tokens": 607825518.0, + "step": 23497 + }, + { + "epoch": 2.5804963760158137, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.395465850830078, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7392986416816711, + "num_tokens": 607850631.0, + "step": 23498 + }, + { + "epoch": 2.5806061937184275, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1658802032470703, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7004902362823486, + "num_tokens": 607880161.0, + "step": 23499 + }, + { + "epoch": 2.5807160114210412, + "ewc_loss": 0.00020599365234375, + "grad_norm": 1.9819660186767578, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.707821786403656, + "num_tokens": 607912535.0, + "step": 23500 + }, + { + "epoch": 2.580825829123655, + "ewc_loss": 0.00020599365234375, + "grad_norm": 1.9640733003616333, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7273459434509277, + "num_tokens": 607949154.0, + "step": 23501 + }, + { + "epoch": 2.5809356468262683, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2882604598999023, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7148840427398682, + "num_tokens": 607978259.0, + "step": 23502 + }, + { + "epoch": 2.581045464528882, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.590355157852173, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7255764007568359, + "num_tokens": 607999828.0, + "step": 23503 + }, + { + "epoch": 2.581155282231496, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4573099613189697, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7234448194503784, + "num_tokens": 608026370.0, + "step": 23504 + }, + { + "epoch": 2.581265099934109, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3429157733917236, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.722212553024292, + "num_tokens": 608052186.0, + "step": 23505 + }, + { + "epoch": 2.581374917636723, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.713965654373169, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7523446083068848, + "num_tokens": 608071921.0, + "step": 23506 + }, + { + "epoch": 2.5814847353393366, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.6729836463928223, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7343630790710449, + "num_tokens": 608094605.0, + "step": 23507 + }, + { + "epoch": 2.5815945530419504, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2942118644714355, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7031475305557251, + "num_tokens": 608121527.0, + "step": 23508 + }, + { + "epoch": 2.581704370744564, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3294286727905273, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7230525016784668, + "num_tokens": 608148738.0, + "step": 23509 + }, + { + "epoch": 2.5818141884471775, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.251218795776367, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7167277932167053, + "num_tokens": 608178076.0, + "step": 23510 + }, + { + "epoch": 2.5819240061497912, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5918848514556885, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7511982321739197, + "num_tokens": 608200723.0, + "step": 23511 + }, + { + "epoch": 2.582033823852405, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.573345899581909, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7374250888824463, + "num_tokens": 608223552.0, + "step": 23512 + }, + { + "epoch": 2.5821436415550187, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.341909170150757, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7182431221008301, + "num_tokens": 608248956.0, + "step": 23513 + }, + { + "epoch": 2.5822534592576325, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1046671867370605, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7284954786300659, + "num_tokens": 608281144.0, + "step": 23514 + }, + { + "epoch": 2.582363276960246, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.9150519371032715, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7408556938171387, + "num_tokens": 608298919.0, + "step": 23515 + }, + { + "epoch": 2.5824730946628596, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.906674385070801, + "learning_rate": 1e-06, + "loss": 0.7572, + "mean_token_accuracy": 0.7611083984375, + "num_tokens": 608317932.0, + "step": 23516 + }, + { + "epoch": 2.5825829123654733, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.418452262878418, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7308213710784912, + "num_tokens": 608343042.0, + "step": 23517 + }, + { + "epoch": 2.582692730068087, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.353360176086426, + "learning_rate": 1e-06, + "loss": 0.7438, + "mean_token_accuracy": 0.758612334728241, + "num_tokens": 608366394.0, + "step": 23518 + }, + { + "epoch": 2.582802547770701, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4909861087799072, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7192744612693787, + "num_tokens": 608391877.0, + "step": 23519 + }, + { + "epoch": 2.582912365473314, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.22546648979187, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7132419347763062, + "num_tokens": 608422930.0, + "step": 23520 + }, + { + "epoch": 2.583022183175928, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.2249515056610107, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.6996632814407349, + "num_tokens": 608452170.0, + "step": 23521 + }, + { + "epoch": 2.5831320008785417, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.587681531906128, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7476909756660461, + "num_tokens": 608474138.0, + "step": 23522 + }, + { + "epoch": 2.5832418185811554, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3389554023742676, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.745526909828186, + "num_tokens": 608500710.0, + "step": 23523 + }, + { + "epoch": 2.583351636283769, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.351416826248169, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7438184022903442, + "num_tokens": 608526318.0, + "step": 23524 + }, + { + "epoch": 2.5834614539863825, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.5555126667022705, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7357687950134277, + "num_tokens": 608549646.0, + "step": 23525 + }, + { + "epoch": 2.5835712716889963, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3818161487579346, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7124297022819519, + "num_tokens": 608576690.0, + "step": 23526 + }, + { + "epoch": 2.58368108939161, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.252336025238037, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7196183204650879, + "num_tokens": 608605824.0, + "step": 23527 + }, + { + "epoch": 2.5837909070942233, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.384237289428711, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7310526967048645, + "num_tokens": 608631410.0, + "step": 23528 + }, + { + "epoch": 2.5839007247968375, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.330827236175537, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7258308529853821, + "num_tokens": 608659695.0, + "step": 23529 + }, + { + "epoch": 2.584010542499451, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4107446670532227, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7216053009033203, + "num_tokens": 608684659.0, + "step": 23530 + }, + { + "epoch": 2.5841203602020646, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3535146713256836, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7311524152755737, + "num_tokens": 608712998.0, + "step": 23531 + }, + { + "epoch": 2.5842301779046783, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1109976768493652, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7210193872451782, + "num_tokens": 608742565.0, + "step": 23532 + }, + { + "epoch": 2.5843399956072917, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.1694326400756836, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7340118885040283, + "num_tokens": 608772595.0, + "step": 23533 + }, + { + "epoch": 2.5844498133099054, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.30808687210083, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7153277397155762, + "num_tokens": 608800380.0, + "step": 23534 + }, + { + "epoch": 2.584559631012519, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.331435441970825, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7089544534683228, + "num_tokens": 608826483.0, + "step": 23535 + }, + { + "epoch": 2.584669448715133, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.4456520080566406, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7213435769081116, + "num_tokens": 608851677.0, + "step": 23536 + }, + { + "epoch": 2.5847792664177467, + "ewc_loss": 0.00020599365234375, + "grad_norm": 2.3529000282287598, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7253239750862122, + "num_tokens": 608877458.0, + "step": 23537 + }, + { + "epoch": 2.58488908412036, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3883514404296875, + "learning_rate": 1e-06, + "loss": 0.7936, + "mean_token_accuracy": 0.7473791837692261, + "num_tokens": 608903584.0, + "step": 23538 + }, + { + "epoch": 2.5849989018229738, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3007705211639404, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7265529632568359, + "num_tokens": 608931746.0, + "step": 23539 + }, + { + "epoch": 2.5851087195255875, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.301090717315674, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.701021671295166, + "num_tokens": 608960365.0, + "step": 23540 + }, + { + "epoch": 2.5852185372282013, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.191659688949585, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.724107027053833, + "num_tokens": 608990885.0, + "step": 23541 + }, + { + "epoch": 2.585328354930815, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.194124937057495, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7275295853614807, + "num_tokens": 609018899.0, + "step": 23542 + }, + { + "epoch": 2.5854381726334283, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4713478088378906, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.72181236743927, + "num_tokens": 609042820.0, + "step": 23543 + }, + { + "epoch": 2.585547990336042, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.172072649002075, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7282963991165161, + "num_tokens": 609071626.0, + "step": 23544 + }, + { + "epoch": 2.585657808038656, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5953240394592285, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7235664129257202, + "num_tokens": 609094227.0, + "step": 23545 + }, + { + "epoch": 2.5857676257412696, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.188603162765503, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7251025438308716, + "num_tokens": 609125138.0, + "step": 23546 + }, + { + "epoch": 2.5858774434438834, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.2902638912200928, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7496142387390137, + "num_tokens": 609153763.0, + "step": 23547 + }, + { + "epoch": 2.5859872611464967, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3838515281677246, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7202494740486145, + "num_tokens": 609181468.0, + "step": 23548 + }, + { + "epoch": 2.5860970788491104, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4995548725128174, + "learning_rate": 1e-06, + "loss": 0.7666, + "mean_token_accuracy": 0.7496930956840515, + "num_tokens": 609203204.0, + "step": 23549 + }, + { + "epoch": 2.586206896551724, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.532297372817993, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7424852848052979, + "num_tokens": 609226339.0, + "step": 23550 + }, + { + "epoch": 2.586316714254338, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4400413036346436, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7243378162384033, + "num_tokens": 609251273.0, + "step": 23551 + }, + { + "epoch": 2.5864265319569517, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5347583293914795, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7258986234664917, + "num_tokens": 609275377.0, + "step": 23552 + }, + { + "epoch": 2.586536349659565, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.40895676612854, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7271139621734619, + "num_tokens": 609302709.0, + "step": 23553 + }, + { + "epoch": 2.586646167362179, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.481717109680176, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7132079601287842, + "num_tokens": 609328684.0, + "step": 23554 + }, + { + "epoch": 2.5867559850647925, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.232247829437256, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6955734491348267, + "num_tokens": 609358995.0, + "step": 23555 + }, + { + "epoch": 2.586865802767406, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4162967205047607, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7171218395233154, + "num_tokens": 609383638.0, + "step": 23556 + }, + { + "epoch": 2.5869756204700196, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.519400119781494, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7190593481063843, + "num_tokens": 609406471.0, + "step": 23557 + }, + { + "epoch": 2.5870854381726334, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3686540126800537, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7136020064353943, + "num_tokens": 609432105.0, + "step": 23558 + }, + { + "epoch": 2.587195255875247, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3133792877197266, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7141354084014893, + "num_tokens": 609460303.0, + "step": 23559 + }, + { + "epoch": 2.587305073577861, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.522341251373291, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7098426818847656, + "num_tokens": 609482805.0, + "step": 23560 + }, + { + "epoch": 2.587414891280474, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.503303289413452, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7388228178024292, + "num_tokens": 609508378.0, + "step": 23561 + }, + { + "epoch": 2.587524708983088, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4528768062591553, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.729548990726471, + "num_tokens": 609532991.0, + "step": 23562 + }, + { + "epoch": 2.5876345266857017, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.1489920616149902, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7072374820709229, + "num_tokens": 609562786.0, + "step": 23563 + }, + { + "epoch": 2.5877443443883155, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.503838539123535, + "learning_rate": 1e-06, + "loss": 0.7392, + "mean_token_accuracy": 0.7580994963645935, + "num_tokens": 609584798.0, + "step": 23564 + }, + { + "epoch": 2.587854162090929, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.1697144508361816, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7314039468765259, + "num_tokens": 609613866.0, + "step": 23565 + }, + { + "epoch": 2.5879639797935425, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3528287410736084, + "learning_rate": 1e-06, + "loss": 0.7816, + "mean_token_accuracy": 0.7646997570991516, + "num_tokens": 609637725.0, + "step": 23566 + }, + { + "epoch": 2.5880737974961563, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.222707986831665, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7329351902008057, + "num_tokens": 609666176.0, + "step": 23567 + }, + { + "epoch": 2.58818361519877, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5481584072113037, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7434374094009399, + "num_tokens": 609688773.0, + "step": 23568 + }, + { + "epoch": 2.588293432901384, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.545492172241211, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7349404096603394, + "num_tokens": 609713472.0, + "step": 23569 + }, + { + "epoch": 2.5884032506039976, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3076512813568115, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7454430460929871, + "num_tokens": 609740360.0, + "step": 23570 + }, + { + "epoch": 2.588513068306611, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5012810230255127, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7417604923248291, + "num_tokens": 609767501.0, + "step": 23571 + }, + { + "epoch": 2.5886228860092246, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.478778123855591, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7371387481689453, + "num_tokens": 609791624.0, + "step": 23572 + }, + { + "epoch": 2.5887327037118384, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.7536120414733887, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.7547752261161804, + "num_tokens": 609811024.0, + "step": 23573 + }, + { + "epoch": 2.588842521414452, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.293046474456787, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7392310500144958, + "num_tokens": 609835938.0, + "step": 23574 + }, + { + "epoch": 2.588952339117066, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.150111436843872, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.729773998260498, + "num_tokens": 609865852.0, + "step": 23575 + }, + { + "epoch": 2.589062156819679, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.2625620365142822, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7247357368469238, + "num_tokens": 609892294.0, + "step": 23576 + }, + { + "epoch": 2.589171974522293, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.252469301223755, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7301160097122192, + "num_tokens": 609918345.0, + "step": 23577 + }, + { + "epoch": 2.5892817922249067, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5234954357147217, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7256380319595337, + "num_tokens": 609943155.0, + "step": 23578 + }, + { + "epoch": 2.58939160992752, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3445727825164795, + "learning_rate": 1e-06, + "loss": 0.7509, + "mean_token_accuracy": 0.7587355375289917, + "num_tokens": 609967669.0, + "step": 23579 + }, + { + "epoch": 2.5895014276301342, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.077080011367798, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7225470542907715, + "num_tokens": 609999713.0, + "step": 23580 + }, + { + "epoch": 2.5896112453327476, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.15189528465271, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7140452861785889, + "num_tokens": 610029806.0, + "step": 23581 + }, + { + "epoch": 2.5897210630353613, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.126627206802368, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7050759792327881, + "num_tokens": 610060118.0, + "step": 23582 + }, + { + "epoch": 2.589830880737975, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4158833026885986, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7199182510375977, + "num_tokens": 610086735.0, + "step": 23583 + }, + { + "epoch": 2.5899406984405884, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.197880506515503, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7080525159835815, + "num_tokens": 610115401.0, + "step": 23584 + }, + { + "epoch": 2.590050516143202, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5414092540740967, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7413114905357361, + "num_tokens": 610138102.0, + "step": 23585 + }, + { + "epoch": 2.590160333845816, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.1627070903778076, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7195238471031189, + "num_tokens": 610169734.0, + "step": 23586 + }, + { + "epoch": 2.5902701515484297, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4057974815368652, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7255135178565979, + "num_tokens": 610194219.0, + "step": 23587 + }, + { + "epoch": 2.5903799692510434, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.463411569595337, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7072095274925232, + "num_tokens": 610218464.0, + "step": 23588 + }, + { + "epoch": 2.5904897869536567, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4695627689361572, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.735338568687439, + "num_tokens": 610244318.0, + "step": 23589 + }, + { + "epoch": 2.5905996046562705, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.6802523136138916, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7296912670135498, + "num_tokens": 610265950.0, + "step": 23590 + }, + { + "epoch": 2.5907094223588842, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.5854270458221436, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7288980484008789, + "num_tokens": 610289450.0, + "step": 23591 + }, + { + "epoch": 2.590819240061498, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.226621150970459, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7265547513961792, + "num_tokens": 610318546.0, + "step": 23592 + }, + { + "epoch": 2.5909290577641118, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.2493457794189453, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7303013801574707, + "num_tokens": 610344578.0, + "step": 23593 + }, + { + "epoch": 2.591038875466725, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.532013177871704, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7454288601875305, + "num_tokens": 610368562.0, + "step": 23594 + }, + { + "epoch": 2.591148693169339, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.400707244873047, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.729202151298523, + "num_tokens": 610393880.0, + "step": 23595 + }, + { + "epoch": 2.5912585108719526, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.112969160079956, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6960679888725281, + "num_tokens": 610427393.0, + "step": 23596 + }, + { + "epoch": 2.5913683285745663, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.6559927463531494, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7462463974952698, + "num_tokens": 610449326.0, + "step": 23597 + }, + { + "epoch": 2.59147814627718, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3897740840911865, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7174140214920044, + "num_tokens": 610477099.0, + "step": 23598 + }, + { + "epoch": 2.5915879639797934, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5064988136291504, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7167993187904358, + "num_tokens": 610502104.0, + "step": 23599 + }, + { + "epoch": 2.591697781682407, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.4605820178985596, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7233232259750366, + "num_tokens": 610526704.0, + "step": 23600 + }, + { + "epoch": 2.591807599385021, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.3608107566833496, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7352191209793091, + "num_tokens": 610552280.0, + "step": 23601 + }, + { + "epoch": 2.5919174170876347, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.9419031143188477, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7497071027755737, + "num_tokens": 610570278.0, + "step": 23602 + }, + { + "epoch": 2.5920272347902484, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.367676258087158, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7207156419754028, + "num_tokens": 610594914.0, + "step": 23603 + }, + { + "epoch": 2.5921370524928617, + "ewc_loss": 0.0002079010009765625, + "grad_norm": 2.506333112716675, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7462586164474487, + "num_tokens": 610618667.0, + "step": 23604 + }, + { + "epoch": 2.5922468701954755, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4527132511138916, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7463768720626831, + "num_tokens": 610641748.0, + "step": 23605 + }, + { + "epoch": 2.5923566878980893, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 1.9405652284622192, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6910045146942139, + "num_tokens": 610678830.0, + "step": 23606 + }, + { + "epoch": 2.5924665056007026, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.171170949935913, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.752091646194458, + "num_tokens": 610706433.0, + "step": 23607 + }, + { + "epoch": 2.5925763233033163, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3464362621307373, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7288532853126526, + "num_tokens": 610732306.0, + "step": 23608 + }, + { + "epoch": 2.59268614100593, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.530956983566284, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7396758794784546, + "num_tokens": 610755381.0, + "step": 23609 + }, + { + "epoch": 2.592795958708544, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.539774179458618, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.710869550704956, + "num_tokens": 610779631.0, + "step": 23610 + }, + { + "epoch": 2.5929057764111576, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.441343307495117, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7011639475822449, + "num_tokens": 610807265.0, + "step": 23611 + }, + { + "epoch": 2.593015594113771, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.092829465866089, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7333155274391174, + "num_tokens": 610837654.0, + "step": 23612 + }, + { + "epoch": 2.5931254118163847, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2368037700653076, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7425073981285095, + "num_tokens": 610863735.0, + "step": 23613 + }, + { + "epoch": 2.5932352295189984, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1465752124786377, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7304149866104126, + "num_tokens": 610894086.0, + "step": 23614 + }, + { + "epoch": 2.593345047221612, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.587678909301758, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7348496913909912, + "num_tokens": 610916455.0, + "step": 23615 + }, + { + "epoch": 2.593454864924226, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.6232495307922363, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7250908613204956, + "num_tokens": 610939189.0, + "step": 23616 + }, + { + "epoch": 2.5935646826268393, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4934983253479004, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7441760301589966, + "num_tokens": 610962545.0, + "step": 23617 + }, + { + "epoch": 2.593674500329453, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.991947889328003, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7382267713546753, + "num_tokens": 610979802.0, + "step": 23618 + }, + { + "epoch": 2.5937843180320668, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.275937080383301, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7278995513916016, + "num_tokens": 611007790.0, + "step": 23619 + }, + { + "epoch": 2.5938941357346805, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.7137932777404785, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7241193056106567, + "num_tokens": 611028766.0, + "step": 23620 + }, + { + "epoch": 2.5940039534372943, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2195627689361572, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7133827805519104, + "num_tokens": 611060140.0, + "step": 23621 + }, + { + "epoch": 2.5941137711399076, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4418110847473145, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.735775351524353, + "num_tokens": 611084267.0, + "step": 23622 + }, + { + "epoch": 2.5942235888425214, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.424060344696045, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7160323858261108, + "num_tokens": 611110761.0, + "step": 23623 + }, + { + "epoch": 2.594333406545135, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3744490146636963, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7041866183280945, + "num_tokens": 611136743.0, + "step": 23624 + }, + { + "epoch": 2.594443224247749, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4191951751708984, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7313016653060913, + "num_tokens": 611162697.0, + "step": 23625 + }, + { + "epoch": 2.5945530419503626, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2762815952301025, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7226893305778503, + "num_tokens": 611190809.0, + "step": 23626 + }, + { + "epoch": 2.594662859652976, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3574891090393066, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.6994237899780273, + "num_tokens": 611219321.0, + "step": 23627 + }, + { + "epoch": 2.5947726773555897, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.71370792388916, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7360888719558716, + "num_tokens": 611240407.0, + "step": 23628 + }, + { + "epoch": 2.5948824950582035, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2081661224365234, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7231103777885437, + "num_tokens": 611269693.0, + "step": 23629 + }, + { + "epoch": 2.594992312760817, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.394960403442383, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7071875333786011, + "num_tokens": 611294073.0, + "step": 23630 + }, + { + "epoch": 2.595102130463431, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2522952556610107, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7191762328147888, + "num_tokens": 611322370.0, + "step": 23631 + }, + { + "epoch": 2.5952119481660443, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.459083080291748, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7205955982208252, + "num_tokens": 611346265.0, + "step": 23632 + }, + { + "epoch": 2.595321765868658, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4324088096618652, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7342023849487305, + "num_tokens": 611371607.0, + "step": 23633 + }, + { + "epoch": 2.595431583571272, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.198317766189575, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7459463477134705, + "num_tokens": 611399678.0, + "step": 23634 + }, + { + "epoch": 2.595541401273885, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.41593599319458, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7245417833328247, + "num_tokens": 611428886.0, + "step": 23635 + }, + { + "epoch": 2.595651218976499, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.6213185787200928, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7254976034164429, + "num_tokens": 611451301.0, + "step": 23636 + }, + { + "epoch": 2.5957610366791126, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2157108783721924, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7291726469993591, + "num_tokens": 611478880.0, + "step": 23637 + }, + { + "epoch": 2.5958708543817264, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2544503211975098, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7252214550971985, + "num_tokens": 611506458.0, + "step": 23638 + }, + { + "epoch": 2.59598067208434, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3202450275421143, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.745734453201294, + "num_tokens": 611532584.0, + "step": 23639 + }, + { + "epoch": 2.5960904897869534, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 1.9327595233917236, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7319474220275879, + "num_tokens": 611568513.0, + "step": 23640 + }, + { + "epoch": 2.596200307489567, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5901763439178467, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7107610702514648, + "num_tokens": 611595791.0, + "step": 23641 + }, + { + "epoch": 2.596310125192181, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3392114639282227, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7149857878684998, + "num_tokens": 611623092.0, + "step": 23642 + }, + { + "epoch": 2.5964199428947947, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.476139545440674, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7362461090087891, + "num_tokens": 611646677.0, + "step": 23643 + }, + { + "epoch": 2.5965297605974085, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.303204298019409, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7379128932952881, + "num_tokens": 611672403.0, + "step": 23644 + }, + { + "epoch": 2.596639578300022, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.230370283126831, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7325171232223511, + "num_tokens": 611702719.0, + "step": 23645 + }, + { + "epoch": 2.5967493960026355, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 1.9224973917007446, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.706718921661377, + "num_tokens": 611740960.0, + "step": 23646 + }, + { + "epoch": 2.5968592137052493, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2392611503601074, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7305841445922852, + "num_tokens": 611770401.0, + "step": 23647 + }, + { + "epoch": 2.596969031407863, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.550001621246338, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7471571564674377, + "num_tokens": 611791834.0, + "step": 23648 + }, + { + "epoch": 2.597078849110477, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4643845558166504, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7274098992347717, + "num_tokens": 611818006.0, + "step": 23649 + }, + { + "epoch": 2.59718866681309, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.829437255859375, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7348555326461792, + "num_tokens": 611836610.0, + "step": 23650 + }, + { + "epoch": 2.597298484515704, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.141613245010376, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7155795693397522, + "num_tokens": 611868349.0, + "step": 23651 + }, + { + "epoch": 2.5974083022183176, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.671314001083374, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7535264492034912, + "num_tokens": 611888622.0, + "step": 23652 + }, + { + "epoch": 2.5975181199209314, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.043360471725464, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7445169687271118, + "num_tokens": 611919472.0, + "step": 23653 + }, + { + "epoch": 2.597627937623545, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1147232055664062, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7265135049819946, + "num_tokens": 611949302.0, + "step": 23654 + }, + { + "epoch": 2.5977377553261585, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2777562141418457, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.734158992767334, + "num_tokens": 611977106.0, + "step": 23655 + }, + { + "epoch": 2.5978475730287722, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3912353515625, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7300039529800415, + "num_tokens": 612002776.0, + "step": 23656 + }, + { + "epoch": 2.597957390731386, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3755102157592773, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7152901887893677, + "num_tokens": 612028747.0, + "step": 23657 + }, + { + "epoch": 2.5980672084339993, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.6839075088500977, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7447187900543213, + "num_tokens": 612051398.0, + "step": 23658 + }, + { + "epoch": 2.5981770261366135, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.525019645690918, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7364673614501953, + "num_tokens": 612076770.0, + "step": 23659 + }, + { + "epoch": 2.598286843839227, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.391658306121826, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.741074800491333, + "num_tokens": 612100720.0, + "step": 23660 + }, + { + "epoch": 2.5983966615418406, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4963650703430176, + "learning_rate": 1e-06, + "loss": 0.7742, + "mean_token_accuracy": 0.7513508796691895, + "num_tokens": 612124654.0, + "step": 23661 + }, + { + "epoch": 2.5985064792444543, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.6047987937927246, + "learning_rate": 1e-06, + "loss": 0.776, + "mean_token_accuracy": 0.7608104944229126, + "num_tokens": 612144913.0, + "step": 23662 + }, + { + "epoch": 2.5986162969470676, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3060553073883057, + "learning_rate": 1e-06, + "loss": 0.7798, + "mean_token_accuracy": 0.7530337572097778, + "num_tokens": 612171849.0, + "step": 23663 + }, + { + "epoch": 2.5987261146496814, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1416709423065186, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7082479000091553, + "num_tokens": 612203713.0, + "step": 23664 + }, + { + "epoch": 2.598835932352295, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1835315227508545, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7270653247833252, + "num_tokens": 612232019.0, + "step": 23665 + }, + { + "epoch": 2.598945750054909, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.661560535430908, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7280913591384888, + "num_tokens": 612254444.0, + "step": 23666 + }, + { + "epoch": 2.5990555677575227, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.333822250366211, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7499529123306274, + "num_tokens": 612280856.0, + "step": 23667 + }, + { + "epoch": 2.599165385460136, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4743268489837646, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7241512537002563, + "num_tokens": 612304793.0, + "step": 23668 + }, + { + "epoch": 2.5992752031627497, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3926010131835938, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7138347625732422, + "num_tokens": 612330615.0, + "step": 23669 + }, + { + "epoch": 2.5993850208653635, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3745343685150146, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.743425726890564, + "num_tokens": 612356540.0, + "step": 23670 + }, + { + "epoch": 2.5994948385679773, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.614967107772827, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7393238544464111, + "num_tokens": 612378341.0, + "step": 23671 + }, + { + "epoch": 2.599604656270591, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.277115821838379, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7194234132766724, + "num_tokens": 612409321.0, + "step": 23672 + }, + { + "epoch": 2.5997144739732043, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.349670648574829, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7067630290985107, + "num_tokens": 612436828.0, + "step": 23673 + }, + { + "epoch": 2.599824291675818, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4341442584991455, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7255797982215881, + "num_tokens": 612461194.0, + "step": 23674 + }, + { + "epoch": 2.599934109378432, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5743048191070557, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7284142374992371, + "num_tokens": 612484439.0, + "step": 23675 + }, + { + "epoch": 2.6000439270810456, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3033134937286377, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7337762117385864, + "num_tokens": 612510719.0, + "step": 23676 + }, + { + "epoch": 2.6001537447836593, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.531005859375, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7402641773223877, + "num_tokens": 612531870.0, + "step": 23677 + }, + { + "epoch": 2.6002635624862727, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3376119136810303, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7417203783988953, + "num_tokens": 612559693.0, + "step": 23678 + }, + { + "epoch": 2.6003733801888864, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5175766944885254, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7336514592170715, + "num_tokens": 612582916.0, + "step": 23679 + }, + { + "epoch": 2.6004831978915, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.555110216140747, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7334801554679871, + "num_tokens": 612607430.0, + "step": 23680 + }, + { + "epoch": 2.600593015594114, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4305074214935303, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7309445142745972, + "num_tokens": 612632203.0, + "step": 23681 + }, + { + "epoch": 2.6007028332967277, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2315516471862793, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7046470046043396, + "num_tokens": 612663563.0, + "step": 23682 + }, + { + "epoch": 2.600812650999341, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5998222827911377, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7345938682556152, + "num_tokens": 612687162.0, + "step": 23683 + }, + { + "epoch": 2.6009224687019548, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.900491952896118, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7269079685211182, + "num_tokens": 612708150.0, + "step": 23684 + }, + { + "epoch": 2.6010322864045685, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5179708003997803, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7137777805328369, + "num_tokens": 612732273.0, + "step": 23685 + }, + { + "epoch": 2.601142104107182, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1351661682128906, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7208273410797119, + "num_tokens": 612762367.0, + "step": 23686 + }, + { + "epoch": 2.6012519218097956, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.676666021347046, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7407448291778564, + "num_tokens": 612784479.0, + "step": 23687 + }, + { + "epoch": 2.6013617395124093, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.476555824279785, + "learning_rate": 1e-06, + "loss": 0.7749, + "mean_token_accuracy": 0.7544698715209961, + "num_tokens": 612807784.0, + "step": 23688 + }, + { + "epoch": 2.601471557215023, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.6249828338623047, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.736126184463501, + "num_tokens": 612830441.0, + "step": 23689 + }, + { + "epoch": 2.601581374917637, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.061469316482544, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7524538040161133, + "num_tokens": 612860287.0, + "step": 23690 + }, + { + "epoch": 2.60169119262025, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5667476654052734, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7220280170440674, + "num_tokens": 612885853.0, + "step": 23691 + }, + { + "epoch": 2.601801010322864, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.337634325027466, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.722281277179718, + "num_tokens": 612913970.0, + "step": 23692 + }, + { + "epoch": 2.6019108280254777, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5441691875457764, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7435474991798401, + "num_tokens": 612936848.0, + "step": 23693 + }, + { + "epoch": 2.6020206457280914, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2858734130859375, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7263804078102112, + "num_tokens": 612962683.0, + "step": 23694 + }, + { + "epoch": 2.602130463430705, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2304694652557373, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7548381090164185, + "num_tokens": 612990744.0, + "step": 23695 + }, + { + "epoch": 2.6022402811333185, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.304940938949585, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.740068793296814, + "num_tokens": 613017058.0, + "step": 23696 + }, + { + "epoch": 2.6023500988359323, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.629695415496826, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7481396794319153, + "num_tokens": 613038577.0, + "step": 23697 + }, + { + "epoch": 2.602459916538546, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2866225242614746, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7171908020973206, + "num_tokens": 613067846.0, + "step": 23698 + }, + { + "epoch": 2.60256973424116, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.7161705493927, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7240563631057739, + "num_tokens": 613089819.0, + "step": 23699 + }, + { + "epoch": 2.6026795519437735, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.644819974899292, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7302423715591431, + "num_tokens": 613111268.0, + "step": 23700 + }, + { + "epoch": 2.602789369646387, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2056257724761963, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7358161211013794, + "num_tokens": 613138421.0, + "step": 23701 + }, + { + "epoch": 2.6028991873490006, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4905099868774414, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7076325416564941, + "num_tokens": 613163254.0, + "step": 23702 + }, + { + "epoch": 2.6030090050516144, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.099274158477783, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7231844663619995, + "num_tokens": 613196287.0, + "step": 23703 + }, + { + "epoch": 2.603118822754228, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.627338171005249, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7202795743942261, + "num_tokens": 613220044.0, + "step": 23704 + }, + { + "epoch": 2.603228640456842, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.617302179336548, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7188888788223267, + "num_tokens": 613241739.0, + "step": 23705 + }, + { + "epoch": 2.603338458159455, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.37416934967041, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.705172598361969, + "num_tokens": 613269674.0, + "step": 23706 + }, + { + "epoch": 2.603448275862069, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.294201135635376, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.726654589176178, + "num_tokens": 613299200.0, + "step": 23707 + }, + { + "epoch": 2.6035580935646827, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2176403999328613, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7479003071784973, + "num_tokens": 613326806.0, + "step": 23708 + }, + { + "epoch": 2.603667911267296, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.035691738128662, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7032502889633179, + "num_tokens": 613363218.0, + "step": 23709 + }, + { + "epoch": 2.60377772896991, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.146409273147583, + "learning_rate": 1e-06, + "loss": 0.7748, + "mean_token_accuracy": 0.7529268860816956, + "num_tokens": 613392091.0, + "step": 23710 + }, + { + "epoch": 2.6038875466725235, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4216725826263428, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7512765526771545, + "num_tokens": 613415949.0, + "step": 23711 + }, + { + "epoch": 2.6039973643751373, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2851955890655518, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7200725078582764, + "num_tokens": 613442712.0, + "step": 23712 + }, + { + "epoch": 2.604107182077751, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.394921064376831, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7216278314590454, + "num_tokens": 613468117.0, + "step": 23713 + }, + { + "epoch": 2.6042169997803644, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.1763267517089844, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7055503129959106, + "num_tokens": 613497539.0, + "step": 23714 + }, + { + "epoch": 2.604326817482978, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.337963819503784, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7370004653930664, + "num_tokens": 613522873.0, + "step": 23715 + }, + { + "epoch": 2.604436635185592, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3680191040039062, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7024234533309937, + "num_tokens": 613549805.0, + "step": 23716 + }, + { + "epoch": 2.6045464528882056, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3373024463653564, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7170675992965698, + "num_tokens": 613576634.0, + "step": 23717 + }, + { + "epoch": 2.6046562705908194, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5358972549438477, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7260550260543823, + "num_tokens": 613600814.0, + "step": 23718 + }, + { + "epoch": 2.6047660882934327, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3630599975585938, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7272629737854004, + "num_tokens": 613626305.0, + "step": 23719 + }, + { + "epoch": 2.6048759059960465, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.274177312850952, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7385559678077698, + "num_tokens": 613654566.0, + "step": 23720 + }, + { + "epoch": 2.60498572369866, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.370032787322998, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7418129444122314, + "num_tokens": 613678702.0, + "step": 23721 + }, + { + "epoch": 2.605095541401274, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.601020336151123, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7005338668823242, + "num_tokens": 613705791.0, + "step": 23722 + }, + { + "epoch": 2.6052053591038877, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2618603706359863, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7191758155822754, + "num_tokens": 613734905.0, + "step": 23723 + }, + { + "epoch": 2.605315176806501, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4885075092315674, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7491422891616821, + "num_tokens": 613758211.0, + "step": 23724 + }, + { + "epoch": 2.605424994509115, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.720996856689453, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7217941284179688, + "num_tokens": 613778836.0, + "step": 23725 + }, + { + "epoch": 2.6055348122117286, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.690502643585205, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.725009560585022, + "num_tokens": 613801624.0, + "step": 23726 + }, + { + "epoch": 2.6056446299143423, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3624603748321533, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7272897958755493, + "num_tokens": 613829644.0, + "step": 23727 + }, + { + "epoch": 2.605754447616956, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.45363712310791, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7369975447654724, + "num_tokens": 613853936.0, + "step": 23728 + }, + { + "epoch": 2.6058642653195694, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.699136734008789, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.749409556388855, + "num_tokens": 613873765.0, + "step": 23729 + }, + { + "epoch": 2.605974083022183, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.156751871109009, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7250758409500122, + "num_tokens": 613901446.0, + "step": 23730 + }, + { + "epoch": 2.606083900724797, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4077024459838867, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7394284009933472, + "num_tokens": 613926484.0, + "step": 23731 + }, + { + "epoch": 2.6061937184274107, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3329310417175293, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7277379035949707, + "num_tokens": 613955178.0, + "step": 23732 + }, + { + "epoch": 2.6063035361300244, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.116276502609253, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7176247835159302, + "num_tokens": 613986262.0, + "step": 23733 + }, + { + "epoch": 2.6064133538326377, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.802962064743042, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.7546374201774597, + "num_tokens": 614004905.0, + "step": 23734 + }, + { + "epoch": 2.6065231715352515, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.520197629928589, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7372162342071533, + "num_tokens": 614028283.0, + "step": 23735 + }, + { + "epoch": 2.6066329892378652, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.9500606060028076, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7420459985733032, + "num_tokens": 614046571.0, + "step": 23736 + }, + { + "epoch": 2.6067428069404786, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.7529921531677246, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7451194524765015, + "num_tokens": 614067079.0, + "step": 23737 + }, + { + "epoch": 2.6068526246430923, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4944474697113037, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7228313088417053, + "num_tokens": 614092766.0, + "step": 23738 + }, + { + "epoch": 2.606962442345706, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5631096363067627, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7363153100013733, + "num_tokens": 614115213.0, + "step": 23739 + }, + { + "epoch": 2.60707226004832, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.7475550174713135, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7385815978050232, + "num_tokens": 614137069.0, + "step": 23740 + }, + { + "epoch": 2.6071820777509336, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.770336866378784, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.755260705947876, + "num_tokens": 614160859.0, + "step": 23741 + }, + { + "epoch": 2.607291895453547, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2745792865753174, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7235103249549866, + "num_tokens": 614188900.0, + "step": 23742 + }, + { + "epoch": 2.6074017131561606, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3733861446380615, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.728134036064148, + "num_tokens": 614217035.0, + "step": 23743 + }, + { + "epoch": 2.6075115308587744, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.5530753135681152, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7270396947860718, + "num_tokens": 614239744.0, + "step": 23744 + }, + { + "epoch": 2.607621348561388, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.3234188556671143, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7342411875724792, + "num_tokens": 614266888.0, + "step": 23745 + }, + { + "epoch": 2.607731166264002, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2305312156677246, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7106261253356934, + "num_tokens": 614295459.0, + "step": 23746 + }, + { + "epoch": 2.6078409839666152, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.401010751724243, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7419369220733643, + "num_tokens": 614319657.0, + "step": 23747 + }, + { + "epoch": 2.607950801669229, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.48000431060791, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7309702038764954, + "num_tokens": 614343494.0, + "step": 23748 + }, + { + "epoch": 2.6080606193718427, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2732882499694824, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7300642728805542, + "num_tokens": 614371395.0, + "step": 23749 + }, + { + "epoch": 2.6081704370744565, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.421100378036499, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7333795428276062, + "num_tokens": 614398280.0, + "step": 23750 + }, + { + "epoch": 2.6082802547770703, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4383509159088135, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7363072633743286, + "num_tokens": 614423501.0, + "step": 23751 + }, + { + "epoch": 2.6083900724796836, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.8061060905456543, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.748938798904419, + "num_tokens": 614443447.0, + "step": 23752 + }, + { + "epoch": 2.6084998901822973, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4306907653808594, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7104743719100952, + "num_tokens": 614470168.0, + "step": 23753 + }, + { + "epoch": 2.608609707884911, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.2433981895446777, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7320270538330078, + "num_tokens": 614496569.0, + "step": 23754 + }, + { + "epoch": 2.608719525587525, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.419003486633301, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7185692191123962, + "num_tokens": 614522567.0, + "step": 23755 + }, + { + "epoch": 2.6088293432901386, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.4144837856292725, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7225750684738159, + "num_tokens": 614547755.0, + "step": 23756 + }, + { + "epoch": 2.608939160992752, + "ewc_loss": 0.00020885467529296875, + "grad_norm": 2.570685625076294, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7192349433898926, + "num_tokens": 614574353.0, + "step": 23757 + }, + { + "epoch": 2.6090489786953657, + "ewc_loss": 0.000209808349609375, + "grad_norm": 2.3699514865875244, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.734950065612793, + "num_tokens": 614600494.0, + "step": 23758 + }, + { + "epoch": 2.6091587963979794, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.588129758834839, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7274618148803711, + "num_tokens": 614624797.0, + "step": 23759 + }, + { + "epoch": 2.6092686141005927, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2592644691467285, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.6972156763076782, + "num_tokens": 614655913.0, + "step": 23760 + }, + { + "epoch": 2.609378431803207, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.536536931991577, + "learning_rate": 1e-06, + "loss": 0.775, + "mean_token_accuracy": 0.7489309310913086, + "num_tokens": 614679388.0, + "step": 23761 + }, + { + "epoch": 2.6094882495058203, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1816017627716064, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7406518459320068, + "num_tokens": 614709117.0, + "step": 23762 + }, + { + "epoch": 2.609598067208434, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.9364967346191406, + "learning_rate": 1e-06, + "loss": 0.7696, + "mean_token_accuracy": 0.7492062449455261, + "num_tokens": 614727989.0, + "step": 23763 + }, + { + "epoch": 2.6097078849110478, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.53062105178833, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7436474561691284, + "num_tokens": 614750503.0, + "step": 23764 + }, + { + "epoch": 2.609817702613661, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5583298206329346, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7282362580299377, + "num_tokens": 614772555.0, + "step": 23765 + }, + { + "epoch": 2.609927520316275, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4029295444488525, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7145054340362549, + "num_tokens": 614798477.0, + "step": 23766 + }, + { + "epoch": 2.6100373380188886, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.381002187728882, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7432089447975159, + "num_tokens": 614821899.0, + "step": 23767 + }, + { + "epoch": 2.6101471557215024, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2235336303710938, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7381114959716797, + "num_tokens": 614853929.0, + "step": 23768 + }, + { + "epoch": 2.610256973424116, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.205221176147461, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7500735521316528, + "num_tokens": 614880210.0, + "step": 23769 + }, + { + "epoch": 2.6103667911267294, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4546844959259033, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.7497026920318604, + "num_tokens": 614903738.0, + "step": 23770 + }, + { + "epoch": 2.610476608829343, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.202788829803467, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7547782063484192, + "num_tokens": 614929478.0, + "step": 23771 + }, + { + "epoch": 2.610586426531957, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.291435956954956, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7234603762626648, + "num_tokens": 614955346.0, + "step": 23772 + }, + { + "epoch": 2.6106962442345707, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2378296852111816, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7381969690322876, + "num_tokens": 614982587.0, + "step": 23773 + }, + { + "epoch": 2.6108060619371845, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.466886281967163, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7190057039260864, + "num_tokens": 615006632.0, + "step": 23774 + }, + { + "epoch": 2.6109158796397978, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.241563081741333, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7326703071594238, + "num_tokens": 615035959.0, + "step": 23775 + }, + { + "epoch": 2.6110256973424115, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2234301567077637, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.713088870048523, + "num_tokens": 615064681.0, + "step": 23776 + }, + { + "epoch": 2.6111355150450253, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.388658285140991, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7296315431594849, + "num_tokens": 615090058.0, + "step": 23777 + }, + { + "epoch": 2.611245332747639, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3774521350860596, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7424862384796143, + "num_tokens": 615115131.0, + "step": 23778 + }, + { + "epoch": 2.611355150450253, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.488701105117798, + "learning_rate": 1e-06, + "loss": 0.781, + "mean_token_accuracy": 0.7529573440551758, + "num_tokens": 615137660.0, + "step": 23779 + }, + { + "epoch": 2.611464968152866, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3073463439941406, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7386139631271362, + "num_tokens": 615164764.0, + "step": 23780 + }, + { + "epoch": 2.61157478585548, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2563557624816895, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7335652112960815, + "num_tokens": 615193518.0, + "step": 23781 + }, + { + "epoch": 2.6116846035580936, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.591113567352295, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7484170198440552, + "num_tokens": 615214833.0, + "step": 23782 + }, + { + "epoch": 2.6117944212607074, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2981083393096924, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7351889610290527, + "num_tokens": 615240146.0, + "step": 23783 + }, + { + "epoch": 2.611904238963321, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3069005012512207, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7353183031082153, + "num_tokens": 615266486.0, + "step": 23784 + }, + { + "epoch": 2.6120140566659344, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2432169914245605, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7299610376358032, + "num_tokens": 615295792.0, + "step": 23785 + }, + { + "epoch": 2.612123874368548, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.265491247177124, + "learning_rate": 1e-06, + "loss": 0.7216, + "mean_token_accuracy": 0.7622257471084595, + "num_tokens": 615322940.0, + "step": 23786 + }, + { + "epoch": 2.612233692071162, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5958547592163086, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.735977292060852, + "num_tokens": 615344774.0, + "step": 23787 + }, + { + "epoch": 2.6123435097737753, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.661428213119507, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7424670457839966, + "num_tokens": 615365924.0, + "step": 23788 + }, + { + "epoch": 2.612453327476389, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.441211223602295, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7352864742279053, + "num_tokens": 615390288.0, + "step": 23789 + }, + { + "epoch": 2.612563145179003, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4567880630493164, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7541964054107666, + "num_tokens": 615413048.0, + "step": 23790 + }, + { + "epoch": 2.6126729628816165, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4067702293395996, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7450484037399292, + "num_tokens": 615436740.0, + "step": 23791 + }, + { + "epoch": 2.6127827805842303, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.248258352279663, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7345589399337769, + "num_tokens": 615462034.0, + "step": 23792 + }, + { + "epoch": 2.6128925982868436, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.351764678955078, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.725568950176239, + "num_tokens": 615488700.0, + "step": 23793 + }, + { + "epoch": 2.6130024159894574, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.463268280029297, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7321280241012573, + "num_tokens": 615513606.0, + "step": 23794 + }, + { + "epoch": 2.613112233692071, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5191750526428223, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7275502681732178, + "num_tokens": 615535227.0, + "step": 23795 + }, + { + "epoch": 2.613222051394685, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2448291778564453, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7161173820495605, + "num_tokens": 615562831.0, + "step": 23796 + }, + { + "epoch": 2.6133318690972986, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1881015300750732, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7054932117462158, + "num_tokens": 615592648.0, + "step": 23797 + }, + { + "epoch": 2.613441686799912, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6260619163513184, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7332508563995361, + "num_tokens": 615614363.0, + "step": 23798 + }, + { + "epoch": 2.6135515045025257, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1276357173919678, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7013784646987915, + "num_tokens": 615647265.0, + "step": 23799 + }, + { + "epoch": 2.6136613222051395, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4511539936065674, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7489728927612305, + "num_tokens": 615671285.0, + "step": 23800 + }, + { + "epoch": 2.6137711399077532, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.127973794937134, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7144143581390381, + "num_tokens": 615704263.0, + "step": 23801 + }, + { + "epoch": 2.613880957610367, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4690260887145996, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7301134467124939, + "num_tokens": 615728820.0, + "step": 23802 + }, + { + "epoch": 2.6139907753129803, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1357290744781494, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7160335779190063, + "num_tokens": 615758594.0, + "step": 23803 + }, + { + "epoch": 2.614100593015594, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2142996788024902, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7046516537666321, + "num_tokens": 615789841.0, + "step": 23804 + }, + { + "epoch": 2.614210410718208, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.560750722885132, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7291858792304993, + "num_tokens": 615811739.0, + "step": 23805 + }, + { + "epoch": 2.6143202284208216, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1574230194091797, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7201696038246155, + "num_tokens": 615843106.0, + "step": 23806 + }, + { + "epoch": 2.6144300461234353, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2408995628356934, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7398255467414856, + "num_tokens": 615870852.0, + "step": 23807 + }, + { + "epoch": 2.6145398638260486, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.475992202758789, + "learning_rate": 1e-06, + "loss": 0.7562, + "mean_token_accuracy": 0.7564038038253784, + "num_tokens": 615893427.0, + "step": 23808 + }, + { + "epoch": 2.6146496815286624, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.227372407913208, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7065900564193726, + "num_tokens": 615922853.0, + "step": 23809 + }, + { + "epoch": 2.614759499231276, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.609792470932007, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7280213236808777, + "num_tokens": 615944594.0, + "step": 23810 + }, + { + "epoch": 2.61486931693389, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.548966407775879, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7361776232719421, + "num_tokens": 615968025.0, + "step": 23811 + }, + { + "epoch": 2.6149791346365037, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4704220294952393, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7098621726036072, + "num_tokens": 615996177.0, + "step": 23812 + }, + { + "epoch": 2.615088952339117, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3025481700897217, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.728744626045227, + "num_tokens": 616023342.0, + "step": 23813 + }, + { + "epoch": 2.6151987700417307, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.084160089492798, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7277321815490723, + "num_tokens": 616052964.0, + "step": 23814 + }, + { + "epoch": 2.6153085877443445, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5212864875793457, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7223025560379028, + "num_tokens": 616078639.0, + "step": 23815 + }, + { + "epoch": 2.615418405446958, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1534390449523926, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.737022876739502, + "num_tokens": 616108430.0, + "step": 23816 + }, + { + "epoch": 2.6155282231495716, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2680022716522217, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7434115409851074, + "num_tokens": 616132651.0, + "step": 23817 + }, + { + "epoch": 2.6156380408521853, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3180441856384277, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7135404348373413, + "num_tokens": 616160477.0, + "step": 23818 + }, + { + "epoch": 2.615747858554799, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.7396669387817383, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7390099763870239, + "num_tokens": 616181366.0, + "step": 23819 + }, + { + "epoch": 2.615857676257413, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.695554733276367, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7246106863021851, + "num_tokens": 616203339.0, + "step": 23820 + }, + { + "epoch": 2.615967493960026, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4227681159973145, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.74381422996521, + "num_tokens": 616226918.0, + "step": 23821 + }, + { + "epoch": 2.61607731166264, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3937416076660156, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7107781767845154, + "num_tokens": 616253500.0, + "step": 23822 + }, + { + "epoch": 2.6161871293652537, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5759849548339844, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7505735754966736, + "num_tokens": 616275443.0, + "step": 23823 + }, + { + "epoch": 2.6162969470678674, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.507850408554077, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7300712466239929, + "num_tokens": 616298483.0, + "step": 23824 + }, + { + "epoch": 2.616406764770481, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5439867973327637, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.730601966381073, + "num_tokens": 616321924.0, + "step": 23825 + }, + { + "epoch": 2.6165165824730945, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3491358757019043, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7277331352233887, + "num_tokens": 616349063.0, + "step": 23826 + }, + { + "epoch": 2.6166264001757082, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.250781297683716, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7240256071090698, + "num_tokens": 616375379.0, + "step": 23827 + }, + { + "epoch": 2.616736217878322, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.371547222137451, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7221283912658691, + "num_tokens": 616401438.0, + "step": 23828 + }, + { + "epoch": 2.6168460355809358, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.221888542175293, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7289357781410217, + "num_tokens": 616430323.0, + "step": 23829 + }, + { + "epoch": 2.6169558532835495, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2560510635375977, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7441568374633789, + "num_tokens": 616457365.0, + "step": 23830 + }, + { + "epoch": 2.617065670986163, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.354095697402954, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7369167804718018, + "num_tokens": 616484086.0, + "step": 23831 + }, + { + "epoch": 2.6171754886887766, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.119335651397705, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.714914083480835, + "num_tokens": 616515260.0, + "step": 23832 + }, + { + "epoch": 2.6172853063913903, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.075439691543579, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7222782969474792, + "num_tokens": 616545524.0, + "step": 23833 + }, + { + "epoch": 2.617395124094004, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.444289207458496, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7333551645278931, + "num_tokens": 616570629.0, + "step": 23834 + }, + { + "epoch": 2.617504941796618, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1808199882507324, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7277671098709106, + "num_tokens": 616600138.0, + "step": 23835 + }, + { + "epoch": 2.617614759499231, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.7499325275421143, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7166789770126343, + "num_tokens": 616621489.0, + "step": 23836 + }, + { + "epoch": 2.617724577201845, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.780519962310791, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.740100085735321, + "num_tokens": 616642458.0, + "step": 23837 + }, + { + "epoch": 2.6178343949044587, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.401761293411255, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7390214204788208, + "num_tokens": 616668748.0, + "step": 23838 + }, + { + "epoch": 2.617944212607072, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1159348487854004, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.72918301820755, + "num_tokens": 616700767.0, + "step": 23839 + }, + { + "epoch": 2.618054030309686, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.385401725769043, + "learning_rate": 1e-06, + "loss": 0.8133, + "mean_token_accuracy": 0.7486310601234436, + "num_tokens": 616725784.0, + "step": 23840 + }, + { + "epoch": 2.6181638480122995, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5162205696105957, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7512071132659912, + "num_tokens": 616749963.0, + "step": 23841 + }, + { + "epoch": 2.6182736657149133, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4684767723083496, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7319268584251404, + "num_tokens": 616773116.0, + "step": 23842 + }, + { + "epoch": 2.618383483417527, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4064581394195557, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7002297639846802, + "num_tokens": 616799662.0, + "step": 23843 + }, + { + "epoch": 2.6184933011201403, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1111252307891846, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7251090407371521, + "num_tokens": 616832181.0, + "step": 23844 + }, + { + "epoch": 2.618603118822754, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.521315336227417, + "learning_rate": 1e-06, + "loss": 0.7234, + "mean_token_accuracy": 0.7659134864807129, + "num_tokens": 616856032.0, + "step": 23845 + }, + { + "epoch": 2.618712936525368, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.406913995742798, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7237254977226257, + "num_tokens": 616882922.0, + "step": 23846 + }, + { + "epoch": 2.6188227542279816, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.579807996749878, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7361972332000732, + "num_tokens": 616905003.0, + "step": 23847 + }, + { + "epoch": 2.6189325719305954, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.29921555519104, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7270539402961731, + "num_tokens": 616934182.0, + "step": 23848 + }, + { + "epoch": 2.6190423896332087, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.402942419052124, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7132896184921265, + "num_tokens": 616962792.0, + "step": 23849 + }, + { + "epoch": 2.6191522073358224, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.211688280105591, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7213530540466309, + "num_tokens": 616993715.0, + "step": 23850 + }, + { + "epoch": 2.619262025038436, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5225350856781006, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7391819953918457, + "num_tokens": 617016131.0, + "step": 23851 + }, + { + "epoch": 2.61937184274105, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0452659130096436, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7171794176101685, + "num_tokens": 617047681.0, + "step": 23852 + }, + { + "epoch": 2.6194816604436637, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2671380043029785, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7469372749328613, + "num_tokens": 617074011.0, + "step": 23853 + }, + { + "epoch": 2.619591478146277, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.35973858833313, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7419069409370422, + "num_tokens": 617098524.0, + "step": 23854 + }, + { + "epoch": 2.6197012958488908, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2289528846740723, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7243146300315857, + "num_tokens": 617125190.0, + "step": 23855 + }, + { + "epoch": 2.6198111135515045, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4792277812957764, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.743691086769104, + "num_tokens": 617149480.0, + "step": 23856 + }, + { + "epoch": 2.6199209312541183, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3302114009857178, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7293909192085266, + "num_tokens": 617175751.0, + "step": 23857 + }, + { + "epoch": 2.620030748956732, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1485648155212402, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7268520593643188, + "num_tokens": 617207187.0, + "step": 23858 + }, + { + "epoch": 2.6201405666593454, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4699201583862305, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7352411150932312, + "num_tokens": 617230627.0, + "step": 23859 + }, + { + "epoch": 2.620250384361959, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4563205242156982, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.6969184875488281, + "num_tokens": 617255474.0, + "step": 23860 + }, + { + "epoch": 2.620360202064573, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2538230419158936, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7344685196876526, + "num_tokens": 617283950.0, + "step": 23861 + }, + { + "epoch": 2.6204700197671866, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.264991521835327, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7247213125228882, + "num_tokens": 617311662.0, + "step": 23862 + }, + { + "epoch": 2.6205798374698004, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2808539867401123, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.74554443359375, + "num_tokens": 617337524.0, + "step": 23863 + }, + { + "epoch": 2.6206896551724137, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.498277187347412, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7304511070251465, + "num_tokens": 617362717.0, + "step": 23864 + }, + { + "epoch": 2.6207994728750275, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3121869564056396, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7229773998260498, + "num_tokens": 617390882.0, + "step": 23865 + }, + { + "epoch": 2.620909290577641, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3389952182769775, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7132292985916138, + "num_tokens": 617419244.0, + "step": 23866 + }, + { + "epoch": 2.6210191082802545, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6578707695007324, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7406437397003174, + "num_tokens": 617440822.0, + "step": 23867 + }, + { + "epoch": 2.6211289259828683, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.196948766708374, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7223100662231445, + "num_tokens": 617472444.0, + "step": 23868 + }, + { + "epoch": 2.621238743685482, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2725822925567627, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.7410457730293274, + "num_tokens": 617498360.0, + "step": 23869 + }, + { + "epoch": 2.621348561388096, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.356562376022339, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.732550323009491, + "num_tokens": 617524776.0, + "step": 23870 + }, + { + "epoch": 2.6214583790907096, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0695576667785645, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7125058770179749, + "num_tokens": 617559133.0, + "step": 23871 + }, + { + "epoch": 2.621568196793323, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0254440307617188, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7201095819473267, + "num_tokens": 617591498.0, + "step": 23872 + }, + { + "epoch": 2.6216780144959366, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1896581649780273, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.72791588306427, + "num_tokens": 617622204.0, + "step": 23873 + }, + { + "epoch": 2.6217878321985504, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6097285747528076, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7211846709251404, + "num_tokens": 617645255.0, + "step": 23874 + }, + { + "epoch": 2.621897649901164, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3336451053619385, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7423524260520935, + "num_tokens": 617668651.0, + "step": 23875 + }, + { + "epoch": 2.622007467603778, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.393395185470581, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7039358019828796, + "num_tokens": 617693808.0, + "step": 23876 + }, + { + "epoch": 2.622117285306391, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.629660129547119, + "learning_rate": 1e-06, + "loss": 0.782, + "mean_token_accuracy": 0.7532833814620972, + "num_tokens": 617714789.0, + "step": 23877 + }, + { + "epoch": 2.622227103009005, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4548914432525635, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.729422390460968, + "num_tokens": 617740029.0, + "step": 23878 + }, + { + "epoch": 2.6223369207116187, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5843045711517334, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7447980642318726, + "num_tokens": 617763305.0, + "step": 23879 + }, + { + "epoch": 2.6224467384142325, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.431673526763916, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7534492015838623, + "num_tokens": 617788975.0, + "step": 23880 + }, + { + "epoch": 2.6225565561168462, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1137001514434814, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7255430221557617, + "num_tokens": 617821155.0, + "step": 23881 + }, + { + "epoch": 2.6226663738194596, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5925724506378174, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7535959482192993, + "num_tokens": 617843279.0, + "step": 23882 + }, + { + "epoch": 2.6227761915220733, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.258242130279541, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.71784508228302, + "num_tokens": 617871053.0, + "step": 23883 + }, + { + "epoch": 2.622886009224687, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0613462924957275, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.738856315612793, + "num_tokens": 617901241.0, + "step": 23884 + }, + { + "epoch": 2.622995826927301, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4718563556671143, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7122137546539307, + "num_tokens": 617925776.0, + "step": 23885 + }, + { + "epoch": 2.6231056446299146, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5533976554870605, + "learning_rate": 1e-06, + "loss": 0.7529, + "mean_token_accuracy": 0.7654078006744385, + "num_tokens": 617947516.0, + "step": 23886 + }, + { + "epoch": 2.623215462332528, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.8822877407073975, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7441097497940063, + "num_tokens": 617966419.0, + "step": 23887 + }, + { + "epoch": 2.6233252800351416, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4120519161224365, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7327293157577515, + "num_tokens": 617991109.0, + "step": 23888 + }, + { + "epoch": 2.6234350977377554, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.253221035003662, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7213358283042908, + "num_tokens": 618020680.0, + "step": 23889 + }, + { + "epoch": 2.6235449154403687, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4508156776428223, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.749825119972229, + "num_tokens": 618043607.0, + "step": 23890 + }, + { + "epoch": 2.623654733142983, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1866185665130615, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7278956174850464, + "num_tokens": 618072111.0, + "step": 23891 + }, + { + "epoch": 2.6237645508455962, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5475659370422363, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7263314723968506, + "num_tokens": 618095676.0, + "step": 23892 + }, + { + "epoch": 2.62387436854821, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2446374893188477, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7238047122955322, + "num_tokens": 618122311.0, + "step": 23893 + }, + { + "epoch": 2.6239841862508237, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4494481086730957, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7371327877044678, + "num_tokens": 618146648.0, + "step": 23894 + }, + { + "epoch": 2.624094003953437, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.611877679824829, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7493349313735962, + "num_tokens": 618167716.0, + "step": 23895 + }, + { + "epoch": 2.624203821656051, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.273393392562866, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7329703569412231, + "num_tokens": 618194484.0, + "step": 23896 + }, + { + "epoch": 2.6243136393586646, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.554933547973633, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7402883172035217, + "num_tokens": 618215614.0, + "step": 23897 + }, + { + "epoch": 2.6244234570612783, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4781785011291504, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7228013277053833, + "num_tokens": 618241722.0, + "step": 23898 + }, + { + "epoch": 2.624533274763892, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.266422986984253, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7251102924346924, + "num_tokens": 618270986.0, + "step": 23899 + }, + { + "epoch": 2.6246430924665054, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.376084089279175, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7346889972686768, + "num_tokens": 618298025.0, + "step": 23900 + }, + { + "epoch": 2.624752910169119, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.679997205734253, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7280781269073486, + "num_tokens": 618320589.0, + "step": 23901 + }, + { + "epoch": 2.624862727871733, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.461878538131714, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7567889094352722, + "num_tokens": 618343548.0, + "step": 23902 + }, + { + "epoch": 2.6249725455743467, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.537839412689209, + "learning_rate": 1e-06, + "loss": 0.7517, + "mean_token_accuracy": 0.7538523077964783, + "num_tokens": 618366871.0, + "step": 23903 + }, + { + "epoch": 2.6250823632769604, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.234703302383423, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7284501791000366, + "num_tokens": 618397442.0, + "step": 23904 + }, + { + "epoch": 2.6251921809795737, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1626484394073486, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7237544059753418, + "num_tokens": 618425474.0, + "step": 23905 + }, + { + "epoch": 2.6253019986821875, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.249525547027588, + "learning_rate": 1e-06, + "loss": 0.7477, + "mean_token_accuracy": 0.7568382024765015, + "num_tokens": 618451059.0, + "step": 23906 + }, + { + "epoch": 2.6254118163848013, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.192366361618042, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7176513671875, + "num_tokens": 618481342.0, + "step": 23907 + }, + { + "epoch": 2.625521634087415, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1308484077453613, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7350214719772339, + "num_tokens": 618512365.0, + "step": 23908 + }, + { + "epoch": 2.6256314517900288, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.238276720046997, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7348107695579529, + "num_tokens": 618540595.0, + "step": 23909 + }, + { + "epoch": 2.625741269492642, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6132264137268066, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7311109304428101, + "num_tokens": 618562278.0, + "step": 23910 + }, + { + "epoch": 2.625851087195256, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.370100975036621, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7380977869033813, + "num_tokens": 618587828.0, + "step": 23911 + }, + { + "epoch": 2.6259609048978696, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3008055686950684, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7170556783676147, + "num_tokens": 618615196.0, + "step": 23912 + }, + { + "epoch": 2.6260707226004834, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1312270164489746, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7527757883071899, + "num_tokens": 618643927.0, + "step": 23913 + }, + { + "epoch": 2.626180540303097, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0780608654022217, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7108601331710815, + "num_tokens": 618676520.0, + "step": 23914 + }, + { + "epoch": 2.6262903580057104, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.486842155456543, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7259415984153748, + "num_tokens": 618701042.0, + "step": 23915 + }, + { + "epoch": 2.626400175708324, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3298346996307373, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.740365743637085, + "num_tokens": 618727378.0, + "step": 23916 + }, + { + "epoch": 2.626509993410938, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.226020574569702, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7274956703186035, + "num_tokens": 618754896.0, + "step": 23917 + }, + { + "epoch": 2.6266198111135513, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5228147506713867, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7369599342346191, + "num_tokens": 618778671.0, + "step": 23918 + }, + { + "epoch": 2.626729628816165, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3189492225646973, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7215819358825684, + "num_tokens": 618804593.0, + "step": 23919 + }, + { + "epoch": 2.6268394465187788, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3931450843811035, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7317416667938232, + "num_tokens": 618829558.0, + "step": 23920 + }, + { + "epoch": 2.6269492642213925, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3608429431915283, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7288429737091064, + "num_tokens": 618856096.0, + "step": 23921 + }, + { + "epoch": 2.6270590819240063, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.167243242263794, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7237539291381836, + "num_tokens": 618887366.0, + "step": 23922 + }, + { + "epoch": 2.6271688996266196, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1183347702026367, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7223619818687439, + "num_tokens": 618918351.0, + "step": 23923 + }, + { + "epoch": 2.6272787173292333, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.549598217010498, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7242547273635864, + "num_tokens": 618944233.0, + "step": 23924 + }, + { + "epoch": 2.627388535031847, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4597384929656982, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7359914183616638, + "num_tokens": 618968428.0, + "step": 23925 + }, + { + "epoch": 2.627498352734461, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.78456449508667, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7466332912445068, + "num_tokens": 618989435.0, + "step": 23926 + }, + { + "epoch": 2.6276081704370746, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.548849582672119, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7244499325752258, + "num_tokens": 619011294.0, + "step": 23927 + }, + { + "epoch": 2.627717988139688, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.691154956817627, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.740214467048645, + "num_tokens": 619031707.0, + "step": 23928 + }, + { + "epoch": 2.6278278058423017, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.358532190322876, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.746665358543396, + "num_tokens": 619056051.0, + "step": 23929 + }, + { + "epoch": 2.6279376235449154, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4571826457977295, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7276439666748047, + "num_tokens": 619078179.0, + "step": 23930 + }, + { + "epoch": 2.628047441247529, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2197041511535645, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7287105321884155, + "num_tokens": 619107033.0, + "step": 23931 + }, + { + "epoch": 2.628157258950143, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.280435562133789, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7212170362472534, + "num_tokens": 619134208.0, + "step": 23932 + }, + { + "epoch": 2.6282670766527563, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.375406265258789, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.727477490901947, + "num_tokens": 619160739.0, + "step": 23933 + }, + { + "epoch": 2.62837689435537, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5013620853424072, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7276855111122131, + "num_tokens": 619182600.0, + "step": 23934 + }, + { + "epoch": 2.628486712057984, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.464303970336914, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7213281393051147, + "num_tokens": 619207172.0, + "step": 23935 + }, + { + "epoch": 2.6285965297605975, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3109588623046875, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7380993962287903, + "num_tokens": 619234152.0, + "step": 23936 + }, + { + "epoch": 2.6287063474632113, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.120633602142334, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7256669402122498, + "num_tokens": 619264240.0, + "step": 23937 + }, + { + "epoch": 2.6288161651658246, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2225096225738525, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7263017892837524, + "num_tokens": 619292047.0, + "step": 23938 + }, + { + "epoch": 2.6289259828684384, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5752503871917725, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7469916939735413, + "num_tokens": 619315084.0, + "step": 23939 + }, + { + "epoch": 2.629035800571052, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.927579879760742, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7266514301300049, + "num_tokens": 619335129.0, + "step": 23940 + }, + { + "epoch": 2.6291456182736654, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.252248764038086, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7253137230873108, + "num_tokens": 619362365.0, + "step": 23941 + }, + { + "epoch": 2.6292554359762796, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.474177360534668, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7233030200004578, + "num_tokens": 619385774.0, + "step": 23942 + }, + { + "epoch": 2.629365253678893, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.325347900390625, + "learning_rate": 1e-06, + "loss": 0.7405, + "mean_token_accuracy": 0.7628592848777771, + "num_tokens": 619411958.0, + "step": 23943 + }, + { + "epoch": 2.6294750713815067, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.128587245941162, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7128721475601196, + "num_tokens": 619446957.0, + "step": 23944 + }, + { + "epoch": 2.6295848890841205, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4434661865234375, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7313675880432129, + "num_tokens": 619477566.0, + "step": 23945 + }, + { + "epoch": 2.629694706786734, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.504950761795044, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7394447922706604, + "num_tokens": 619501131.0, + "step": 23946 + }, + { + "epoch": 2.6298045244893475, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.0028510093688965, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7272061705589294, + "num_tokens": 619535315.0, + "step": 23947 + }, + { + "epoch": 2.6299143421919613, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.621497869491577, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.735139787197113, + "num_tokens": 619556423.0, + "step": 23948 + }, + { + "epoch": 2.630024159894575, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1180288791656494, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7109723091125488, + "num_tokens": 619587941.0, + "step": 23949 + }, + { + "epoch": 2.630133977597189, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4370810985565186, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7354882955551147, + "num_tokens": 619612379.0, + "step": 23950 + }, + { + "epoch": 2.630243795299802, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5454835891723633, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7443685531616211, + "num_tokens": 619634800.0, + "step": 23951 + }, + { + "epoch": 2.630353613002416, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2982747554779053, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7318412065505981, + "num_tokens": 619661653.0, + "step": 23952 + }, + { + "epoch": 2.6304634307050296, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2428884506225586, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.732528805732727, + "num_tokens": 619688426.0, + "step": 23953 + }, + { + "epoch": 2.6305732484076434, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.515846014022827, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7292311191558838, + "num_tokens": 619711954.0, + "step": 23954 + }, + { + "epoch": 2.630683066110257, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.202255964279175, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7182877063751221, + "num_tokens": 619740104.0, + "step": 23955 + }, + { + "epoch": 2.6307928838128705, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.416926383972168, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7273937463760376, + "num_tokens": 619763446.0, + "step": 23956 + }, + { + "epoch": 2.630902701515484, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.409322500228882, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.712058424949646, + "num_tokens": 619789381.0, + "step": 23957 + }, + { + "epoch": 2.631012519218098, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4805474281311035, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7329803109169006, + "num_tokens": 619815071.0, + "step": 23958 + }, + { + "epoch": 2.6311223369207117, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.5566656589508057, + "learning_rate": 1e-06, + "loss": 0.7969, + "mean_token_accuracy": 0.7447306513786316, + "num_tokens": 619836300.0, + "step": 23959 + }, + { + "epoch": 2.6312321546233255, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.342127561569214, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7158957123756409, + "num_tokens": 619864827.0, + "step": 23960 + }, + { + "epoch": 2.631341972325939, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.644979476928711, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7454051971435547, + "num_tokens": 619886514.0, + "step": 23961 + }, + { + "epoch": 2.6314517900285526, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.353759765625, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7316334843635559, + "num_tokens": 619911601.0, + "step": 23962 + }, + { + "epoch": 2.6315616077311663, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.7138922214508057, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7367365956306458, + "num_tokens": 619933542.0, + "step": 23963 + }, + { + "epoch": 2.63167142543378, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.350013494491577, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7328147292137146, + "num_tokens": 619962953.0, + "step": 23964 + }, + { + "epoch": 2.631781243136394, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.362635374069214, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7286795377731323, + "num_tokens": 619990685.0, + "step": 23965 + }, + { + "epoch": 2.631891060839007, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4397497177124023, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7287030220031738, + "num_tokens": 620016539.0, + "step": 23966 + }, + { + "epoch": 2.632000878541621, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.044325828552246, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7181004285812378, + "num_tokens": 620048256.0, + "step": 23967 + }, + { + "epoch": 2.6321106962442347, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6749258041381836, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7349698543548584, + "num_tokens": 620071741.0, + "step": 23968 + }, + { + "epoch": 2.632220513946848, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.4274468421936035, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7266762256622314, + "num_tokens": 620096596.0, + "step": 23969 + }, + { + "epoch": 2.6323303316494617, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.6726877689361572, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.734214186668396, + "num_tokens": 620118757.0, + "step": 23970 + }, + { + "epoch": 2.6324401493520755, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.27299427986145, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.6984696984291077, + "num_tokens": 620152284.0, + "step": 23971 + }, + { + "epoch": 2.6325499670546892, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.628828287124634, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7247449159622192, + "num_tokens": 620174262.0, + "step": 23972 + }, + { + "epoch": 2.632659784757303, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.214578866958618, + "learning_rate": 1e-06, + "loss": 1.0083, + "mean_token_accuracy": 0.686906099319458, + "num_tokens": 620205473.0, + "step": 23973 + }, + { + "epoch": 2.6327696024599163, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3146727085113525, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7142859697341919, + "num_tokens": 620231736.0, + "step": 23974 + }, + { + "epoch": 2.63287942016253, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.243849992752075, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7509858012199402, + "num_tokens": 620260991.0, + "step": 23975 + }, + { + "epoch": 2.632989237865144, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.1820387840270996, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7304201126098633, + "num_tokens": 620288854.0, + "step": 23976 + }, + { + "epoch": 2.6330990555677576, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.2870168685913086, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.708667516708374, + "num_tokens": 620317596.0, + "step": 23977 + }, + { + "epoch": 2.6332088732703713, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.3439290523529053, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7552441954612732, + "num_tokens": 620341641.0, + "step": 23978 + }, + { + "epoch": 2.6333186909729847, + "ewc_loss": 0.00021076202392578125, + "grad_norm": 2.19980788230896, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7312093377113342, + "num_tokens": 620371222.0, + "step": 23979 + }, + { + "epoch": 2.6334285086755984, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.47171950340271, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7421789765357971, + "num_tokens": 620395234.0, + "step": 23980 + }, + { + "epoch": 2.633538326378212, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.200129747390747, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7325893044471741, + "num_tokens": 620423716.0, + "step": 23981 + }, + { + "epoch": 2.633648144080826, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.676309823989868, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7296137809753418, + "num_tokens": 620445412.0, + "step": 23982 + }, + { + "epoch": 2.6337579617834397, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5186853408813477, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7502533197402954, + "num_tokens": 620468398.0, + "step": 23983 + }, + { + "epoch": 2.633867779486053, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.417607545852661, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7313594222068787, + "num_tokens": 620492287.0, + "step": 23984 + }, + { + "epoch": 2.6339775971886668, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.514908790588379, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7392706871032715, + "num_tokens": 620514595.0, + "step": 23985 + }, + { + "epoch": 2.6340874148912805, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2127058506011963, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7324808239936829, + "num_tokens": 620543280.0, + "step": 23986 + }, + { + "epoch": 2.6341972325938943, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.6008810997009277, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7303855419158936, + "num_tokens": 620564893.0, + "step": 23987 + }, + { + "epoch": 2.634307050296508, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.363473415374756, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7168519496917725, + "num_tokens": 620592556.0, + "step": 23988 + }, + { + "epoch": 2.6344168679991213, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.7559220790863037, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7518007755279541, + "num_tokens": 620612564.0, + "step": 23989 + }, + { + "epoch": 2.634526685701735, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3606340885162354, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7435543537139893, + "num_tokens": 620637624.0, + "step": 23990 + }, + { + "epoch": 2.634636503404349, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4071638584136963, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.726605236530304, + "num_tokens": 620662855.0, + "step": 23991 + }, + { + "epoch": 2.6347463211069626, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.556264877319336, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7499116063117981, + "num_tokens": 620684595.0, + "step": 23992 + }, + { + "epoch": 2.6348561388095764, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4268102645874023, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7302258014678955, + "num_tokens": 620710732.0, + "step": 23993 + }, + { + "epoch": 2.6349659565121897, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.198215961456299, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7103939056396484, + "num_tokens": 620740712.0, + "step": 23994 + }, + { + "epoch": 2.6350757742148034, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.434704065322876, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7354990243911743, + "num_tokens": 620764298.0, + "step": 23995 + }, + { + "epoch": 2.635185591917417, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2813382148742676, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.721966028213501, + "num_tokens": 620790394.0, + "step": 23996 + }, + { + "epoch": 2.6352954096200305, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2906062602996826, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7116579413414001, + "num_tokens": 620817893.0, + "step": 23997 + }, + { + "epoch": 2.6354052273226443, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.13330340385437, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7453088760375977, + "num_tokens": 620844533.0, + "step": 23998 + }, + { + "epoch": 2.635515045025258, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.728010416030884, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7247250080108643, + "num_tokens": 620867060.0, + "step": 23999 + }, + { + "epoch": 2.6356248627278718, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.324159860610962, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7293562293052673, + "num_tokens": 620893365.0, + "step": 24000 + }, + { + "epoch": 2.6357346804304855, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.262206792831421, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.742834746837616, + "num_tokens": 620920208.0, + "step": 24001 + }, + { + "epoch": 2.635844498133099, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.652893543243408, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7228525876998901, + "num_tokens": 620942974.0, + "step": 24002 + }, + { + "epoch": 2.6359543158357126, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2691256999969482, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7157487869262695, + "num_tokens": 620970769.0, + "step": 24003 + }, + { + "epoch": 2.6360641335383264, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.146740436553955, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.704718828201294, + "num_tokens": 621003008.0, + "step": 24004 + }, + { + "epoch": 2.63617395124094, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.210573196411133, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7257287502288818, + "num_tokens": 621033307.0, + "step": 24005 + }, + { + "epoch": 2.636283768943554, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1623904705047607, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.727544903755188, + "num_tokens": 621063311.0, + "step": 24006 + }, + { + "epoch": 2.636393586646167, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2824089527130127, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7242275476455688, + "num_tokens": 621091081.0, + "step": 24007 + }, + { + "epoch": 2.636503404348781, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.8616220951080322, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7162379026412964, + "num_tokens": 621111113.0, + "step": 24008 + }, + { + "epoch": 2.6366132220513947, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3737409114837646, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7265447378158569, + "num_tokens": 621137657.0, + "step": 24009 + }, + { + "epoch": 2.6367230397540085, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.679921865463257, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.746422290802002, + "num_tokens": 621157945.0, + "step": 24010 + }, + { + "epoch": 2.636832857456622, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.713979721069336, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7149292826652527, + "num_tokens": 621178900.0, + "step": 24011 + }, + { + "epoch": 2.6369426751592355, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3941168785095215, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7321174740791321, + "num_tokens": 621204561.0, + "step": 24012 + }, + { + "epoch": 2.6370524928618493, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.584808826446533, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7339814305305481, + "num_tokens": 621226032.0, + "step": 24013 + }, + { + "epoch": 2.637162310564463, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.783430814743042, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7238656878471375, + "num_tokens": 621245472.0, + "step": 24014 + }, + { + "epoch": 2.637272128267077, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.575918436050415, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7227563858032227, + "num_tokens": 621266922.0, + "step": 24015 + }, + { + "epoch": 2.6373819459696906, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2484700679779053, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7205288410186768, + "num_tokens": 621296209.0, + "step": 24016 + }, + { + "epoch": 2.637491763672304, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.6628401279449463, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7171701192855835, + "num_tokens": 621318693.0, + "step": 24017 + }, + { + "epoch": 2.6376015813749176, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2785563468933105, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.6983656883239746, + "num_tokens": 621347752.0, + "step": 24018 + }, + { + "epoch": 2.6377113990775314, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3715336322784424, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7576562166213989, + "num_tokens": 621373631.0, + "step": 24019 + }, + { + "epoch": 2.6378212167801447, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4874684810638428, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7406004667282104, + "num_tokens": 621398262.0, + "step": 24020 + }, + { + "epoch": 2.637931034482759, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2836244106292725, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7549535036087036, + "num_tokens": 621423360.0, + "step": 24021 + }, + { + "epoch": 2.638040852185372, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2348711490631104, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7325630187988281, + "num_tokens": 621451716.0, + "step": 24022 + }, + { + "epoch": 2.638150669887986, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5464553833007812, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.71979159116745, + "num_tokens": 621475198.0, + "step": 24023 + }, + { + "epoch": 2.6382604875905997, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4936330318450928, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7274062633514404, + "num_tokens": 621499649.0, + "step": 24024 + }, + { + "epoch": 2.638370305293213, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3503715991973877, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7395142316818237, + "num_tokens": 621524844.0, + "step": 24025 + }, + { + "epoch": 2.638480122995827, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1787302494049072, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7042205929756165, + "num_tokens": 621556794.0, + "step": 24026 + }, + { + "epoch": 2.6385899406984406, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.760394811630249, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7305613160133362, + "num_tokens": 621576931.0, + "step": 24027 + }, + { + "epoch": 2.6386997584010543, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.236663818359375, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7245786786079407, + "num_tokens": 621603944.0, + "step": 24028 + }, + { + "epoch": 2.638809576103668, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1847100257873535, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.713234543800354, + "num_tokens": 621635189.0, + "step": 24029 + }, + { + "epoch": 2.6389193938062814, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.649017333984375, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7508227229118347, + "num_tokens": 621656516.0, + "step": 24030 + }, + { + "epoch": 2.639029211508895, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4066429138183594, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7291511297225952, + "num_tokens": 621681421.0, + "step": 24031 + }, + { + "epoch": 2.639139029211509, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.421009063720703, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7014440298080444, + "num_tokens": 621705867.0, + "step": 24032 + }, + { + "epoch": 2.6392488469141226, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2537734508514404, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7349061369895935, + "num_tokens": 621731823.0, + "step": 24033 + }, + { + "epoch": 2.6393586646167364, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.8864495754241943, + "learning_rate": 1e-06, + "loss": 0.7708, + "mean_token_accuracy": 0.7484750151634216, + "num_tokens": 621751871.0, + "step": 24034 + }, + { + "epoch": 2.6394684823193497, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.242518424987793, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7317087650299072, + "num_tokens": 621778959.0, + "step": 24035 + }, + { + "epoch": 2.6395783000219635, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.313328981399536, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7185169458389282, + "num_tokens": 621806512.0, + "step": 24036 + }, + { + "epoch": 2.6396881177245772, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 3.002141237258911, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7424707412719727, + "num_tokens": 621823949.0, + "step": 24037 + }, + { + "epoch": 2.639797935427191, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2036304473876953, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7418454885482788, + "num_tokens": 621851867.0, + "step": 24038 + }, + { + "epoch": 2.6399077531298047, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3431484699249268, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7354764938354492, + "num_tokens": 621877830.0, + "step": 24039 + }, + { + "epoch": 2.640017570832418, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.377121686935425, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7139729261398315, + "num_tokens": 621904075.0, + "step": 24040 + }, + { + "epoch": 2.640127388535032, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.770102024078369, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7467700242996216, + "num_tokens": 621923382.0, + "step": 24041 + }, + { + "epoch": 2.6402372062376456, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.26391863822937, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7294973134994507, + "num_tokens": 621953238.0, + "step": 24042 + }, + { + "epoch": 2.6403470239402593, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.8326287269592285, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7421184778213501, + "num_tokens": 621972582.0, + "step": 24043 + }, + { + "epoch": 2.640456841642873, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4630424976348877, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7297540903091431, + "num_tokens": 621997752.0, + "step": 24044 + }, + { + "epoch": 2.6405666593454864, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.624502658843994, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7412276864051819, + "num_tokens": 622020119.0, + "step": 24045 + }, + { + "epoch": 2.6406764770481, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.595621347427368, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7339385747909546, + "num_tokens": 622043496.0, + "step": 24046 + }, + { + "epoch": 2.640786294750714, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3051679134368896, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7175265550613403, + "num_tokens": 622070300.0, + "step": 24047 + }, + { + "epoch": 2.6408961124533272, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5372908115386963, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7281949520111084, + "num_tokens": 622094411.0, + "step": 24048 + }, + { + "epoch": 2.641005930155941, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.0716934204101562, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7073038816452026, + "num_tokens": 622128698.0, + "step": 24049 + }, + { + "epoch": 2.6411157478585547, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.615084409713745, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7320083379745483, + "num_tokens": 622148936.0, + "step": 24050 + }, + { + "epoch": 2.6412255655611685, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.199692726135254, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.734069287776947, + "num_tokens": 622176301.0, + "step": 24051 + }, + { + "epoch": 2.6413353832637823, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.0808777809143066, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7121126651763916, + "num_tokens": 622208911.0, + "step": 24052 + }, + { + "epoch": 2.6414452009663956, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3516268730163574, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7136385440826416, + "num_tokens": 622236356.0, + "step": 24053 + }, + { + "epoch": 2.6415550186690093, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3277478218078613, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7274969816207886, + "num_tokens": 622262816.0, + "step": 24054 + }, + { + "epoch": 2.641664836371623, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.576185941696167, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7147434949874878, + "num_tokens": 622284963.0, + "step": 24055 + }, + { + "epoch": 2.641774654074237, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.708515167236328, + "learning_rate": 1e-06, + "loss": 0.7562, + "mean_token_accuracy": 0.7562664151191711, + "num_tokens": 622305749.0, + "step": 24056 + }, + { + "epoch": 2.6418844717768506, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.348802328109741, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.7491825222969055, + "num_tokens": 622331647.0, + "step": 24057 + }, + { + "epoch": 2.641994289479464, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.500054359436035, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7223374843597412, + "num_tokens": 622354461.0, + "step": 24058 + }, + { + "epoch": 2.6421041071820777, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4606518745422363, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7231168150901794, + "num_tokens": 622378606.0, + "step": 24059 + }, + { + "epoch": 2.6422139248846914, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3161191940307617, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.730087399482727, + "num_tokens": 622404972.0, + "step": 24060 + }, + { + "epoch": 2.642323742587305, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1843035221099854, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7142064571380615, + "num_tokens": 622437224.0, + "step": 24061 + }, + { + "epoch": 2.642433560289919, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.095320463180542, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6902048587799072, + "num_tokens": 622470711.0, + "step": 24062 + }, + { + "epoch": 2.6425433779925322, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4772605895996094, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7436522841453552, + "num_tokens": 622495174.0, + "step": 24063 + }, + { + "epoch": 2.642653195695146, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5375583171844482, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.749205470085144, + "num_tokens": 622518425.0, + "step": 24064 + }, + { + "epoch": 2.6427630133977598, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1181905269622803, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7196875810623169, + "num_tokens": 622549915.0, + "step": 24065 + }, + { + "epoch": 2.6428728311003735, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.607534408569336, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7032609581947327, + "num_tokens": 622572444.0, + "step": 24066 + }, + { + "epoch": 2.6429826488029873, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.6910431385040283, + "learning_rate": 1e-06, + "loss": 0.7793, + "mean_token_accuracy": 0.7483153939247131, + "num_tokens": 622593190.0, + "step": 24067 + }, + { + "epoch": 2.6430924665056006, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4718761444091797, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7281579971313477, + "num_tokens": 622620392.0, + "step": 24068 + }, + { + "epoch": 2.6432022842082143, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.317617177963257, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7248232960700989, + "num_tokens": 622649585.0, + "step": 24069 + }, + { + "epoch": 2.643312101910828, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1937780380249023, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7369815707206726, + "num_tokens": 622676984.0, + "step": 24070 + }, + { + "epoch": 2.6434219196134414, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.212341547012329, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7199475169181824, + "num_tokens": 622705084.0, + "step": 24071 + }, + { + "epoch": 2.6435317373160556, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4068644046783447, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7247490882873535, + "num_tokens": 622732563.0, + "step": 24072 + }, + { + "epoch": 2.643641555018669, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.581857681274414, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7451490759849548, + "num_tokens": 622757536.0, + "step": 24073 + }, + { + "epoch": 2.6437513727212827, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.1974449157714844, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7256432771682739, + "num_tokens": 622787075.0, + "step": 24074 + }, + { + "epoch": 2.6438611904238964, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4817516803741455, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.724321722984314, + "num_tokens": 622811946.0, + "step": 24075 + }, + { + "epoch": 2.6439710081265098, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.287773847579956, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7326978445053101, + "num_tokens": 622839809.0, + "step": 24076 + }, + { + "epoch": 2.6440808258291235, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.739401340484619, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7416476011276245, + "num_tokens": 622858369.0, + "step": 24077 + }, + { + "epoch": 2.6441906435317373, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.6112375259399414, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7417380809783936, + "num_tokens": 622879353.0, + "step": 24078 + }, + { + "epoch": 2.644300461234351, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.555773973464966, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7138240337371826, + "num_tokens": 622906053.0, + "step": 24079 + }, + { + "epoch": 2.644410278936965, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.233987331390381, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7194037437438965, + "num_tokens": 622935393.0, + "step": 24080 + }, + { + "epoch": 2.644520096639578, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.36338472366333, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7213219404220581, + "num_tokens": 622961271.0, + "step": 24081 + }, + { + "epoch": 2.644629914342192, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4665448665618896, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7447209358215332, + "num_tokens": 622988400.0, + "step": 24082 + }, + { + "epoch": 2.6447397320448056, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.5870561599731445, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7282511591911316, + "num_tokens": 623011022.0, + "step": 24083 + }, + { + "epoch": 2.6448495497474194, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.7364723682403564, + "learning_rate": 1e-06, + "loss": 0.754, + "mean_token_accuracy": 0.7683950066566467, + "num_tokens": 623030796.0, + "step": 24084 + }, + { + "epoch": 2.644959367450033, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.151918411254883, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7146446704864502, + "num_tokens": 623060043.0, + "step": 24085 + }, + { + "epoch": 2.6450691851526464, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.387277841567993, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7475077509880066, + "num_tokens": 623082033.0, + "step": 24086 + }, + { + "epoch": 2.64517900285526, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.3868603706359863, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.717272937297821, + "num_tokens": 623109445.0, + "step": 24087 + }, + { + "epoch": 2.645288820557874, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2587883472442627, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7150914669036865, + "num_tokens": 623138438.0, + "step": 24088 + }, + { + "epoch": 2.6453986382604877, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.4144296646118164, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7392760515213013, + "num_tokens": 623162892.0, + "step": 24089 + }, + { + "epoch": 2.6455084559631015, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 3.0303666591644287, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7426469326019287, + "num_tokens": 623181200.0, + "step": 24090 + }, + { + "epoch": 2.645618273665715, + "ewc_loss": 0.0002117156982421875, + "grad_norm": 2.2910609245300293, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7285777926445007, + "num_tokens": 623209572.0, + "step": 24091 + }, + { + "epoch": 2.6457280913683285, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.548715829849243, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7173832058906555, + "num_tokens": 623233462.0, + "step": 24092 + }, + { + "epoch": 2.6458379090709423, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.229560136795044, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7277947068214417, + "num_tokens": 623260152.0, + "step": 24093 + }, + { + "epoch": 2.645947726773556, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.137291193008423, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7204570174217224, + "num_tokens": 623290976.0, + "step": 24094 + }, + { + "epoch": 2.64605754447617, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.566295623779297, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7131890654563904, + "num_tokens": 623314911.0, + "step": 24095 + }, + { + "epoch": 2.646167362178783, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3013224601745605, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7332303524017334, + "num_tokens": 623342413.0, + "step": 24096 + }, + { + "epoch": 2.646277179881397, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1759443283081055, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7142041921615601, + "num_tokens": 623371308.0, + "step": 24097 + }, + { + "epoch": 2.6463869975840106, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.131204128265381, + "learning_rate": 1e-06, + "loss": 0.7558, + "mean_token_accuracy": 0.7590975761413574, + "num_tokens": 623398913.0, + "step": 24098 + }, + { + "epoch": 2.646496815286624, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.149815559387207, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7122292518615723, + "num_tokens": 623430023.0, + "step": 24099 + }, + { + "epoch": 2.6466066329892377, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1099815368652344, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7158425450325012, + "num_tokens": 623461192.0, + "step": 24100 + }, + { + "epoch": 2.6467164506918515, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4010298252105713, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7413555383682251, + "num_tokens": 623487347.0, + "step": 24101 + }, + { + "epoch": 2.646826268394465, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.412034273147583, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7195968627929688, + "num_tokens": 623511378.0, + "step": 24102 + }, + { + "epoch": 2.646936086097079, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4296910762786865, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7209604382514954, + "num_tokens": 623535263.0, + "step": 24103 + }, + { + "epoch": 2.6470459037996923, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3668465614318848, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7193230390548706, + "num_tokens": 623562398.0, + "step": 24104 + }, + { + "epoch": 2.647155721502306, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.553105354309082, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.732016384601593, + "num_tokens": 623586222.0, + "step": 24105 + }, + { + "epoch": 2.64726553920492, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.697004795074463, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.725016176700592, + "num_tokens": 623608154.0, + "step": 24106 + }, + { + "epoch": 2.6473753569075336, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4728736877441406, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7049657106399536, + "num_tokens": 623634219.0, + "step": 24107 + }, + { + "epoch": 2.6474851746101473, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.422412633895874, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.724799394607544, + "num_tokens": 623661239.0, + "step": 24108 + }, + { + "epoch": 2.6475949923127606, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.224813461303711, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.7511739730834961, + "num_tokens": 623687698.0, + "step": 24109 + }, + { + "epoch": 2.6477048100153744, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4464192390441895, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7320653200149536, + "num_tokens": 623712239.0, + "step": 24110 + }, + { + "epoch": 2.647814627717988, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.5962600708007812, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7464304566383362, + "num_tokens": 623734006.0, + "step": 24111 + }, + { + "epoch": 2.647924445420602, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.317284345626831, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7242700457572937, + "num_tokens": 623761280.0, + "step": 24112 + }, + { + "epoch": 2.6480342631232157, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2876784801483154, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7185161709785461, + "num_tokens": 623788872.0, + "step": 24113 + }, + { + "epoch": 2.648144080825829, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3825793266296387, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.708571195602417, + "num_tokens": 623815823.0, + "step": 24114 + }, + { + "epoch": 2.6482538985284427, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.436708450317383, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7265428304672241, + "num_tokens": 623841142.0, + "step": 24115 + }, + { + "epoch": 2.6483637162310565, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.300507068634033, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6941494941711426, + "num_tokens": 623868966.0, + "step": 24116 + }, + { + "epoch": 2.6484735339336702, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.560875177383423, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7343544363975525, + "num_tokens": 623889833.0, + "step": 24117 + }, + { + "epoch": 2.648583351636284, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3953614234924316, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7144160270690918, + "num_tokens": 623916392.0, + "step": 24118 + }, + { + "epoch": 2.6486931693388973, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.388876438140869, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7278820276260376, + "num_tokens": 623941110.0, + "step": 24119 + }, + { + "epoch": 2.648802987041511, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.353774309158325, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.739566445350647, + "num_tokens": 623965608.0, + "step": 24120 + }, + { + "epoch": 2.648912804744125, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.409001111984253, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.730260968208313, + "num_tokens": 623992636.0, + "step": 24121 + }, + { + "epoch": 2.649022622446738, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3839504718780518, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7432770729064941, + "num_tokens": 624016363.0, + "step": 24122 + }, + { + "epoch": 2.6491324401493523, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.282052755355835, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.744961142539978, + "num_tokens": 624043771.0, + "step": 24123 + }, + { + "epoch": 2.6492422578519657, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2080130577087402, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7220178842544556, + "num_tokens": 624072386.0, + "step": 24124 + }, + { + "epoch": 2.6493520755545794, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3800888061523438, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7293497323989868, + "num_tokens": 624097076.0, + "step": 24125 + }, + { + "epoch": 2.649461893257193, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.404599189758301, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7480472326278687, + "num_tokens": 624120131.0, + "step": 24126 + }, + { + "epoch": 2.6495717109598065, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3083910942077637, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7399643659591675, + "num_tokens": 624148254.0, + "step": 24127 + }, + { + "epoch": 2.6496815286624202, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2248408794403076, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7128095030784607, + "num_tokens": 624176985.0, + "step": 24128 + }, + { + "epoch": 2.649791346365034, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4397542476654053, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7328612804412842, + "num_tokens": 624202640.0, + "step": 24129 + }, + { + "epoch": 2.6499011640676478, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.5027542114257812, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7466776371002197, + "num_tokens": 624225318.0, + "step": 24130 + }, + { + "epoch": 2.6500109817702615, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.49955153465271, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7333942651748657, + "num_tokens": 624249215.0, + "step": 24131 + }, + { + "epoch": 2.650120799472875, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.569852828979492, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7339715957641602, + "num_tokens": 624271947.0, + "step": 24132 + }, + { + "epoch": 2.6502306171754886, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4289462566375732, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.736586332321167, + "num_tokens": 624294819.0, + "step": 24133 + }, + { + "epoch": 2.6503404348781023, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.490374803543091, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.730781078338623, + "num_tokens": 624319693.0, + "step": 24134 + }, + { + "epoch": 2.650450252580716, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.0931103229522705, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7284273505210876, + "num_tokens": 624350989.0, + "step": 24135 + }, + { + "epoch": 2.65056007028333, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1678378582000732, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7553998231887817, + "num_tokens": 624380414.0, + "step": 24136 + }, + { + "epoch": 2.650669887985943, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3691627979278564, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7190922498703003, + "num_tokens": 624406399.0, + "step": 24137 + }, + { + "epoch": 2.650779705688557, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.281851291656494, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7087012529373169, + "num_tokens": 624433901.0, + "step": 24138 + }, + { + "epoch": 2.6508895233911707, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.402621269226074, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7399560809135437, + "num_tokens": 624458660.0, + "step": 24139 + }, + { + "epoch": 2.6509993410937844, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.7088544368743896, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.735475480556488, + "num_tokens": 624481456.0, + "step": 24140 + }, + { + "epoch": 2.651109158796398, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.5885159969329834, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7313469648361206, + "num_tokens": 624504676.0, + "step": 24141 + }, + { + "epoch": 2.6512189764990115, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2095658779144287, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6953845024108887, + "num_tokens": 624535033.0, + "step": 24142 + }, + { + "epoch": 2.6513287942016253, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6044552326202393, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7411383390426636, + "num_tokens": 624557535.0, + "step": 24143 + }, + { + "epoch": 2.651438611904239, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.302900791168213, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7305848598480225, + "num_tokens": 624583930.0, + "step": 24144 + }, + { + "epoch": 2.6515484296068528, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.614696979522705, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7309406995773315, + "num_tokens": 624607216.0, + "step": 24145 + }, + { + "epoch": 2.6516582473094665, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4130609035491943, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.744550347328186, + "num_tokens": 624629725.0, + "step": 24146 + }, + { + "epoch": 2.65176806501208, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3648648262023926, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7264530658721924, + "num_tokens": 624654995.0, + "step": 24147 + }, + { + "epoch": 2.6518778827146936, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.553354263305664, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7502654790878296, + "num_tokens": 624677595.0, + "step": 24148 + }, + { + "epoch": 2.6519877004173074, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4028375148773193, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7152940034866333, + "num_tokens": 624705954.0, + "step": 24149 + }, + { + "epoch": 2.6520975181199207, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2312586307525635, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7154381275177002, + "num_tokens": 624735000.0, + "step": 24150 + }, + { + "epoch": 2.6522073358225344, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.7309958934783936, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7197598218917847, + "num_tokens": 624756511.0, + "step": 24151 + }, + { + "epoch": 2.652317153525148, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2151901721954346, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7121519446372986, + "num_tokens": 624784308.0, + "step": 24152 + }, + { + "epoch": 2.652426971227762, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.494901180267334, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7281860113143921, + "num_tokens": 624809534.0, + "step": 24153 + }, + { + "epoch": 2.6525367889303757, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3594682216644287, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.743668794631958, + "num_tokens": 624834936.0, + "step": 24154 + }, + { + "epoch": 2.652646606632989, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1854662895202637, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.714002788066864, + "num_tokens": 624864487.0, + "step": 24155 + }, + { + "epoch": 2.6527564243356028, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4644615650177, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7103713750839233, + "num_tokens": 624889573.0, + "step": 24156 + }, + { + "epoch": 2.6528662420382165, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.252699375152588, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7071435451507568, + "num_tokens": 624918369.0, + "step": 24157 + }, + { + "epoch": 2.6529760597408303, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3589138984680176, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7360990047454834, + "num_tokens": 624944122.0, + "step": 24158 + }, + { + "epoch": 2.653085877443444, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6076676845550537, + "learning_rate": 1e-06, + "loss": 0.7494, + "mean_token_accuracy": 0.7597048282623291, + "num_tokens": 624964762.0, + "step": 24159 + }, + { + "epoch": 2.6531956951460574, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.415161371231079, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7171547412872314, + "num_tokens": 624989315.0, + "step": 24160 + }, + { + "epoch": 2.653305512848671, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.43400239944458, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7098730206489563, + "num_tokens": 625016020.0, + "step": 24161 + }, + { + "epoch": 2.653415330551285, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.072028398513794, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7302497625350952, + "num_tokens": 625046600.0, + "step": 24162 + }, + { + "epoch": 2.6535251482538986, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3405187129974365, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7273567318916321, + "num_tokens": 625073179.0, + "step": 24163 + }, + { + "epoch": 2.6536349659565124, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2175683975219727, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7444941997528076, + "num_tokens": 625100817.0, + "step": 24164 + }, + { + "epoch": 2.6537447836591257, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.090576410293579, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.705446183681488, + "num_tokens": 625132091.0, + "step": 24165 + }, + { + "epoch": 2.6538546013617395, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.5885727405548096, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7295841574668884, + "num_tokens": 625156293.0, + "step": 24166 + }, + { + "epoch": 2.653964419064353, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.27639102935791, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7088294625282288, + "num_tokens": 625185119.0, + "step": 24167 + }, + { + "epoch": 2.654074236766967, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.229942798614502, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7194150686264038, + "num_tokens": 625216115.0, + "step": 24168 + }, + { + "epoch": 2.6541840544695807, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.300075054168701, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7287131547927856, + "num_tokens": 625243558.0, + "step": 24169 + }, + { + "epoch": 2.654293872172194, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.53155517578125, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7363817095756531, + "num_tokens": 625266309.0, + "step": 24170 + }, + { + "epoch": 2.654403689874808, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1783273220062256, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7118625640869141, + "num_tokens": 625295082.0, + "step": 24171 + }, + { + "epoch": 2.6545135075774215, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.512789249420166, + "learning_rate": 1e-06, + "loss": 0.782, + "mean_token_accuracy": 0.7498255968093872, + "num_tokens": 625319200.0, + "step": 24172 + }, + { + "epoch": 2.6546233252800353, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3686821460723877, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7542536854743958, + "num_tokens": 625344812.0, + "step": 24173 + }, + { + "epoch": 2.654733142982649, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.272284746170044, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7310892343521118, + "num_tokens": 625371736.0, + "step": 24174 + }, + { + "epoch": 2.6548429606852624, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2540266513824463, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7372136116027832, + "num_tokens": 625400843.0, + "step": 24175 + }, + { + "epoch": 2.654952778387876, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.78240704536438, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7236305475234985, + "num_tokens": 625421941.0, + "step": 24176 + }, + { + "epoch": 2.65506259609049, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.7328107357025146, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7338800430297852, + "num_tokens": 625443964.0, + "step": 24177 + }, + { + "epoch": 2.655172413793103, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.7186453342437744, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7455433011054993, + "num_tokens": 625463927.0, + "step": 24178 + }, + { + "epoch": 2.655282231495717, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.18501353263855, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7366368770599365, + "num_tokens": 625492348.0, + "step": 24179 + }, + { + "epoch": 2.6553920491983307, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6124143600463867, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.738686203956604, + "num_tokens": 625513604.0, + "step": 24180 + }, + { + "epoch": 2.6555018669009445, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6793341636657715, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7317534685134888, + "num_tokens": 625534927.0, + "step": 24181 + }, + { + "epoch": 2.6556116846035582, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.52640700340271, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7468494772911072, + "num_tokens": 625558477.0, + "step": 24182 + }, + { + "epoch": 2.6557215023061715, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.264899492263794, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7349886894226074, + "num_tokens": 625585865.0, + "step": 24183 + }, + { + "epoch": 2.6558313200087853, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.365476369857788, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7245857119560242, + "num_tokens": 625610298.0, + "step": 24184 + }, + { + "epoch": 2.655941137711399, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.376121759414673, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7345341444015503, + "num_tokens": 625635271.0, + "step": 24185 + }, + { + "epoch": 2.656050955414013, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2311148643493652, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7011421322822571, + "num_tokens": 625667778.0, + "step": 24186 + }, + { + "epoch": 2.6561607731166266, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.307509660720825, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7222025394439697, + "num_tokens": 625693015.0, + "step": 24187 + }, + { + "epoch": 2.65627059081924, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1780593395233154, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7173094153404236, + "num_tokens": 625726287.0, + "step": 24188 + }, + { + "epoch": 2.6563804085218536, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.736194610595703, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7300492525100708, + "num_tokens": 625747179.0, + "step": 24189 + }, + { + "epoch": 2.6564902262244674, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.423219919204712, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7231932878494263, + "num_tokens": 625772030.0, + "step": 24190 + }, + { + "epoch": 2.656600043927081, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.30841064453125, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7242205142974854, + "num_tokens": 625800036.0, + "step": 24191 + }, + { + "epoch": 2.656709861629695, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2440571784973145, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7222601175308228, + "num_tokens": 625828947.0, + "step": 24192 + }, + { + "epoch": 2.6568196793323082, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.348205804824829, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7193149328231812, + "num_tokens": 625855482.0, + "step": 24193 + }, + { + "epoch": 2.656929497034922, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.2356507778167725, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7468501925468445, + "num_tokens": 625881288.0, + "step": 24194 + }, + { + "epoch": 2.6570393147375357, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.456778049468994, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.716640830039978, + "num_tokens": 625905574.0, + "step": 24195 + }, + { + "epoch": 2.6571491324401495, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1377456188201904, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7304887771606445, + "num_tokens": 625934805.0, + "step": 24196 + }, + { + "epoch": 2.6572589501427633, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6319665908813477, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7396005392074585, + "num_tokens": 625957195.0, + "step": 24197 + }, + { + "epoch": 2.6573687678453766, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6133036613464355, + "learning_rate": 1e-06, + "loss": 0.6985, + "mean_token_accuracy": 0.7713475227355957, + "num_tokens": 625976035.0, + "step": 24198 + }, + { + "epoch": 2.6574785855479903, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.636274576187134, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.721337616443634, + "num_tokens": 625999073.0, + "step": 24199 + }, + { + "epoch": 2.657588403250604, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.312227487564087, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7266712784767151, + "num_tokens": 626026380.0, + "step": 24200 + }, + { + "epoch": 2.6576982209532174, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4194464683532715, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7196005582809448, + "num_tokens": 626052800.0, + "step": 24201 + }, + { + "epoch": 2.6578080386558316, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4806971549987793, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7264626026153564, + "num_tokens": 626076458.0, + "step": 24202 + }, + { + "epoch": 2.657917856358445, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6118814945220947, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7466360330581665, + "num_tokens": 626097548.0, + "step": 24203 + }, + { + "epoch": 2.6580276740610587, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4724321365356445, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7411638498306274, + "num_tokens": 626121697.0, + "step": 24204 + }, + { + "epoch": 2.6581374917636724, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.5830187797546387, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7355551719665527, + "num_tokens": 626143533.0, + "step": 24205 + }, + { + "epoch": 2.6582473094662857, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3111612796783447, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7178000807762146, + "num_tokens": 626169981.0, + "step": 24206 + }, + { + "epoch": 2.6583571271688995, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3262269496917725, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7237966060638428, + "num_tokens": 626197217.0, + "step": 24207 + }, + { + "epoch": 2.6584669448715132, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.0265910625457764, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7315053939819336, + "num_tokens": 626229495.0, + "step": 24208 + }, + { + "epoch": 2.658576762574127, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1559314727783203, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.724209189414978, + "num_tokens": 626257022.0, + "step": 24209 + }, + { + "epoch": 2.6586865802767408, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.480691909790039, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7198227643966675, + "num_tokens": 626280771.0, + "step": 24210 + }, + { + "epoch": 2.658796397979354, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4673757553100586, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7201020121574402, + "num_tokens": 626305088.0, + "step": 24211 + }, + { + "epoch": 2.658906215681968, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.7558679580688477, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7176280617713928, + "num_tokens": 626327391.0, + "step": 24212 + }, + { + "epoch": 2.6590160333845816, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3545303344726562, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7380620241165161, + "num_tokens": 626354456.0, + "step": 24213 + }, + { + "epoch": 2.6591258510871953, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.234283924102783, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7403215765953064, + "num_tokens": 626380975.0, + "step": 24214 + }, + { + "epoch": 2.659235668789809, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.213986873626709, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7484176158905029, + "num_tokens": 626408294.0, + "step": 24215 + }, + { + "epoch": 2.6593454864924224, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.272843837738037, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7338884472846985, + "num_tokens": 626436301.0, + "step": 24216 + }, + { + "epoch": 2.659455304195036, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3203890323638916, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7197349071502686, + "num_tokens": 626461838.0, + "step": 24217 + }, + { + "epoch": 2.65956512189765, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.1361191272735596, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7180267572402954, + "num_tokens": 626491759.0, + "step": 24218 + }, + { + "epoch": 2.6596749396002637, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4036200046539307, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7351123094558716, + "num_tokens": 626516598.0, + "step": 24219 + }, + { + "epoch": 2.6597847573028774, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.468085289001465, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7111548781394958, + "num_tokens": 626542760.0, + "step": 24220 + }, + { + "epoch": 2.6598945750054908, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3443832397460938, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7359730005264282, + "num_tokens": 626568063.0, + "step": 24221 + }, + { + "epoch": 2.6600043927081045, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 1.9249539375305176, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7342935800552368, + "num_tokens": 626602420.0, + "step": 24222 + }, + { + "epoch": 2.6601142104107183, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.6297712326049805, + "learning_rate": 1e-06, + "loss": 0.7641, + "mean_token_accuracy": 0.7532007694244385, + "num_tokens": 626622888.0, + "step": 24223 + }, + { + "epoch": 2.660224028113332, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 1.9871628284454346, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7412947416305542, + "num_tokens": 626655392.0, + "step": 24224 + }, + { + "epoch": 2.660333845815946, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.0817532539367676, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7165614366531372, + "num_tokens": 626687978.0, + "step": 24225 + }, + { + "epoch": 2.660443663518559, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.4413528442382812, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7305687665939331, + "num_tokens": 626714453.0, + "step": 24226 + }, + { + "epoch": 2.660553481221173, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.601461172103882, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7338857054710388, + "num_tokens": 626738129.0, + "step": 24227 + }, + { + "epoch": 2.6606632989237866, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.530259847640991, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7328243255615234, + "num_tokens": 626762084.0, + "step": 24228 + }, + { + "epoch": 2.6607731166264, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.498692750930786, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.729059636592865, + "num_tokens": 626786533.0, + "step": 24229 + }, + { + "epoch": 2.6608829343290137, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.610642194747925, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7354406118392944, + "num_tokens": 626811012.0, + "step": 24230 + }, + { + "epoch": 2.6609927520316274, + "ewc_loss": 0.00021266937255859375, + "grad_norm": 2.3664002418518066, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7529009580612183, + "num_tokens": 626835039.0, + "step": 24231 + }, + { + "epoch": 2.661102569734241, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.360529661178589, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7297860980033875, + "num_tokens": 626859162.0, + "step": 24232 + }, + { + "epoch": 2.661212387436855, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.228802442550659, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7080422639846802, + "num_tokens": 626890271.0, + "step": 24233 + }, + { + "epoch": 2.6613222051394683, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.5185532569885254, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7232984304428101, + "num_tokens": 626916816.0, + "step": 24234 + }, + { + "epoch": 2.661432022842082, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.0946075916290283, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7111208438873291, + "num_tokens": 626948544.0, + "step": 24235 + }, + { + "epoch": 2.661541840544696, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.693662643432617, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7182415127754211, + "num_tokens": 626969038.0, + "step": 24236 + }, + { + "epoch": 2.6616516582473095, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.1083343029022217, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7212879061698914, + "num_tokens": 627001751.0, + "step": 24237 + }, + { + "epoch": 2.6617614759499233, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4453511238098145, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.726219654083252, + "num_tokens": 627025947.0, + "step": 24238 + }, + { + "epoch": 2.6618712936525366, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.745138168334961, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7367792725563049, + "num_tokens": 627045642.0, + "step": 24239 + }, + { + "epoch": 2.6619811113551504, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.224777936935425, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7243731021881104, + "num_tokens": 627072345.0, + "step": 24240 + }, + { + "epoch": 2.662090929057764, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.269648551940918, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7362004518508911, + "num_tokens": 627100585.0, + "step": 24241 + }, + { + "epoch": 2.662200746760378, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.322781562805176, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7136874794960022, + "num_tokens": 627126318.0, + "step": 24242 + }, + { + "epoch": 2.6623105644629916, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2866642475128174, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7148661017417908, + "num_tokens": 627153684.0, + "step": 24243 + }, + { + "epoch": 2.662420382165605, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.249706745147705, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7278950214385986, + "num_tokens": 627182507.0, + "step": 24244 + }, + { + "epoch": 2.6625301998682187, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2072386741638184, + "learning_rate": 1e-06, + "loss": 0.7193, + "mean_token_accuracy": 0.7703922986984253, + "num_tokens": 627210042.0, + "step": 24245 + }, + { + "epoch": 2.6626400175708325, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2831788063049316, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.7598305344581604, + "num_tokens": 627235725.0, + "step": 24246 + }, + { + "epoch": 2.662749835273446, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2236225605010986, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7177106142044067, + "num_tokens": 627264374.0, + "step": 24247 + }, + { + "epoch": 2.66285965297606, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.176593065261841, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7142826318740845, + "num_tokens": 627292718.0, + "step": 24248 + }, + { + "epoch": 2.6629694706786733, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.198611259460449, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7030515670776367, + "num_tokens": 627323275.0, + "step": 24249 + }, + { + "epoch": 2.663079288381287, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.541316032409668, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.74640953540802, + "num_tokens": 627347055.0, + "step": 24250 + }, + { + "epoch": 2.663189106083901, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2547085285186768, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7171010375022888, + "num_tokens": 627375292.0, + "step": 24251 + }, + { + "epoch": 2.663298923786514, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.310760974884033, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7098517417907715, + "num_tokens": 627404543.0, + "step": 24252 + }, + { + "epoch": 2.6634087414891283, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.1913230419158936, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7205565571784973, + "num_tokens": 627435946.0, + "step": 24253 + }, + { + "epoch": 2.6635185591917416, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.338233709335327, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7201173901557922, + "num_tokens": 627463374.0, + "step": 24254 + }, + { + "epoch": 2.6636283768943554, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.6455559730529785, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7325820326805115, + "num_tokens": 627486138.0, + "step": 24255 + }, + { + "epoch": 2.663738194596969, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3123786449432373, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7300907373428345, + "num_tokens": 627512833.0, + "step": 24256 + }, + { + "epoch": 2.6638480122995825, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.0674359798431396, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7157273292541504, + "num_tokens": 627545399.0, + "step": 24257 + }, + { + "epoch": 2.663957830002196, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2956438064575195, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.719809889793396, + "num_tokens": 627571910.0, + "step": 24258 + }, + { + "epoch": 2.66406764770481, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3861258029937744, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7203713655471802, + "num_tokens": 627600017.0, + "step": 24259 + }, + { + "epoch": 2.6641774654074237, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2185885906219482, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7250884175300598, + "num_tokens": 627630488.0, + "step": 24260 + }, + { + "epoch": 2.6642872831100375, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.521122694015503, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7346794605255127, + "num_tokens": 627654978.0, + "step": 24261 + }, + { + "epoch": 2.664397100812651, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.672912836074829, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7207791805267334, + "num_tokens": 627676512.0, + "step": 24262 + }, + { + "epoch": 2.6645069185152646, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4577431678771973, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7079883813858032, + "num_tokens": 627702865.0, + "step": 24263 + }, + { + "epoch": 2.6646167362178783, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3878045082092285, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.735734224319458, + "num_tokens": 627729759.0, + "step": 24264 + }, + { + "epoch": 2.664726553920492, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.1085779666900635, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7140841484069824, + "num_tokens": 627761297.0, + "step": 24265 + }, + { + "epoch": 2.664836371623106, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.579249858856201, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7392164468765259, + "num_tokens": 627783809.0, + "step": 24266 + }, + { + "epoch": 2.664946189325719, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4675896167755127, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7379087805747986, + "num_tokens": 627808469.0, + "step": 24267 + }, + { + "epoch": 2.665056007028333, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.329092025756836, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7449175119400024, + "num_tokens": 627832696.0, + "step": 24268 + }, + { + "epoch": 2.6651658247309467, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.428776502609253, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7461485862731934, + "num_tokens": 627857509.0, + "step": 24269 + }, + { + "epoch": 2.6652756424335604, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.295233726501465, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7516984343528748, + "num_tokens": 627883020.0, + "step": 24270 + }, + { + "epoch": 2.665385460136174, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.405951499938965, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7275297045707703, + "num_tokens": 627907355.0, + "step": 24271 + }, + { + "epoch": 2.6654952778387875, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.507456064224243, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7326622605323792, + "num_tokens": 627934318.0, + "step": 24272 + }, + { + "epoch": 2.6656050955414012, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.348543882369995, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7386188507080078, + "num_tokens": 627962342.0, + "step": 24273 + }, + { + "epoch": 2.665714913244015, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.445464611053467, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7069307565689087, + "num_tokens": 627989770.0, + "step": 24274 + }, + { + "epoch": 2.6658247309466288, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.6350443363189697, + "learning_rate": 1e-06, + "loss": 0.7318, + "mean_token_accuracy": 0.7605372667312622, + "num_tokens": 628009019.0, + "step": 24275 + }, + { + "epoch": 2.6659345486492425, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.023871421813965, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7005072832107544, + "num_tokens": 628044962.0, + "step": 24276 + }, + { + "epoch": 2.666044366351856, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.450472116470337, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.743621826171875, + "num_tokens": 628069155.0, + "step": 24277 + }, + { + "epoch": 2.6661541840544696, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4812583923339844, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7238501310348511, + "num_tokens": 628094640.0, + "step": 24278 + }, + { + "epoch": 2.6662640017570833, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4332518577575684, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7203871607780457, + "num_tokens": 628118746.0, + "step": 24279 + }, + { + "epoch": 2.6663738194596966, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.24849271774292, + "learning_rate": 1e-06, + "loss": 0.7548, + "mean_token_accuracy": 0.7595064640045166, + "num_tokens": 628143374.0, + "step": 24280 + }, + { + "epoch": 2.6664836371623104, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.0719809532165527, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.742280900478363, + "num_tokens": 628175194.0, + "step": 24281 + }, + { + "epoch": 2.666593454864924, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4288744926452637, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7391363382339478, + "num_tokens": 628199375.0, + "step": 24282 + }, + { + "epoch": 2.666703272567538, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.357630491256714, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7369880676269531, + "num_tokens": 628224584.0, + "step": 24283 + }, + { + "epoch": 2.6668130902701517, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.5492727756500244, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7231556177139282, + "num_tokens": 628246848.0, + "step": 24284 + }, + { + "epoch": 2.666922907972765, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3495171070098877, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7016518115997314, + "num_tokens": 628274470.0, + "step": 24285 + }, + { + "epoch": 2.6670327256753787, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3809268474578857, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7477308511734009, + "num_tokens": 628299899.0, + "step": 24286 + }, + { + "epoch": 2.6671425433779925, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.449223041534424, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7131668925285339, + "num_tokens": 628323819.0, + "step": 24287 + }, + { + "epoch": 2.6672523610806063, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.547912359237671, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7355830669403076, + "num_tokens": 628346219.0, + "step": 24288 + }, + { + "epoch": 2.66736217878322, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.354496955871582, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7311117649078369, + "num_tokens": 628370534.0, + "step": 24289 + }, + { + "epoch": 2.6674719964858333, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.325084924697876, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7195053100585938, + "num_tokens": 628397247.0, + "step": 24290 + }, + { + "epoch": 2.667581814188447, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.201108932495117, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7332512140274048, + "num_tokens": 628424134.0, + "step": 24291 + }, + { + "epoch": 2.667691631891061, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2744009494781494, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7509688138961792, + "num_tokens": 628451886.0, + "step": 24292 + }, + { + "epoch": 2.6678014495936746, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3641936779022217, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7353176474571228, + "num_tokens": 628475401.0, + "step": 24293 + }, + { + "epoch": 2.6679112672962884, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.5755562782287598, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7295321226119995, + "num_tokens": 628496463.0, + "step": 24294 + }, + { + "epoch": 2.6680210849989017, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.7129721641540527, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7420452833175659, + "num_tokens": 628518543.0, + "step": 24295 + }, + { + "epoch": 2.6681309027015154, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2967655658721924, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7244868874549866, + "num_tokens": 628546183.0, + "step": 24296 + }, + { + "epoch": 2.668240720404129, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.469928503036499, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7328683733940125, + "num_tokens": 628569383.0, + "step": 24297 + }, + { + "epoch": 2.668350538106743, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.461367607116699, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7411391139030457, + "num_tokens": 628592940.0, + "step": 24298 + }, + { + "epoch": 2.6684603558093567, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3196425437927246, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7239539623260498, + "num_tokens": 628620055.0, + "step": 24299 + }, + { + "epoch": 2.66857017351197, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2915334701538086, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7340724468231201, + "num_tokens": 628645968.0, + "step": 24300 + }, + { + "epoch": 2.6686799912145838, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.504267692565918, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7272717952728271, + "num_tokens": 628667921.0, + "step": 24301 + }, + { + "epoch": 2.6687898089171975, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4088385105133057, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7390434145927429, + "num_tokens": 628693110.0, + "step": 24302 + }, + { + "epoch": 2.668899626619811, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4958901405334473, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7200618982315063, + "num_tokens": 628716271.0, + "step": 24303 + }, + { + "epoch": 2.669009444322425, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.349984884262085, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7246809005737305, + "num_tokens": 628741671.0, + "step": 24304 + }, + { + "epoch": 2.6691192620250384, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3397083282470703, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.711212158203125, + "num_tokens": 628768746.0, + "step": 24305 + }, + { + "epoch": 2.669229079727652, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.517103910446167, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.7511610984802246, + "num_tokens": 628792673.0, + "step": 24306 + }, + { + "epoch": 2.669338897430266, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.322850465774536, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.719099760055542, + "num_tokens": 628817965.0, + "step": 24307 + }, + { + "epoch": 2.669448715132879, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.0202138423919678, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7041568756103516, + "num_tokens": 628852368.0, + "step": 24308 + }, + { + "epoch": 2.669558532835493, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.6399405002593994, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7332441806793213, + "num_tokens": 628876487.0, + "step": 24309 + }, + { + "epoch": 2.6696683505381067, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.211808204650879, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7272995710372925, + "num_tokens": 628904968.0, + "step": 24310 + }, + { + "epoch": 2.6697781682407205, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.098111867904663, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7044315338134766, + "num_tokens": 628936609.0, + "step": 24311 + }, + { + "epoch": 2.669887985943334, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.626335859298706, + "learning_rate": 1e-06, + "loss": 0.7297, + "mean_token_accuracy": 0.7601919174194336, + "num_tokens": 628956272.0, + "step": 24312 + }, + { + "epoch": 2.6699978036459475, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.543829917907715, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.734248161315918, + "num_tokens": 628978794.0, + "step": 24313 + }, + { + "epoch": 2.6701076213485613, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.550583839416504, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.721923828125, + "num_tokens": 629004138.0, + "step": 24314 + }, + { + "epoch": 2.670217439051175, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4262139797210693, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7269989252090454, + "num_tokens": 629030168.0, + "step": 24315 + }, + { + "epoch": 2.670327256753789, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3116302490234375, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7268173098564148, + "num_tokens": 629057981.0, + "step": 24316 + }, + { + "epoch": 2.6704370744564025, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2008135318756104, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7358220219612122, + "num_tokens": 629086282.0, + "step": 24317 + }, + { + "epoch": 2.670546892159016, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.483552932739258, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7384049892425537, + "num_tokens": 629111573.0, + "step": 24318 + }, + { + "epoch": 2.6706567098616296, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.56514048576355, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7326354384422302, + "num_tokens": 629134667.0, + "step": 24319 + }, + { + "epoch": 2.6707665275642434, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1039016246795654, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7168396711349487, + "num_tokens": 629169421.0, + "step": 24320 + }, + { + "epoch": 2.670876345266857, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6838698387145996, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7434355616569519, + "num_tokens": 629190802.0, + "step": 24321 + }, + { + "epoch": 2.670986162969471, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.538240671157837, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7248628735542297, + "num_tokens": 629214455.0, + "step": 24322 + }, + { + "epoch": 2.671095980672084, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3512027263641357, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7400181293487549, + "num_tokens": 629239574.0, + "step": 24323 + }, + { + "epoch": 2.671205798374698, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2835395336151123, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7113170623779297, + "num_tokens": 629266362.0, + "step": 24324 + }, + { + "epoch": 2.6713156160773117, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4088613986968994, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7340987920761108, + "num_tokens": 629290907.0, + "step": 24325 + }, + { + "epoch": 2.6714254337799255, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3839943408966064, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7339233160018921, + "num_tokens": 629317945.0, + "step": 24326 + }, + { + "epoch": 2.6715352514825392, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.5189130306243896, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7334773540496826, + "num_tokens": 629340176.0, + "step": 24327 + }, + { + "epoch": 2.6716450691851525, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.358989715576172, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.6982067823410034, + "num_tokens": 629367833.0, + "step": 24328 + }, + { + "epoch": 2.6717548868877663, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3757708072662354, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7213336229324341, + "num_tokens": 629395065.0, + "step": 24329 + }, + { + "epoch": 2.67186470459038, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4425511360168457, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7336615324020386, + "num_tokens": 629420042.0, + "step": 24330 + }, + { + "epoch": 2.6719745222929934, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.8739664554595947, + "learning_rate": 1e-06, + "loss": 0.7502, + "mean_token_accuracy": 0.7596189975738525, + "num_tokens": 629438858.0, + "step": 24331 + }, + { + "epoch": 2.672084339995607, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4352002143859863, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7365448474884033, + "num_tokens": 629463482.0, + "step": 24332 + }, + { + "epoch": 2.672194157698221, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4218404293060303, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7216977477073669, + "num_tokens": 629489477.0, + "step": 24333 + }, + { + "epoch": 2.6723039754008346, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.12276029586792, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7476378679275513, + "num_tokens": 629516528.0, + "step": 24334 + }, + { + "epoch": 2.6724137931034484, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.5600547790527344, + "learning_rate": 1e-06, + "loss": 0.7678, + "mean_token_accuracy": 0.7500354051589966, + "num_tokens": 629538327.0, + "step": 24335 + }, + { + "epoch": 2.6725236108060617, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.87426495552063, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7385015487670898, + "num_tokens": 629558109.0, + "step": 24336 + }, + { + "epoch": 2.6726334285086755, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4796035289764404, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7498953342437744, + "num_tokens": 629581221.0, + "step": 24337 + }, + { + "epoch": 2.6727432462112892, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.278991222381592, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7196396589279175, + "num_tokens": 629610989.0, + "step": 24338 + }, + { + "epoch": 2.672853063913903, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3018882274627686, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.6973634362220764, + "num_tokens": 629639140.0, + "step": 24339 + }, + { + "epoch": 2.6729628816165167, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.7845680713653564, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7440235614776611, + "num_tokens": 629658917.0, + "step": 24340 + }, + { + "epoch": 2.67307269931913, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.646925210952759, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7397762537002563, + "num_tokens": 629683062.0, + "step": 24341 + }, + { + "epoch": 2.673182517021744, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.18383526802063, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7049484848976135, + "num_tokens": 629717786.0, + "step": 24342 + }, + { + "epoch": 2.6732923347243576, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.64440655708313, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7421087622642517, + "num_tokens": 629738750.0, + "step": 24343 + }, + { + "epoch": 2.6734021524269713, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.719884157180786, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7190495729446411, + "num_tokens": 629759552.0, + "step": 24344 + }, + { + "epoch": 2.673511970129585, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3522486686706543, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7421046495437622, + "num_tokens": 629787628.0, + "step": 24345 + }, + { + "epoch": 2.6736217878321984, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.351222515106201, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7361382842063904, + "num_tokens": 629812867.0, + "step": 24346 + }, + { + "epoch": 2.673731605534812, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.338233470916748, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7228460311889648, + "num_tokens": 629839150.0, + "step": 24347 + }, + { + "epoch": 2.673841423237426, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.3670997619628906, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7261189222335815, + "num_tokens": 629863973.0, + "step": 24348 + }, + { + "epoch": 2.6739512409400397, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.459124803543091, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.728510856628418, + "num_tokens": 629888021.0, + "step": 24349 + }, + { + "epoch": 2.6740610586426534, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.1028904914855957, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.70916348695755, + "num_tokens": 629920231.0, + "step": 24350 + }, + { + "epoch": 2.6741708763452667, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2509512901306152, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7257692813873291, + "num_tokens": 629947745.0, + "step": 24351 + }, + { + "epoch": 2.6742806940478805, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.6491000652313232, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7309896945953369, + "num_tokens": 629969679.0, + "step": 24352 + }, + { + "epoch": 2.6743905117504942, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2727017402648926, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7084683775901794, + "num_tokens": 630001784.0, + "step": 24353 + }, + { + "epoch": 2.674500329453108, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.224248170852661, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7138581275939941, + "num_tokens": 630033023.0, + "step": 24354 + }, + { + "epoch": 2.6746101471557218, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.6457595825195312, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7234593629837036, + "num_tokens": 630059274.0, + "step": 24355 + }, + { + "epoch": 2.674719964858335, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4184746742248535, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7197253704071045, + "num_tokens": 630085439.0, + "step": 24356 + }, + { + "epoch": 2.674829782560949, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.361705780029297, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7156844139099121, + "num_tokens": 630111706.0, + "step": 24357 + }, + { + "epoch": 2.6749396002635626, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2360010147094727, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7256366014480591, + "num_tokens": 630140800.0, + "step": 24358 + }, + { + "epoch": 2.675049417966176, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.1756508350372314, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7204778790473938, + "num_tokens": 630170430.0, + "step": 24359 + }, + { + "epoch": 2.6751592356687897, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4989171028137207, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7418534159660339, + "num_tokens": 630192434.0, + "step": 24360 + }, + { + "epoch": 2.6752690533714034, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.448274612426758, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7364039421081543, + "num_tokens": 630216439.0, + "step": 24361 + }, + { + "epoch": 2.675378871074017, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.2392048835754395, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7133681774139404, + "num_tokens": 630244907.0, + "step": 24362 + }, + { + "epoch": 2.675488688776631, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.201767921447754, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7297764420509338, + "num_tokens": 630273273.0, + "step": 24363 + }, + { + "epoch": 2.6755985064792442, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.321904420852661, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7264966368675232, + "num_tokens": 630299721.0, + "step": 24364 + }, + { + "epoch": 2.675708324181858, + "ewc_loss": 0.00021457672119140625, + "grad_norm": 2.4212353229522705, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7219147682189941, + "num_tokens": 630325033.0, + "step": 24365 + }, + { + "epoch": 2.6758181418844718, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3705015182495117, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7215585112571716, + "num_tokens": 630351318.0, + "step": 24366 + }, + { + "epoch": 2.6759279595870855, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.282790422439575, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.722403347492218, + "num_tokens": 630378922.0, + "step": 24367 + }, + { + "epoch": 2.6760377772896993, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1315126419067383, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7152913808822632, + "num_tokens": 630408757.0, + "step": 24368 + }, + { + "epoch": 2.6761475949923126, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3699252605438232, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7424102425575256, + "num_tokens": 630435441.0, + "step": 24369 + }, + { + "epoch": 2.6762574126949263, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3133718967437744, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.713283896446228, + "num_tokens": 630462981.0, + "step": 24370 + }, + { + "epoch": 2.67636723039754, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5703389644622803, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7487954497337341, + "num_tokens": 630484003.0, + "step": 24371 + }, + { + "epoch": 2.676477048100154, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4668495655059814, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7262290716171265, + "num_tokens": 630510191.0, + "step": 24372 + }, + { + "epoch": 2.6765868658027676, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.602388620376587, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7356481552124023, + "num_tokens": 630532773.0, + "step": 24373 + }, + { + "epoch": 2.676696683505381, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3990073204040527, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7417718172073364, + "num_tokens": 630557174.0, + "step": 24374 + }, + { + "epoch": 2.6768065012079947, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.475929021835327, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.741373598575592, + "num_tokens": 630580489.0, + "step": 24375 + }, + { + "epoch": 2.6769163189106084, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2497365474700928, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7272947430610657, + "num_tokens": 630609870.0, + "step": 24376 + }, + { + "epoch": 2.677026136613222, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6139495372772217, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7300487160682678, + "num_tokens": 630630976.0, + "step": 24377 + }, + { + "epoch": 2.677135954315836, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.7511959075927734, + "learning_rate": 1e-06, + "loss": 0.7896, + "mean_token_accuracy": 0.7459706664085388, + "num_tokens": 630651152.0, + "step": 24378 + }, + { + "epoch": 2.6772457720184493, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.276914596557617, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7101883888244629, + "num_tokens": 630679285.0, + "step": 24379 + }, + { + "epoch": 2.677355589721063, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.281203269958496, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7154141664505005, + "num_tokens": 630706399.0, + "step": 24380 + }, + { + "epoch": 2.677465407423677, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.253047227859497, + "learning_rate": 1e-06, + "loss": 0.7242, + "mean_token_accuracy": 0.7636429071426392, + "num_tokens": 630732261.0, + "step": 24381 + }, + { + "epoch": 2.67757522512629, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.430878162384033, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7413629293441772, + "num_tokens": 630755814.0, + "step": 24382 + }, + { + "epoch": 2.6776850428289043, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3154547214508057, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7193170785903931, + "num_tokens": 630782783.0, + "step": 24383 + }, + { + "epoch": 2.6777948605315176, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0520482063293457, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7080825567245483, + "num_tokens": 630815627.0, + "step": 24384 + }, + { + "epoch": 2.6779046782341314, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 3.054805278778076, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7525977492332458, + "num_tokens": 630832948.0, + "step": 24385 + }, + { + "epoch": 2.678014495936745, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3431007862091064, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7364405989646912, + "num_tokens": 630860164.0, + "step": 24386 + }, + { + "epoch": 2.6781243136393584, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3034861087799072, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7256844639778137, + "num_tokens": 630889955.0, + "step": 24387 + }, + { + "epoch": 2.678234131341972, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.28696346282959, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7165663242340088, + "num_tokens": 630917747.0, + "step": 24388 + }, + { + "epoch": 2.678343949044586, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2775931358337402, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7180377244949341, + "num_tokens": 630945446.0, + "step": 24389 + }, + { + "epoch": 2.6784537667471997, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.335697650909424, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7060177326202393, + "num_tokens": 630973759.0, + "step": 24390 + }, + { + "epoch": 2.6785635844498135, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3168745040893555, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7192597389221191, + "num_tokens": 631001578.0, + "step": 24391 + }, + { + "epoch": 2.6786734021524268, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0440456867218018, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7287822961807251, + "num_tokens": 631032517.0, + "step": 24392 + }, + { + "epoch": 2.6787832198550405, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6570136547088623, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7295277118682861, + "num_tokens": 631054209.0, + "step": 24393 + }, + { + "epoch": 2.6788930375576543, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.955394983291626, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7504096627235413, + "num_tokens": 631073210.0, + "step": 24394 + }, + { + "epoch": 2.679002855260268, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.282796859741211, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7328975200653076, + "num_tokens": 631099917.0, + "step": 24395 + }, + { + "epoch": 2.679112672962882, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3244969844818115, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7003545761108398, + "num_tokens": 631128902.0, + "step": 24396 + }, + { + "epoch": 2.679222490665495, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.225245237350464, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7072970867156982, + "num_tokens": 631157803.0, + "step": 24397 + }, + { + "epoch": 2.679332308368109, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.712786912918091, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7383768558502197, + "num_tokens": 631177743.0, + "step": 24398 + }, + { + "epoch": 2.6794421260707226, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.403136730194092, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7443259954452515, + "num_tokens": 631201585.0, + "step": 24399 + }, + { + "epoch": 2.6795519437733364, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 1.9872114658355713, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7374157905578613, + "num_tokens": 631235262.0, + "step": 24400 + }, + { + "epoch": 2.67966176147595, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.592561721801758, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7226739525794983, + "num_tokens": 631257057.0, + "step": 24401 + }, + { + "epoch": 2.6797715791785635, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0384938716888428, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.7139030694961548, + "num_tokens": 631287863.0, + "step": 24402 + }, + { + "epoch": 2.679881396881177, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4132957458496094, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7236138582229614, + "num_tokens": 631312985.0, + "step": 24403 + }, + { + "epoch": 2.679991214583791, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5724666118621826, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7296420335769653, + "num_tokens": 631337562.0, + "step": 24404 + }, + { + "epoch": 2.6801010322864047, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.7534568309783936, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7445343732833862, + "num_tokens": 631357719.0, + "step": 24405 + }, + { + "epoch": 2.6802108499890185, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0970590114593506, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.727757453918457, + "num_tokens": 631389130.0, + "step": 24406 + }, + { + "epoch": 2.680320667691632, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.667030096054077, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7352355718612671, + "num_tokens": 631410423.0, + "step": 24407 + }, + { + "epoch": 2.6804304853942456, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1964075565338135, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7499101161956787, + "num_tokens": 631437968.0, + "step": 24408 + }, + { + "epoch": 2.6805403030968593, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2567942142486572, + "learning_rate": 1e-06, + "loss": 0.7415, + "mean_token_accuracy": 0.7591887712478638, + "num_tokens": 631463678.0, + "step": 24409 + }, + { + "epoch": 2.6806501207994726, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1885275840759277, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7352211475372314, + "num_tokens": 631494062.0, + "step": 24410 + }, + { + "epoch": 2.6807599385020864, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3999598026275635, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7336442470550537, + "num_tokens": 631519307.0, + "step": 24411 + }, + { + "epoch": 2.6808697562047, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5704078674316406, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7410125136375427, + "num_tokens": 631542890.0, + "step": 24412 + }, + { + "epoch": 2.680979573907314, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.126427412033081, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.701367199420929, + "num_tokens": 631577584.0, + "step": 24413 + }, + { + "epoch": 2.6810893916099277, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3958516120910645, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7194052934646606, + "num_tokens": 631602628.0, + "step": 24414 + }, + { + "epoch": 2.681199209312541, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5063059329986572, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7399046421051025, + "num_tokens": 631625565.0, + "step": 24415 + }, + { + "epoch": 2.6813090270151547, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.564276933670044, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7365912795066833, + "num_tokens": 631648466.0, + "step": 24416 + }, + { + "epoch": 2.6814188447177685, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3908419609069824, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7211582660675049, + "num_tokens": 631672970.0, + "step": 24417 + }, + { + "epoch": 2.6815286624203822, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2952566146850586, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7160291075706482, + "num_tokens": 631700133.0, + "step": 24418 + }, + { + "epoch": 2.681638480122996, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3657944202423096, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.704024076461792, + "num_tokens": 631726860.0, + "step": 24419 + }, + { + "epoch": 2.6817482978256093, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1630566120147705, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7242201566696167, + "num_tokens": 631754956.0, + "step": 24420 + }, + { + "epoch": 2.681858115528223, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.589824676513672, + "learning_rate": 1e-06, + "loss": 0.7497, + "mean_token_accuracy": 0.7578340768814087, + "num_tokens": 631775773.0, + "step": 24421 + }, + { + "epoch": 2.681967933230837, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5988118648529053, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7364729642868042, + "num_tokens": 631798551.0, + "step": 24422 + }, + { + "epoch": 2.6820777509334506, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2413203716278076, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7284587621688843, + "num_tokens": 631826646.0, + "step": 24423 + }, + { + "epoch": 2.6821875686360643, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 1.9632785320281982, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7574893832206726, + "num_tokens": 631858987.0, + "step": 24424 + }, + { + "epoch": 2.6822973863386776, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.456860303878784, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7254383563995361, + "num_tokens": 631882069.0, + "step": 24425 + }, + { + "epoch": 2.6824072040412914, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2110755443573, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7268978953361511, + "num_tokens": 631911029.0, + "step": 24426 + }, + { + "epoch": 2.682517021743905, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4271576404571533, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6991657018661499, + "num_tokens": 631937513.0, + "step": 24427 + }, + { + "epoch": 2.682626839446519, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.201693296432495, + "learning_rate": 1e-06, + "loss": 0.7819, + "mean_token_accuracy": 0.744723916053772, + "num_tokens": 631964628.0, + "step": 24428 + }, + { + "epoch": 2.6827366571491327, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5921459197998047, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7252411842346191, + "num_tokens": 631987694.0, + "step": 24429 + }, + { + "epoch": 2.682846474851746, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.308680772781372, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7137953042984009, + "num_tokens": 632018211.0, + "step": 24430 + }, + { + "epoch": 2.6829562925543597, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.223113775253296, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7196769118309021, + "num_tokens": 632047901.0, + "step": 24431 + }, + { + "epoch": 2.6830661102569735, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.634448289871216, + "learning_rate": 1e-06, + "loss": 0.7471, + "mean_token_accuracy": 0.7578391432762146, + "num_tokens": 632068329.0, + "step": 24432 + }, + { + "epoch": 2.683175927959587, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4085657596588135, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7222564220428467, + "num_tokens": 632094069.0, + "step": 24433 + }, + { + "epoch": 2.683285745662201, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.188030242919922, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7269132137298584, + "num_tokens": 632123887.0, + "step": 24434 + }, + { + "epoch": 2.6833955633648143, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.325894832611084, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6954324245452881, + "num_tokens": 632151927.0, + "step": 24435 + }, + { + "epoch": 2.683505381067428, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.7560579776763916, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7428217530250549, + "num_tokens": 632172273.0, + "step": 24436 + }, + { + "epoch": 2.683615198770042, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.35137939453125, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7240613102912903, + "num_tokens": 632200268.0, + "step": 24437 + }, + { + "epoch": 2.683725016472655, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1243574619293213, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.712648332118988, + "num_tokens": 632232815.0, + "step": 24438 + }, + { + "epoch": 2.683834834175269, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4219765663146973, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.728538990020752, + "num_tokens": 632256901.0, + "step": 24439 + }, + { + "epoch": 2.6839446518778827, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3237884044647217, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7288190722465515, + "num_tokens": 632282897.0, + "step": 24440 + }, + { + "epoch": 2.6840544695804964, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5370776653289795, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.731740415096283, + "num_tokens": 632307063.0, + "step": 24441 + }, + { + "epoch": 2.68416428728311, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2066423892974854, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7322036027908325, + "num_tokens": 632336231.0, + "step": 24442 + }, + { + "epoch": 2.6842741049857235, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1913325786590576, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7364020347595215, + "num_tokens": 632365393.0, + "step": 24443 + }, + { + "epoch": 2.6843839226883373, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.383315324783325, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7024869322776794, + "num_tokens": 632392173.0, + "step": 24444 + }, + { + "epoch": 2.684493740390951, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3966727256774902, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7135350704193115, + "num_tokens": 632419194.0, + "step": 24445 + }, + { + "epoch": 2.6846035580935648, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0331614017486572, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7271426320075989, + "num_tokens": 632452581.0, + "step": 24446 + }, + { + "epoch": 2.6847133757961785, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0439295768737793, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7056567072868347, + "num_tokens": 632485790.0, + "step": 24447 + }, + { + "epoch": 2.684823193498792, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.7443127632141113, + "learning_rate": 1e-06, + "loss": 0.7823, + "mean_token_accuracy": 0.7510516047477722, + "num_tokens": 632504942.0, + "step": 24448 + }, + { + "epoch": 2.6849330112014056, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3752901554107666, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7468842267990112, + "num_tokens": 632528163.0, + "step": 24449 + }, + { + "epoch": 2.6850428289040194, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2090156078338623, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7205812931060791, + "num_tokens": 632556386.0, + "step": 24450 + }, + { + "epoch": 2.685152646606633, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6499576568603516, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7361047863960266, + "num_tokens": 632578355.0, + "step": 24451 + }, + { + "epoch": 2.685262464309247, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.369840145111084, + "learning_rate": 1e-06, + "loss": 0.7669, + "mean_token_accuracy": 0.7555761337280273, + "num_tokens": 632603086.0, + "step": 24452 + }, + { + "epoch": 2.68537228201186, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.30879807472229, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7314360737800598, + "num_tokens": 632631197.0, + "step": 24453 + }, + { + "epoch": 2.685482099714474, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.253908634185791, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7193871736526489, + "num_tokens": 632659157.0, + "step": 24454 + }, + { + "epoch": 2.6855919174170877, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.391752004623413, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7088015079498291, + "num_tokens": 632684876.0, + "step": 24455 + }, + { + "epoch": 2.6857017351197014, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.8118035793304443, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7485668659210205, + "num_tokens": 632704324.0, + "step": 24456 + }, + { + "epoch": 2.685811552822315, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.205472946166992, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7349324822425842, + "num_tokens": 632732603.0, + "step": 24457 + }, + { + "epoch": 2.6859213705249285, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.046117067337036, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7217565774917603, + "num_tokens": 632766272.0, + "step": 24458 + }, + { + "epoch": 2.6860311882275423, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2968668937683105, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.727709174156189, + "num_tokens": 632793284.0, + "step": 24459 + }, + { + "epoch": 2.686141005930156, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.23622465133667, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.6990396976470947, + "num_tokens": 632822643.0, + "step": 24460 + }, + { + "epoch": 2.6862508236327693, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3757219314575195, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7215083837509155, + "num_tokens": 632848903.0, + "step": 24461 + }, + { + "epoch": 2.686360641335383, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.504384756088257, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7521030306816101, + "num_tokens": 632871012.0, + "step": 24462 + }, + { + "epoch": 2.686470459037997, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.55537748336792, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7418810129165649, + "num_tokens": 632893130.0, + "step": 24463 + }, + { + "epoch": 2.6865802767406106, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2318027019500732, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7111890316009521, + "num_tokens": 632921317.0, + "step": 24464 + }, + { + "epoch": 2.6866900944432244, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2617952823638916, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7342837452888489, + "num_tokens": 632949586.0, + "step": 24465 + }, + { + "epoch": 2.6867999121458377, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.478860378265381, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7092069387435913, + "num_tokens": 632975352.0, + "step": 24466 + }, + { + "epoch": 2.6869097298484514, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.118119955062866, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7464755773544312, + "num_tokens": 633004096.0, + "step": 24467 + }, + { + "epoch": 2.687019547551065, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0514028072357178, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7088750600814819, + "num_tokens": 633035934.0, + "step": 24468 + }, + { + "epoch": 2.687129365253679, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1137540340423584, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7377080917358398, + "num_tokens": 633066967.0, + "step": 24469 + }, + { + "epoch": 2.6872391829562927, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5048937797546387, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7255685329437256, + "num_tokens": 633091102.0, + "step": 24470 + }, + { + "epoch": 2.687349000658906, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3178911209106445, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7487297058105469, + "num_tokens": 633115293.0, + "step": 24471 + }, + { + "epoch": 2.68745881836152, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5985076427459717, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7383503317832947, + "num_tokens": 633138373.0, + "step": 24472 + }, + { + "epoch": 2.6875686360641335, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4168009757995605, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7369788289070129, + "num_tokens": 633162004.0, + "step": 24473 + }, + { + "epoch": 2.6876784537667473, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.435823440551758, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7376424074172974, + "num_tokens": 633185919.0, + "step": 24474 + }, + { + "epoch": 2.687788271469361, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.262799024581909, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7280784249305725, + "num_tokens": 633214269.0, + "step": 24475 + }, + { + "epoch": 2.6878980891719744, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2731072902679443, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.722678542137146, + "num_tokens": 633242678.0, + "step": 24476 + }, + { + "epoch": 2.688007906874588, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.603992462158203, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7170155048370361, + "num_tokens": 633264820.0, + "step": 24477 + }, + { + "epoch": 2.688117724577202, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 1.9345821142196655, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7137507200241089, + "num_tokens": 633298596.0, + "step": 24478 + }, + { + "epoch": 2.6882275422798156, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2271485328674316, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.740143358707428, + "num_tokens": 633325792.0, + "step": 24479 + }, + { + "epoch": 2.6883373599824294, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.543931722640991, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7370489239692688, + "num_tokens": 633348353.0, + "step": 24480 + }, + { + "epoch": 2.6884471776850427, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.464134931564331, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7408826947212219, + "num_tokens": 633371318.0, + "step": 24481 + }, + { + "epoch": 2.6885569953876565, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.10280179977417, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7271375060081482, + "num_tokens": 633400630.0, + "step": 24482 + }, + { + "epoch": 2.6886668130902702, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.401822566986084, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7540757656097412, + "num_tokens": 633425392.0, + "step": 24483 + }, + { + "epoch": 2.6887766307928835, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5725362300872803, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7286799550056458, + "num_tokens": 633450601.0, + "step": 24484 + }, + { + "epoch": 2.6888864484954977, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4273974895477295, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7349278926849365, + "num_tokens": 633475397.0, + "step": 24485 + }, + { + "epoch": 2.688996266198111, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4241349697113037, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7295509576797485, + "num_tokens": 633500874.0, + "step": 24486 + }, + { + "epoch": 2.689106083900725, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.204343318939209, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7121233940124512, + "num_tokens": 633530432.0, + "step": 24487 + }, + { + "epoch": 2.6892159016033386, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.338094472885132, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7376585006713867, + "num_tokens": 633557693.0, + "step": 24488 + }, + { + "epoch": 2.689325719305952, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2978031635284424, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7392721176147461, + "num_tokens": 633584872.0, + "step": 24489 + }, + { + "epoch": 2.6894355370085656, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.7649295330047607, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7523870468139648, + "num_tokens": 633605096.0, + "step": 24490 + }, + { + "epoch": 2.6895453547111794, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.509845018386841, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7383374571800232, + "num_tokens": 633628223.0, + "step": 24491 + }, + { + "epoch": 2.689655172413793, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3949673175811768, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.734904408454895, + "num_tokens": 633654837.0, + "step": 24492 + }, + { + "epoch": 2.689764990116407, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5262951850891113, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.723961353302002, + "num_tokens": 633678396.0, + "step": 24493 + }, + { + "epoch": 2.68987480781902, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2512295246124268, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7191143035888672, + "num_tokens": 633709747.0, + "step": 24494 + }, + { + "epoch": 2.689984625521634, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.334052324295044, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7012945413589478, + "num_tokens": 633738224.0, + "step": 24495 + }, + { + "epoch": 2.6900944432242477, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2177350521087646, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7028824090957642, + "num_tokens": 633765948.0, + "step": 24496 + }, + { + "epoch": 2.6902042609268615, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5620338916778564, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7400588989257812, + "num_tokens": 633789191.0, + "step": 24497 + }, + { + "epoch": 2.6903140786294752, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.886371612548828, + "learning_rate": 1e-06, + "loss": 0.7255, + "mean_token_accuracy": 0.763444721698761, + "num_tokens": 633806814.0, + "step": 24498 + }, + { + "epoch": 2.6904238963320886, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2932636737823486, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7345820665359497, + "num_tokens": 633836908.0, + "step": 24499 + }, + { + "epoch": 2.6905337140347023, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5765697956085205, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.754167914390564, + "num_tokens": 633859780.0, + "step": 24500 + }, + { + "epoch": 2.690643531737316, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.379760980606079, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7058221101760864, + "num_tokens": 633885859.0, + "step": 24501 + }, + { + "epoch": 2.69075334943993, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.54679012298584, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.731313943862915, + "num_tokens": 633909121.0, + "step": 24502 + }, + { + "epoch": 2.6908631671425436, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.330617904663086, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.6991087794303894, + "num_tokens": 633938167.0, + "step": 24503 + }, + { + "epoch": 2.690972984845157, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.202998638153076, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7444123029708862, + "num_tokens": 633965109.0, + "step": 24504 + }, + { + "epoch": 2.6910828025477707, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3953919410705566, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7310017347335815, + "num_tokens": 633988185.0, + "step": 24505 + }, + { + "epoch": 2.6911926202503844, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.533527374267578, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7142751216888428, + "num_tokens": 634012442.0, + "step": 24506 + }, + { + "epoch": 2.691302437952998, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.548494338989258, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7384923696517944, + "num_tokens": 634036849.0, + "step": 24507 + }, + { + "epoch": 2.691412255655612, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.494204521179199, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7170246243476868, + "num_tokens": 634061913.0, + "step": 24508 + }, + { + "epoch": 2.6915220733582252, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3520398139953613, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7030965089797974, + "num_tokens": 634088505.0, + "step": 24509 + }, + { + "epoch": 2.691631891060839, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.1536149978637695, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7232488989830017, + "num_tokens": 634119812.0, + "step": 24510 + }, + { + "epoch": 2.6917417087634528, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.389869213104248, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7340234518051147, + "num_tokens": 634144470.0, + "step": 24511 + }, + { + "epoch": 2.691851526466066, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.0438132286071777, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7185016870498657, + "num_tokens": 634178419.0, + "step": 24512 + }, + { + "epoch": 2.69196134416868, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4239273071289062, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7192633152008057, + "num_tokens": 634204048.0, + "step": 24513 + }, + { + "epoch": 2.6920711618712936, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.661222219467163, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7541934251785278, + "num_tokens": 634224837.0, + "step": 24514 + }, + { + "epoch": 2.6921809795739073, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2581160068511963, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7236688137054443, + "num_tokens": 634251640.0, + "step": 24515 + }, + { + "epoch": 2.692290797276521, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.178164482116699, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6857367753982544, + "num_tokens": 634285192.0, + "step": 24516 + }, + { + "epoch": 2.6924006149791344, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.24172306060791, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7283986210823059, + "num_tokens": 634313160.0, + "step": 24517 + }, + { + "epoch": 2.692510432681748, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6060352325439453, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7580711841583252, + "num_tokens": 634334500.0, + "step": 24518 + }, + { + "epoch": 2.692620250384362, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.268702983856201, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7239582538604736, + "num_tokens": 634361753.0, + "step": 24519 + }, + { + "epoch": 2.6927300680869757, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.32846736907959, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7408702969551086, + "num_tokens": 634387044.0, + "step": 24520 + }, + { + "epoch": 2.6928398857895894, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2026143074035645, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7135162353515625, + "num_tokens": 634416653.0, + "step": 24521 + }, + { + "epoch": 2.6929497034922028, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.542306423187256, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7398825883865356, + "num_tokens": 634440434.0, + "step": 24522 + }, + { + "epoch": 2.6930595211948165, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2637829780578613, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7240007519721985, + "num_tokens": 634466287.0, + "step": 24523 + }, + { + "epoch": 2.6931693388974303, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.375325918197632, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7431721091270447, + "num_tokens": 634491980.0, + "step": 24524 + }, + { + "epoch": 2.693279156600044, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4248476028442383, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7383486032485962, + "num_tokens": 634515161.0, + "step": 24525 + }, + { + "epoch": 2.693388974302658, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.6708199977874756, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7300547361373901, + "num_tokens": 634536852.0, + "step": 24526 + }, + { + "epoch": 2.693498792005271, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4246554374694824, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7399659156799316, + "num_tokens": 634561649.0, + "step": 24527 + }, + { + "epoch": 2.693608609707885, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2215569019317627, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7455693483352661, + "num_tokens": 634586384.0, + "step": 24528 + }, + { + "epoch": 2.6937184274104986, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.127108097076416, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.728303074836731, + "num_tokens": 634617927.0, + "step": 24529 + }, + { + "epoch": 2.6938282451131124, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.299158811569214, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7416731119155884, + "num_tokens": 634644921.0, + "step": 24530 + }, + { + "epoch": 2.693938062815726, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.8221569061279297, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7519159913063049, + "num_tokens": 634665892.0, + "step": 24531 + }, + { + "epoch": 2.6940478805183394, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.4560599327087402, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7354732751846313, + "num_tokens": 634696236.0, + "step": 24532 + }, + { + "epoch": 2.694157698220953, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.202348470687866, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7393749952316284, + "num_tokens": 634725664.0, + "step": 24533 + }, + { + "epoch": 2.694267515923567, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.4982497692108154, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.734322726726532, + "num_tokens": 634749180.0, + "step": 24534 + }, + { + "epoch": 2.6943773336261807, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.6745333671569824, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7350107431411743, + "num_tokens": 634769158.0, + "step": 24535 + }, + { + "epoch": 2.6944871513287945, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.4787135124206543, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7299878597259521, + "num_tokens": 634791663.0, + "step": 24536 + }, + { + "epoch": 2.6945969690314078, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.433708667755127, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.733254611492157, + "num_tokens": 634817045.0, + "step": 24537 + }, + { + "epoch": 2.6947067867340215, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.732445001602173, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7345033884048462, + "num_tokens": 634838364.0, + "step": 24538 + }, + { + "epoch": 2.6948166044366353, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2928946018218994, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7106732130050659, + "num_tokens": 634867696.0, + "step": 24539 + }, + { + "epoch": 2.6949264221392486, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.7054405212402344, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7265214920043945, + "num_tokens": 634887761.0, + "step": 24540 + }, + { + "epoch": 2.6950362398418624, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.5754103660583496, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7358835935592651, + "num_tokens": 634910623.0, + "step": 24541 + }, + { + "epoch": 2.695146057544476, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.612859010696411, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7291473150253296, + "num_tokens": 634933794.0, + "step": 24542 + }, + { + "epoch": 2.69525587524709, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.2404725551605225, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7017432451248169, + "num_tokens": 634963445.0, + "step": 24543 + }, + { + "epoch": 2.6953656929497036, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.527108669281006, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7414926290512085, + "num_tokens": 634984866.0, + "step": 24544 + }, + { + "epoch": 2.695475510652317, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3200554847717285, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7367318868637085, + "num_tokens": 635012109.0, + "step": 24545 + }, + { + "epoch": 2.6955853283549307, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.268449068069458, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7188159823417664, + "num_tokens": 635040466.0, + "step": 24546 + }, + { + "epoch": 2.6956951460575445, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.541576385498047, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7371731996536255, + "num_tokens": 635064335.0, + "step": 24547 + }, + { + "epoch": 2.695804963760158, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.381060838699341, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7201145887374878, + "num_tokens": 635091295.0, + "step": 24548 + }, + { + "epoch": 2.695914781462772, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.201686143875122, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7152413725852966, + "num_tokens": 635122405.0, + "step": 24549 + }, + { + "epoch": 2.6960245991653853, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.744694948196411, + "learning_rate": 1e-06, + "loss": 0.7579, + "mean_token_accuracy": 0.7547299265861511, + "num_tokens": 635143309.0, + "step": 24550 + }, + { + "epoch": 2.696134416867999, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.256927251815796, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7345234751701355, + "num_tokens": 635172218.0, + "step": 24551 + }, + { + "epoch": 2.696244234570613, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.0392751693725586, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7070960998535156, + "num_tokens": 635206805.0, + "step": 24552 + }, + { + "epoch": 2.6963540522732266, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.3303046226501465, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7256470322608948, + "num_tokens": 635233987.0, + "step": 24553 + }, + { + "epoch": 2.6964638699758403, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.5689761638641357, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7082227468490601, + "num_tokens": 635257790.0, + "step": 24554 + }, + { + "epoch": 2.6965736876784536, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.573483943939209, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7361612915992737, + "num_tokens": 635280021.0, + "step": 24555 + }, + { + "epoch": 2.6966835053810674, + "ewc_loss": 0.0002155303955078125, + "grad_norm": 2.268312692642212, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7136452198028564, + "num_tokens": 635309456.0, + "step": 24556 + }, + { + "epoch": 2.696793323083681, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.5273783206939697, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7352465391159058, + "num_tokens": 635331597.0, + "step": 24557 + }, + { + "epoch": 2.696903140786295, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.888963460922241, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.733873188495636, + "num_tokens": 635352324.0, + "step": 24558 + }, + { + "epoch": 2.6970129584889087, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.5673158168792725, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7422800660133362, + "num_tokens": 635373940.0, + "step": 24559 + }, + { + "epoch": 2.697122776191522, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.0251410007476807, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7205743789672852, + "num_tokens": 635407931.0, + "step": 24560 + }, + { + "epoch": 2.6972325938941357, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.1498398780822754, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7243667840957642, + "num_tokens": 635437344.0, + "step": 24561 + }, + { + "epoch": 2.6973424115967495, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.4674618244171143, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7171394228935242, + "num_tokens": 635462421.0, + "step": 24562 + }, + { + "epoch": 2.697452229299363, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.81596302986145, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7461605668067932, + "num_tokens": 635481273.0, + "step": 24563 + }, + { + "epoch": 2.697562047001977, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.431769847869873, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7245959639549255, + "num_tokens": 635505387.0, + "step": 24564 + }, + { + "epoch": 2.6976718647045903, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.313521385192871, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7098889350891113, + "num_tokens": 635533811.0, + "step": 24565 + }, + { + "epoch": 2.697781682407204, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2895655632019043, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.730961799621582, + "num_tokens": 635561271.0, + "step": 24566 + }, + { + "epoch": 2.697891500109818, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.4585814476013184, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.720227837562561, + "num_tokens": 635584208.0, + "step": 24567 + }, + { + "epoch": 2.698001317812431, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.25553035736084, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7531594038009644, + "num_tokens": 635611111.0, + "step": 24568 + }, + { + "epoch": 2.698111135515045, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2800731658935547, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7366257309913635, + "num_tokens": 635638897.0, + "step": 24569 + }, + { + "epoch": 2.6982209532176586, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.113278865814209, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7304041385650635, + "num_tokens": 635670072.0, + "step": 24570 + }, + { + "epoch": 2.6983307709202724, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.857588768005371, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7471508979797363, + "num_tokens": 635688613.0, + "step": 24571 + }, + { + "epoch": 2.698440588622886, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.539276361465454, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7411247491836548, + "num_tokens": 635711296.0, + "step": 24572 + }, + { + "epoch": 2.6985504063254995, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2130916118621826, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7304661273956299, + "num_tokens": 635741297.0, + "step": 24573 + }, + { + "epoch": 2.6986602240281132, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.1335160732269287, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7170724868774414, + "num_tokens": 635776310.0, + "step": 24574 + }, + { + "epoch": 2.698770041730727, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.415165424346924, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7251071333885193, + "num_tokens": 635799526.0, + "step": 24575 + }, + { + "epoch": 2.6988798594333407, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.3694167137145996, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7179498672485352, + "num_tokens": 635827342.0, + "step": 24576 + }, + { + "epoch": 2.6989896771359545, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.3490192890167236, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7297022342681885, + "num_tokens": 635853904.0, + "step": 24577 + }, + { + "epoch": 2.699099494838568, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.3475351333618164, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.727960467338562, + "num_tokens": 635882300.0, + "step": 24578 + }, + { + "epoch": 2.6992093125411816, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2016797065734863, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7419501543045044, + "num_tokens": 635911264.0, + "step": 24579 + }, + { + "epoch": 2.6993191302437953, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.2255020141601562, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7336747646331787, + "num_tokens": 635937409.0, + "step": 24580 + }, + { + "epoch": 2.699428947946409, + "ewc_loss": 0.00021648406982421875, + "grad_norm": 2.34515380859375, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7233337163925171, + "num_tokens": 635962176.0, + "step": 24581 + }, + { + "epoch": 2.699538765649023, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.501378059387207, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7384845018386841, + "num_tokens": 635985408.0, + "step": 24582 + }, + { + "epoch": 2.699648583351636, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5414304733276367, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7482922077178955, + "num_tokens": 636008460.0, + "step": 24583 + }, + { + "epoch": 2.69975840105425, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4709644317626953, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.715458869934082, + "num_tokens": 636033600.0, + "step": 24584 + }, + { + "epoch": 2.6998682187568637, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.105182647705078, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7023195624351501, + "num_tokens": 636064531.0, + "step": 24585 + }, + { + "epoch": 2.6999780364594774, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4455254077911377, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7341392040252686, + "num_tokens": 636088199.0, + "step": 24586 + }, + { + "epoch": 2.700087854162091, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3798940181732178, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7346705198287964, + "num_tokens": 636114960.0, + "step": 24587 + }, + { + "epoch": 2.7001976718647045, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.360837697982788, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7099185585975647, + "num_tokens": 636142518.0, + "step": 24588 + }, + { + "epoch": 2.7003074895673183, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2188239097595215, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7281010150909424, + "num_tokens": 636170558.0, + "step": 24589 + }, + { + "epoch": 2.700417307269932, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4403114318847656, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7369372248649597, + "num_tokens": 636195178.0, + "step": 24590 + }, + { + "epoch": 2.7005271249725453, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.55945086479187, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7009629011154175, + "num_tokens": 636219507.0, + "step": 24591 + }, + { + "epoch": 2.700636942675159, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2316770553588867, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7138222455978394, + "num_tokens": 636249262.0, + "step": 24592 + }, + { + "epoch": 2.700746760377773, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1985011100769043, + "learning_rate": 1e-06, + "loss": 1.0205, + "mean_token_accuracy": 0.6939084529876709, + "num_tokens": 636282270.0, + "step": 24593 + }, + { + "epoch": 2.7008565780803866, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.369746685028076, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.708059549331665, + "num_tokens": 636309437.0, + "step": 24594 + }, + { + "epoch": 2.7009663957830004, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4386632442474365, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7460546493530273, + "num_tokens": 636334045.0, + "step": 24595 + }, + { + "epoch": 2.7010762134856137, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.511272430419922, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7264180183410645, + "num_tokens": 636356655.0, + "step": 24596 + }, + { + "epoch": 2.7011860311882274, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.370875120162964, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7517192959785461, + "num_tokens": 636380594.0, + "step": 24597 + }, + { + "epoch": 2.701295848890841, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2282330989837646, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7250195741653442, + "num_tokens": 636409274.0, + "step": 24598 + }, + { + "epoch": 2.701405666593455, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.508023977279663, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7433716058731079, + "num_tokens": 636431881.0, + "step": 24599 + }, + { + "epoch": 2.7015154842960687, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5773258209228516, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.719053328037262, + "num_tokens": 636456702.0, + "step": 24600 + }, + { + "epoch": 2.701625301998682, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.51617431640625, + "learning_rate": 1e-06, + "loss": 0.7504, + "mean_token_accuracy": 0.7624180316925049, + "num_tokens": 636477902.0, + "step": 24601 + }, + { + "epoch": 2.7017351197012958, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3201162815093994, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7264804244041443, + "num_tokens": 636506103.0, + "step": 24602 + }, + { + "epoch": 2.7018449374039095, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1999056339263916, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7336562871932983, + "num_tokens": 636534483.0, + "step": 24603 + }, + { + "epoch": 2.7019547551065233, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3566324710845947, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7325657606124878, + "num_tokens": 636560499.0, + "step": 24604 + }, + { + "epoch": 2.702064572809137, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.360105514526367, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7285633087158203, + "num_tokens": 636587021.0, + "step": 24605 + }, + { + "epoch": 2.7021743905117503, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3569376468658447, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7319414615631104, + "num_tokens": 636611808.0, + "step": 24606 + }, + { + "epoch": 2.702284208214364, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.137160539627075, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7251349687576294, + "num_tokens": 636642235.0, + "step": 24607 + }, + { + "epoch": 2.702394025916978, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3632853031158447, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7210516929626465, + "num_tokens": 636668339.0, + "step": 24608 + }, + { + "epoch": 2.7025038436195916, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4052581787109375, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7218806147575378, + "num_tokens": 636694031.0, + "step": 24609 + }, + { + "epoch": 2.7026136613222054, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2761056423187256, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7352569103240967, + "num_tokens": 636721196.0, + "step": 24610 + }, + { + "epoch": 2.7027234790248187, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.455911874771118, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7213070392608643, + "num_tokens": 636744077.0, + "step": 24611 + }, + { + "epoch": 2.7028332967274324, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3197925090789795, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7254227995872498, + "num_tokens": 636771047.0, + "step": 24612 + }, + { + "epoch": 2.702943114430046, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2640621662139893, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7426018714904785, + "num_tokens": 636797984.0, + "step": 24613 + }, + { + "epoch": 2.7030529321326595, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.597977638244629, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.740906834602356, + "num_tokens": 636820040.0, + "step": 24614 + }, + { + "epoch": 2.7031627498352737, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2725489139556885, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7497237920761108, + "num_tokens": 636847213.0, + "step": 24615 + }, + { + "epoch": 2.703272567537887, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4450879096984863, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7316344976425171, + "num_tokens": 636873425.0, + "step": 24616 + }, + { + "epoch": 2.703382385240501, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4904744625091553, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7347111105918884, + "num_tokens": 636895224.0, + "step": 24617 + }, + { + "epoch": 2.7034922029431145, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.152609348297119, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7082220911979675, + "num_tokens": 636925505.0, + "step": 24618 + }, + { + "epoch": 2.703602020645728, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3022947311401367, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7318594455718994, + "num_tokens": 636950985.0, + "step": 24619 + }, + { + "epoch": 2.7037118383483416, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.6217901706695557, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7352519035339355, + "num_tokens": 636972902.0, + "step": 24620 + }, + { + "epoch": 2.7038216560509554, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5007989406585693, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.728844165802002, + "num_tokens": 636995027.0, + "step": 24621 + }, + { + "epoch": 2.703931473753569, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2325191497802734, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7110697627067566, + "num_tokens": 637025443.0, + "step": 24622 + }, + { + "epoch": 2.704041291456183, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3605990409851074, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7340998649597168, + "num_tokens": 637053295.0, + "step": 24623 + }, + { + "epoch": 2.704151109158796, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3947317600250244, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7343049645423889, + "num_tokens": 637078599.0, + "step": 24624 + }, + { + "epoch": 2.70426092686141, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1175875663757324, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7211964130401611, + "num_tokens": 637109623.0, + "step": 24625 + }, + { + "epoch": 2.7043707445640237, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4073987007141113, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7275295257568359, + "num_tokens": 637132960.0, + "step": 24626 + }, + { + "epoch": 2.7044805622666375, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3178586959838867, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7396059632301331, + "num_tokens": 637159563.0, + "step": 24627 + }, + { + "epoch": 2.7045903799692512, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1149027347564697, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7437185645103455, + "num_tokens": 637189795.0, + "step": 24628 + }, + { + "epoch": 2.7047001976718645, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2331035137176514, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7157506942749023, + "num_tokens": 637217854.0, + "step": 24629 + }, + { + "epoch": 2.7048100153744783, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.252868413925171, + "learning_rate": 1e-06, + "loss": 0.7406, + "mean_token_accuracy": 0.7593973875045776, + "num_tokens": 637246068.0, + "step": 24630 + }, + { + "epoch": 2.704919833077092, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.243274450302124, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7173069715499878, + "num_tokens": 637275907.0, + "step": 24631 + }, + { + "epoch": 2.705029650779706, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5632219314575195, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7271736860275269, + "num_tokens": 637298231.0, + "step": 24632 + }, + { + "epoch": 2.7051394684823196, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.350712537765503, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7404102087020874, + "num_tokens": 637324818.0, + "step": 24633 + }, + { + "epoch": 2.705249286184933, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2114901542663574, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7407872080802917, + "num_tokens": 637351693.0, + "step": 24634 + }, + { + "epoch": 2.7053591038875466, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4039499759674072, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7501873970031738, + "num_tokens": 637376108.0, + "step": 24635 + }, + { + "epoch": 2.7054689215901604, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5604116916656494, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7261070609092712, + "num_tokens": 637402645.0, + "step": 24636 + }, + { + "epoch": 2.705578739292774, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4315249919891357, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7157862186431885, + "num_tokens": 637427056.0, + "step": 24637 + }, + { + "epoch": 2.705688556995388, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.203251361846924, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7199227213859558, + "num_tokens": 637457793.0, + "step": 24638 + }, + { + "epoch": 2.705798374698001, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.35518479347229, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7451608180999756, + "num_tokens": 637481962.0, + "step": 24639 + }, + { + "epoch": 2.705908192400615, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.7426440715789795, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7468873262405396, + "num_tokens": 637502153.0, + "step": 24640 + }, + { + "epoch": 2.7060180101032287, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3060667514801025, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7142145037651062, + "num_tokens": 637529072.0, + "step": 24641 + }, + { + "epoch": 2.706127827805842, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.78963303565979, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7509684562683105, + "num_tokens": 637549332.0, + "step": 24642 + }, + { + "epoch": 2.706237645508456, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.7140824794769287, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.7570885419845581, + "num_tokens": 637571205.0, + "step": 24643 + }, + { + "epoch": 2.7063474632110696, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.356069803237915, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7192076444625854, + "num_tokens": 637599018.0, + "step": 24644 + }, + { + "epoch": 2.7064572809136833, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3997819423675537, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7118808031082153, + "num_tokens": 637626618.0, + "step": 24645 + }, + { + "epoch": 2.706567098616297, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.178820848464966, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7236456871032715, + "num_tokens": 637655494.0, + "step": 24646 + }, + { + "epoch": 2.7066769163189104, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3919310569763184, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7275500297546387, + "num_tokens": 637680457.0, + "step": 24647 + }, + { + "epoch": 2.706786734021524, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.270357608795166, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7312865257263184, + "num_tokens": 637708205.0, + "step": 24648 + }, + { + "epoch": 2.706896551724138, + "ewc_loss": 0.000217437744140625, + "grad_norm": 3.157276153564453, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.749362587928772, + "num_tokens": 637725835.0, + "step": 24649 + }, + { + "epoch": 2.7070063694267517, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.115008592605591, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7276027798652649, + "num_tokens": 637757318.0, + "step": 24650 + }, + { + "epoch": 2.7071161871293654, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1854965686798096, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7230190634727478, + "num_tokens": 637786045.0, + "step": 24651 + }, + { + "epoch": 2.7072260048319787, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.317169666290283, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.732366144657135, + "num_tokens": 637814344.0, + "step": 24652 + }, + { + "epoch": 2.7073358225345925, + "ewc_loss": 0.000217437744140625, + "grad_norm": 1.9643656015396118, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7162142992019653, + "num_tokens": 637851542.0, + "step": 24653 + }, + { + "epoch": 2.7074456402372062, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.310211420059204, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6966994404792786, + "num_tokens": 637880353.0, + "step": 24654 + }, + { + "epoch": 2.70755545793982, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3402061462402344, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7257901430130005, + "num_tokens": 637905596.0, + "step": 24655 + }, + { + "epoch": 2.7076652756424338, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.422255754470825, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7258326411247253, + "num_tokens": 637930958.0, + "step": 24656 + }, + { + "epoch": 2.707775093345047, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5986666679382324, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.741122841835022, + "num_tokens": 637952486.0, + "step": 24657 + }, + { + "epoch": 2.707884911047661, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3919472694396973, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7084547281265259, + "num_tokens": 637980785.0, + "step": 24658 + }, + { + "epoch": 2.7079947287502746, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1679091453552246, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7189241647720337, + "num_tokens": 638012640.0, + "step": 24659 + }, + { + "epoch": 2.7081045464528883, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.597196578979492, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7488420009613037, + "num_tokens": 638035306.0, + "step": 24660 + }, + { + "epoch": 2.708214364155502, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.288654088973999, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7266541719436646, + "num_tokens": 638064193.0, + "step": 24661 + }, + { + "epoch": 2.7083241818581154, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.454022169113159, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7458465099334717, + "num_tokens": 638088972.0, + "step": 24662 + }, + { + "epoch": 2.708433999560729, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.677494525909424, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7382590770721436, + "num_tokens": 638110260.0, + "step": 24663 + }, + { + "epoch": 2.708543817263343, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.6141908168792725, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7356734275817871, + "num_tokens": 638132932.0, + "step": 24664 + }, + { + "epoch": 2.7086536349659562, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4079325199127197, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.7038832306861877, + "num_tokens": 638159454.0, + "step": 24665 + }, + { + "epoch": 2.7087634526685704, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4153804779052734, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7379338145256042, + "num_tokens": 638183545.0, + "step": 24666 + }, + { + "epoch": 2.7088732703711838, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.045559883117676, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7295600771903992, + "num_tokens": 638215187.0, + "step": 24667 + }, + { + "epoch": 2.7089830880737975, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4477744102478027, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.726197361946106, + "num_tokens": 638240218.0, + "step": 24668 + }, + { + "epoch": 2.7090929057764113, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1232352256774902, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7301641702651978, + "num_tokens": 638271344.0, + "step": 24669 + }, + { + "epoch": 2.7092027234790246, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.0065691471099854, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7246352434158325, + "num_tokens": 638303577.0, + "step": 24670 + }, + { + "epoch": 2.7093125411816383, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.315406560897827, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7131761312484741, + "num_tokens": 638330973.0, + "step": 24671 + }, + { + "epoch": 2.709422358884252, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.25209379196167, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7307378649711609, + "num_tokens": 638357139.0, + "step": 24672 + }, + { + "epoch": 2.709532176586866, + "ewc_loss": 0.000217437744140625, + "grad_norm": 3.1882638931274414, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7359713315963745, + "num_tokens": 638373815.0, + "step": 24673 + }, + { + "epoch": 2.7096419942894796, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5217444896698, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7043848037719727, + "num_tokens": 638399938.0, + "step": 24674 + }, + { + "epoch": 2.709751811992093, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5198163986206055, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.741710364818573, + "num_tokens": 638422808.0, + "step": 24675 + }, + { + "epoch": 2.7098616296947067, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1113779544830322, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7561920285224915, + "num_tokens": 638450788.0, + "step": 24676 + }, + { + "epoch": 2.7099714473973204, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.350231409072876, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.728782594203949, + "num_tokens": 638478890.0, + "step": 24677 + }, + { + "epoch": 2.710081265099934, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.143005609512329, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7356255650520325, + "num_tokens": 638510335.0, + "step": 24678 + }, + { + "epoch": 2.710191082802548, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4614675045013428, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7318514585494995, + "num_tokens": 638533854.0, + "step": 24679 + }, + { + "epoch": 2.7103009005051613, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2985925674438477, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7109808921813965, + "num_tokens": 638561507.0, + "step": 24680 + }, + { + "epoch": 2.710410718207775, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4401934146881104, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7338955402374268, + "num_tokens": 638586013.0, + "step": 24681 + }, + { + "epoch": 2.7105205359103888, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.374138116836548, + "learning_rate": 1e-06, + "loss": 0.7829, + "mean_token_accuracy": 0.7513118982315063, + "num_tokens": 638609740.0, + "step": 24682 + }, + { + "epoch": 2.7106303536130025, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4367759227752686, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.721106767654419, + "num_tokens": 638636714.0, + "step": 24683 + }, + { + "epoch": 2.7107401713156163, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.4476065635681152, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7430850863456726, + "num_tokens": 638661166.0, + "step": 24684 + }, + { + "epoch": 2.7108499890182296, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.1551010608673096, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.719147801399231, + "num_tokens": 638689317.0, + "step": 24685 + }, + { + "epoch": 2.7109598067208434, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.441258192062378, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7091642022132874, + "num_tokens": 638714501.0, + "step": 24686 + }, + { + "epoch": 2.711069624423457, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.495028018951416, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.716705322265625, + "num_tokens": 638739101.0, + "step": 24687 + }, + { + "epoch": 2.711179442126071, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.331284284591675, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7143197059631348, + "num_tokens": 638766889.0, + "step": 24688 + }, + { + "epoch": 2.7112892598286846, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5571768283843994, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7325084209442139, + "num_tokens": 638789471.0, + "step": 24689 + }, + { + "epoch": 2.711399077531298, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.6069138050079346, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7364069819450378, + "num_tokens": 638810942.0, + "step": 24690 + }, + { + "epoch": 2.7115088952339117, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.107102870941162, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7304890155792236, + "num_tokens": 638843824.0, + "step": 24691 + }, + { + "epoch": 2.7116187129365255, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.003944158554077, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7281431555747986, + "num_tokens": 638874574.0, + "step": 24692 + }, + { + "epoch": 2.7117285306391388, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.222490072250366, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7259324193000793, + "num_tokens": 638903284.0, + "step": 24693 + }, + { + "epoch": 2.7118383483417525, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5076284408569336, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7170329093933105, + "num_tokens": 638928789.0, + "step": 24694 + }, + { + "epoch": 2.7119481660443663, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.118964433670044, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7068912386894226, + "num_tokens": 638960128.0, + "step": 24695 + }, + { + "epoch": 2.71205798374698, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2836999893188477, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.732843816280365, + "num_tokens": 638986557.0, + "step": 24696 + }, + { + "epoch": 2.712167801449594, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.038360595703125, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7134010791778564, + "num_tokens": 639019260.0, + "step": 24697 + }, + { + "epoch": 2.712277619152207, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.0505685806274414, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7242943048477173, + "num_tokens": 639051229.0, + "step": 24698 + }, + { + "epoch": 2.712387436854821, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.247575044631958, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7242771983146667, + "num_tokens": 639080680.0, + "step": 24699 + }, + { + "epoch": 2.7124972545574346, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.0898423194885254, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7381193041801453, + "num_tokens": 639112770.0, + "step": 24700 + }, + { + "epoch": 2.7126070722600484, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.7960004806518555, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7324739694595337, + "num_tokens": 639132757.0, + "step": 24701 + }, + { + "epoch": 2.712716889962662, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.457707643508911, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7459962964057922, + "num_tokens": 639155224.0, + "step": 24702 + }, + { + "epoch": 2.7128267076652754, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.3586883544921875, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7285049557685852, + "num_tokens": 639180807.0, + "step": 24703 + }, + { + "epoch": 2.712936525367889, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.8056435585021973, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7227345705032349, + "num_tokens": 639200014.0, + "step": 24704 + }, + { + "epoch": 2.713046343070503, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.739284038543701, + "learning_rate": 1e-06, + "loss": 0.7582, + "mean_token_accuracy": 0.7565664052963257, + "num_tokens": 639219260.0, + "step": 24705 + }, + { + "epoch": 2.7131561607731167, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.166822910308838, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7194638848304749, + "num_tokens": 639250650.0, + "step": 24706 + }, + { + "epoch": 2.7132659784757305, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.418384313583374, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7366426587104797, + "num_tokens": 639275641.0, + "step": 24707 + }, + { + "epoch": 2.713375796178344, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.5641252994537354, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7335343360900879, + "num_tokens": 639297757.0, + "step": 24708 + }, + { + "epoch": 2.7134856138809575, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.2002999782562256, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7303723096847534, + "num_tokens": 639325714.0, + "step": 24709 + }, + { + "epoch": 2.7135954315835713, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.674168586730957, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7401808500289917, + "num_tokens": 639345003.0, + "step": 24710 + }, + { + "epoch": 2.713705249286185, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.390723943710327, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7305504083633423, + "num_tokens": 639369979.0, + "step": 24711 + }, + { + "epoch": 2.713815066988799, + "ewc_loss": 0.000217437744140625, + "grad_norm": 2.257657289505005, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7262388467788696, + "num_tokens": 639398053.0, + "step": 24712 + }, + { + "epoch": 2.713924884691412, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.163872480392456, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7127748727798462, + "num_tokens": 639428209.0, + "step": 24713 + }, + { + "epoch": 2.714034702394026, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.185108184814453, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7391325235366821, + "num_tokens": 639454434.0, + "step": 24714 + }, + { + "epoch": 2.7141445200966396, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4125170707702637, + "learning_rate": 1e-06, + "loss": 0.7335, + "mean_token_accuracy": 0.7624862194061279, + "num_tokens": 639476779.0, + "step": 24715 + }, + { + "epoch": 2.7142543377992534, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3054559230804443, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7372841238975525, + "num_tokens": 639503392.0, + "step": 24716 + }, + { + "epoch": 2.714364155501867, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.354292869567871, + "learning_rate": 1e-06, + "loss": 0.7694, + "mean_token_accuracy": 0.7568113803863525, + "num_tokens": 639527342.0, + "step": 24717 + }, + { + "epoch": 2.7144739732044805, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5140578746795654, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7404054403305054, + "num_tokens": 639550694.0, + "step": 24718 + }, + { + "epoch": 2.7145837909070942, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 1.9971725940704346, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.729179322719574, + "num_tokens": 639585039.0, + "step": 24719 + }, + { + "epoch": 2.714693608609708, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6410882472991943, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7452304363250732, + "num_tokens": 639607137.0, + "step": 24720 + }, + { + "epoch": 2.7148034263123213, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.195913791656494, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7303329706192017, + "num_tokens": 639635501.0, + "step": 24721 + }, + { + "epoch": 2.714913244014935, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.257362127304077, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.715538501739502, + "num_tokens": 639665461.0, + "step": 24722 + }, + { + "epoch": 2.715023061717549, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.280182123184204, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.714390754699707, + "num_tokens": 639693491.0, + "step": 24723 + }, + { + "epoch": 2.7151328794201626, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.672804117202759, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.733254611492157, + "num_tokens": 639715777.0, + "step": 24724 + }, + { + "epoch": 2.7152426971227763, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3925507068634033, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7289487719535828, + "num_tokens": 639741348.0, + "step": 24725 + }, + { + "epoch": 2.7153525148253896, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4989311695098877, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7327412366867065, + "num_tokens": 639764709.0, + "step": 24726 + }, + { + "epoch": 2.7154623325280034, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.143216133117676, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7431545257568359, + "num_tokens": 639793037.0, + "step": 24727 + }, + { + "epoch": 2.715572150230617, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.344271659851074, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7130281925201416, + "num_tokens": 639820215.0, + "step": 24728 + }, + { + "epoch": 2.715681967933231, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5485148429870605, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7248877882957458, + "num_tokens": 639842429.0, + "step": 24729 + }, + { + "epoch": 2.7157917856358447, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.197998046875, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7257853746414185, + "num_tokens": 639873513.0, + "step": 24730 + }, + { + "epoch": 2.715901603338458, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3813891410827637, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7107200622558594, + "num_tokens": 639899481.0, + "step": 24731 + }, + { + "epoch": 2.7160114210410717, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4026687145233154, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7276871204376221, + "num_tokens": 639924799.0, + "step": 24732 + }, + { + "epoch": 2.7161212387436855, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2871947288513184, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7256948947906494, + "num_tokens": 639951117.0, + "step": 24733 + }, + { + "epoch": 2.7162310564462993, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.329692840576172, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7338957190513611, + "num_tokens": 639974759.0, + "step": 24734 + }, + { + "epoch": 2.716340874148913, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.7077927589416504, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.722791850566864, + "num_tokens": 639997730.0, + "step": 24735 + }, + { + "epoch": 2.7164506918515263, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5946743488311768, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7351675033569336, + "num_tokens": 640020029.0, + "step": 24736 + }, + { + "epoch": 2.71656050955414, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.321525812149048, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7335253953933716, + "num_tokens": 640048122.0, + "step": 24737 + }, + { + "epoch": 2.716670327256754, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.185241937637329, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7102137804031372, + "num_tokens": 640077578.0, + "step": 24738 + }, + { + "epoch": 2.7167801449593676, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.345613479614258, + "learning_rate": 1e-06, + "loss": 0.7367, + "mean_token_accuracy": 0.7625002861022949, + "num_tokens": 640102825.0, + "step": 24739 + }, + { + "epoch": 2.7168899626619814, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2959160804748535, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7384997606277466, + "num_tokens": 640132048.0, + "step": 24740 + }, + { + "epoch": 2.7169997803645947, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6810216903686523, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7296444773674011, + "num_tokens": 640156823.0, + "step": 24741 + }, + { + "epoch": 2.7171095980672084, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.529115915298462, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7219563126564026, + "num_tokens": 640180505.0, + "step": 24742 + }, + { + "epoch": 2.717219415769822, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6862711906433105, + "learning_rate": 1e-06, + "loss": 0.7577, + "mean_token_accuracy": 0.7586798667907715, + "num_tokens": 640199882.0, + "step": 24743 + }, + { + "epoch": 2.7173292334724355, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3203577995300293, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7401571869850159, + "num_tokens": 640228809.0, + "step": 24744 + }, + { + "epoch": 2.7174390511750497, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.380091667175293, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7160216569900513, + "num_tokens": 640256072.0, + "step": 24745 + }, + { + "epoch": 2.717548868877663, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3055431842803955, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7499077320098877, + "num_tokens": 640282519.0, + "step": 24746 + }, + { + "epoch": 2.7176586865802768, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3783199787139893, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7171240448951721, + "num_tokens": 640307450.0, + "step": 24747 + }, + { + "epoch": 2.7177685042828905, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5274202823638916, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7343218326568604, + "num_tokens": 640329891.0, + "step": 24748 + }, + { + "epoch": 2.717878321985504, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0353589057922363, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6997635364532471, + "num_tokens": 640364993.0, + "step": 24749 + }, + { + "epoch": 2.7179881396881176, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.320648431777954, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7290163040161133, + "num_tokens": 640392942.0, + "step": 24750 + }, + { + "epoch": 2.7180979573907313, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3049752712249756, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7216157913208008, + "num_tokens": 640419676.0, + "step": 24751 + }, + { + "epoch": 2.718207775093345, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.249300241470337, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7378733158111572, + "num_tokens": 640448091.0, + "step": 24752 + }, + { + "epoch": 2.718317592795959, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3237392902374268, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7159121632575989, + "num_tokens": 640477778.0, + "step": 24753 + }, + { + "epoch": 2.718427410498572, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4632694721221924, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7310110330581665, + "num_tokens": 640500398.0, + "step": 24754 + }, + { + "epoch": 2.718537228201186, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2570300102233887, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7400388121604919, + "num_tokens": 640528757.0, + "step": 24755 + }, + { + "epoch": 2.7186470459037997, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3558948040008545, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.705979585647583, + "num_tokens": 640554774.0, + "step": 24756 + }, + { + "epoch": 2.7187568636064134, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.279682159423828, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7201399207115173, + "num_tokens": 640582467.0, + "step": 24757 + }, + { + "epoch": 2.718866681309027, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.186020612716675, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7201380729675293, + "num_tokens": 640614279.0, + "step": 24758 + }, + { + "epoch": 2.7189764990116405, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2217297554016113, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7220068573951721, + "num_tokens": 640642335.0, + "step": 24759 + }, + { + "epoch": 2.7190863167142543, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.294410467147827, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7100871801376343, + "num_tokens": 640669322.0, + "step": 24760 + }, + { + "epoch": 2.719196134416868, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.9193294048309326, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7302721738815308, + "num_tokens": 640687288.0, + "step": 24761 + }, + { + "epoch": 2.719305952119482, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.509828805923462, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7282721400260925, + "num_tokens": 640710975.0, + "step": 24762 + }, + { + "epoch": 2.7194157698220955, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 1.9281251430511475, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7257688045501709, + "num_tokens": 640747982.0, + "step": 24763 + }, + { + "epoch": 2.719525587524709, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.352900266647339, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7541989684104919, + "num_tokens": 640775056.0, + "step": 24764 + }, + { + "epoch": 2.7196354052273226, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.272472620010376, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7178702354431152, + "num_tokens": 640802794.0, + "step": 24765 + }, + { + "epoch": 2.7197452229299364, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3977487087249756, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7176854014396667, + "num_tokens": 640828970.0, + "step": 24766 + }, + { + "epoch": 2.71985504063255, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.342766761779785, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7357255816459656, + "num_tokens": 640856830.0, + "step": 24767 + }, + { + "epoch": 2.719964858335164, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.176924228668213, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7082716226577759, + "num_tokens": 640888052.0, + "step": 24768 + }, + { + "epoch": 2.720074676037777, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.386777400970459, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.744118332862854, + "num_tokens": 640912422.0, + "step": 24769 + }, + { + "epoch": 2.720184493740391, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6088132858276367, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7353814840316772, + "num_tokens": 640933340.0, + "step": 24770 + }, + { + "epoch": 2.7202943114430047, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.7063183784484863, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7518947124481201, + "num_tokens": 640953858.0, + "step": 24771 + }, + { + "epoch": 2.720404129145618, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3890440464019775, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7317514419555664, + "num_tokens": 640979591.0, + "step": 24772 + }, + { + "epoch": 2.720513946848232, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2454826831817627, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7292009592056274, + "num_tokens": 641005981.0, + "step": 24773 + }, + { + "epoch": 2.7206237645508455, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4930942058563232, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7441569566726685, + "num_tokens": 641030596.0, + "step": 24774 + }, + { + "epoch": 2.7207335822534593, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.352180004119873, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7060955166816711, + "num_tokens": 641057299.0, + "step": 24775 + }, + { + "epoch": 2.720843399956073, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.522901773452759, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7279499173164368, + "num_tokens": 641083248.0, + "step": 24776 + }, + { + "epoch": 2.7209532176586864, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.400325059890747, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.701850414276123, + "num_tokens": 641110826.0, + "step": 24777 + }, + { + "epoch": 2.7210630353613, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3926124572753906, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7461588382720947, + "num_tokens": 641138748.0, + "step": 24778 + }, + { + "epoch": 2.721172853063914, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3240911960601807, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7170208692550659, + "num_tokens": 641167288.0, + "step": 24779 + }, + { + "epoch": 2.7212826707665276, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2231173515319824, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7219300270080566, + "num_tokens": 641196385.0, + "step": 24780 + }, + { + "epoch": 2.7213924884691414, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.385810136795044, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7225790619850159, + "num_tokens": 641222366.0, + "step": 24781 + }, + { + "epoch": 2.7215023061717547, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0398244857788086, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7237981557846069, + "num_tokens": 641255955.0, + "step": 24782 + }, + { + "epoch": 2.7216121238743685, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5067853927612305, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7162424325942993, + "num_tokens": 641280874.0, + "step": 24783 + }, + { + "epoch": 2.721721941576982, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.492572546005249, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7377086877822876, + "num_tokens": 641303592.0, + "step": 24784 + }, + { + "epoch": 2.721831759279596, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.050158739089966, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6952086687088013, + "num_tokens": 641336893.0, + "step": 24785 + }, + { + "epoch": 2.7219415769822097, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5064666271209717, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7373931407928467, + "num_tokens": 641359312.0, + "step": 24786 + }, + { + "epoch": 2.722051394684823, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4571118354797363, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7488995790481567, + "num_tokens": 641383693.0, + "step": 24787 + }, + { + "epoch": 2.722161212387437, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.522857189178467, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7321383953094482, + "num_tokens": 641407736.0, + "step": 24788 + }, + { + "epoch": 2.7222710300900506, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.343801259994507, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7034174799919128, + "num_tokens": 641433221.0, + "step": 24789 + }, + { + "epoch": 2.7223808477926643, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.503188371658325, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7273915410041809, + "num_tokens": 641456588.0, + "step": 24790 + }, + { + "epoch": 2.722490665495278, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0306344032287598, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7107611298561096, + "num_tokens": 641491270.0, + "step": 24791 + }, + { + "epoch": 2.7226004831978914, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1164469718933105, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7197773456573486, + "num_tokens": 641524657.0, + "step": 24792 + }, + { + "epoch": 2.722710300900505, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.29133939743042, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7112439870834351, + "num_tokens": 641551827.0, + "step": 24793 + }, + { + "epoch": 2.722820118603119, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.860109567642212, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7569612264633179, + "num_tokens": 641570975.0, + "step": 24794 + }, + { + "epoch": 2.722929936305732, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.651718854904175, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7363133430480957, + "num_tokens": 641591883.0, + "step": 24795 + }, + { + "epoch": 2.7230397540083464, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5239858627319336, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7278643846511841, + "num_tokens": 641615483.0, + "step": 24796 + }, + { + "epoch": 2.7231495717109597, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3130042552948, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.749225378036499, + "num_tokens": 641640463.0, + "step": 24797 + }, + { + "epoch": 2.7232593894135735, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.262072801589966, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7170171737670898, + "num_tokens": 641669341.0, + "step": 24798 + }, + { + "epoch": 2.7233692071161872, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.444614887237549, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7096550464630127, + "num_tokens": 641695953.0, + "step": 24799 + }, + { + "epoch": 2.7234790248188006, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5038278102874756, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7255116105079651, + "num_tokens": 641720964.0, + "step": 24800 + }, + { + "epoch": 2.7235888425214143, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1085364818573, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7167927026748657, + "num_tokens": 641752724.0, + "step": 24801 + }, + { + "epoch": 2.723698660224028, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6404051780700684, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7516148686408997, + "num_tokens": 641774184.0, + "step": 24802 + }, + { + "epoch": 2.723808477926642, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.236694097518921, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7185544371604919, + "num_tokens": 641802550.0, + "step": 24803 + }, + { + "epoch": 2.7239182956292556, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.668912410736084, + "learning_rate": 1e-06, + "loss": 0.7591, + "mean_token_accuracy": 0.7550339102745056, + "num_tokens": 641822945.0, + "step": 24804 + }, + { + "epoch": 2.724028113331869, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3986127376556396, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7452619671821594, + "num_tokens": 641848120.0, + "step": 24805 + }, + { + "epoch": 2.7241379310344827, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2543089389801025, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7465346455574036, + "num_tokens": 641874104.0, + "step": 24806 + }, + { + "epoch": 2.7242477487370964, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4303665161132812, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.738274335861206, + "num_tokens": 641898951.0, + "step": 24807 + }, + { + "epoch": 2.72435756643971, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.501725673675537, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7457300424575806, + "num_tokens": 641921362.0, + "step": 24808 + }, + { + "epoch": 2.724467384142324, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 3.063964366912842, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7429357767105103, + "num_tokens": 641938341.0, + "step": 24809 + }, + { + "epoch": 2.7245772018449372, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0655806064605713, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7099940776824951, + "num_tokens": 641970748.0, + "step": 24810 + }, + { + "epoch": 2.724687019547551, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2320175170898438, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7088461518287659, + "num_tokens": 642001075.0, + "step": 24811 + }, + { + "epoch": 2.7247968372501647, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3237464427948, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7488449811935425, + "num_tokens": 642026261.0, + "step": 24812 + }, + { + "epoch": 2.7249066549527785, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2622973918914795, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7173447012901306, + "num_tokens": 642055981.0, + "step": 24813 + }, + { + "epoch": 2.7250164726553923, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3141260147094727, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7062926292419434, + "num_tokens": 642082195.0, + "step": 24814 + }, + { + "epoch": 2.7251262903580056, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.114948034286499, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7265437841415405, + "num_tokens": 642114405.0, + "step": 24815 + }, + { + "epoch": 2.7252361080606193, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3298656940460205, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7186861038208008, + "num_tokens": 642139251.0, + "step": 24816 + }, + { + "epoch": 2.725345925763233, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.515141010284424, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7394084334373474, + "num_tokens": 642163813.0, + "step": 24817 + }, + { + "epoch": 2.725455743465847, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2737810611724854, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7355708479881287, + "num_tokens": 642189645.0, + "step": 24818 + }, + { + "epoch": 2.7255655611684606, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1518943309783936, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.732089102268219, + "num_tokens": 642219610.0, + "step": 24819 + }, + { + "epoch": 2.725675378871074, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6608529090881348, + "learning_rate": 1e-06, + "loss": 0.828, + "mean_token_accuracy": 0.7357461452484131, + "num_tokens": 642240897.0, + "step": 24820 + }, + { + "epoch": 2.7257851965736877, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.285809278488159, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.722583532333374, + "num_tokens": 642270152.0, + "step": 24821 + }, + { + "epoch": 2.7258950142763014, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.207941770553589, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.728413999080658, + "num_tokens": 642303524.0, + "step": 24822 + }, + { + "epoch": 2.7260048319789147, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.541862726211548, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7257981896400452, + "num_tokens": 642326904.0, + "step": 24823 + }, + { + "epoch": 2.7261146496815285, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.44629168510437, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7272149920463562, + "num_tokens": 642352381.0, + "step": 24824 + }, + { + "epoch": 2.7262244673841423, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3834547996520996, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7288223505020142, + "num_tokens": 642378034.0, + "step": 24825 + }, + { + "epoch": 2.726334285086756, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3796658515930176, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7394023537635803, + "num_tokens": 642403044.0, + "step": 24826 + }, + { + "epoch": 2.7264441027893698, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4763917922973633, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.722460150718689, + "num_tokens": 642426545.0, + "step": 24827 + }, + { + "epoch": 2.726553920491983, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.301609516143799, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7534395456314087, + "num_tokens": 642452117.0, + "step": 24828 + }, + { + "epoch": 2.726663738194597, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4908313751220703, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7165732383728027, + "num_tokens": 642478164.0, + "step": 24829 + }, + { + "epoch": 2.7267735558972106, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2122881412506104, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7063307762145996, + "num_tokens": 642506938.0, + "step": 24830 + }, + { + "epoch": 2.7268833735998244, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.095792531967163, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.711033284664154, + "num_tokens": 642542506.0, + "step": 24831 + }, + { + "epoch": 2.726993191302438, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5592474937438965, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7361515760421753, + "num_tokens": 642563542.0, + "step": 24832 + }, + { + "epoch": 2.7271030090050514, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0695948600769043, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.725050687789917, + "num_tokens": 642595227.0, + "step": 24833 + }, + { + "epoch": 2.727212826707665, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.923621892929077, + "learning_rate": 1e-06, + "loss": 0.777, + "mean_token_accuracy": 0.7479203939437866, + "num_tokens": 642613641.0, + "step": 24834 + }, + { + "epoch": 2.727322644410279, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.238767147064209, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.723617434501648, + "num_tokens": 642640041.0, + "step": 24835 + }, + { + "epoch": 2.7274324621128927, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.7451088428497314, + "learning_rate": 1e-06, + "loss": 0.7552, + "mean_token_accuracy": 0.7626888155937195, + "num_tokens": 642658862.0, + "step": 24836 + }, + { + "epoch": 2.7275422798155065, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.486168384552002, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7369766235351562, + "num_tokens": 642683284.0, + "step": 24837 + }, + { + "epoch": 2.7276520975181198, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1120667457580566, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7427483797073364, + "num_tokens": 642711859.0, + "step": 24838 + }, + { + "epoch": 2.7277619152207335, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 1.994691014289856, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7160482406616211, + "num_tokens": 642749230.0, + "step": 24839 + }, + { + "epoch": 2.7278717329233473, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3701624870300293, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.716026246547699, + "num_tokens": 642775552.0, + "step": 24840 + }, + { + "epoch": 2.727981550625961, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5242977142333984, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7534857392311096, + "num_tokens": 642797175.0, + "step": 24841 + }, + { + "epoch": 2.728091368328575, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3750648498535156, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7273716330528259, + "num_tokens": 642823312.0, + "step": 24842 + }, + { + "epoch": 2.728201186031188, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.269850015640259, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7237746715545654, + "num_tokens": 642851144.0, + "step": 24843 + }, + { + "epoch": 2.728311003733802, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.352339029312134, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7367390990257263, + "num_tokens": 642878449.0, + "step": 24844 + }, + { + "epoch": 2.7284208214364156, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5087666511535645, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.7530531883239746, + "num_tokens": 642900043.0, + "step": 24845 + }, + { + "epoch": 2.728530639139029, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.06669020652771, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7059640884399414, + "num_tokens": 642933975.0, + "step": 24846 + }, + { + "epoch": 2.728640456841643, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.562962532043457, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7366272211074829, + "num_tokens": 642957037.0, + "step": 24847 + }, + { + "epoch": 2.7287502745442564, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.285569667816162, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7371309995651245, + "num_tokens": 642983748.0, + "step": 24848 + }, + { + "epoch": 2.72886009224687, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0386035442352295, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7315021753311157, + "num_tokens": 643015345.0, + "step": 24849 + }, + { + "epoch": 2.728969909949484, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.67527437210083, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7282359004020691, + "num_tokens": 643037919.0, + "step": 24850 + }, + { + "epoch": 2.7290797276520973, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1565752029418945, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7419819831848145, + "num_tokens": 643066006.0, + "step": 24851 + }, + { + "epoch": 2.729189545354711, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2224035263061523, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7462611198425293, + "num_tokens": 643092663.0, + "step": 24852 + }, + { + "epoch": 2.729299363057325, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2301487922668457, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7276060581207275, + "num_tokens": 643122114.0, + "step": 24853 + }, + { + "epoch": 2.7294091807599385, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.518303155899048, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.742093563079834, + "num_tokens": 643143373.0, + "step": 24854 + }, + { + "epoch": 2.7295189984625523, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3684678077697754, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7486518621444702, + "num_tokens": 643166868.0, + "step": 24855 + }, + { + "epoch": 2.7296288161651656, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4054393768310547, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7205058336257935, + "num_tokens": 643195898.0, + "step": 24856 + }, + { + "epoch": 2.7297386338677794, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1116487979888916, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.705231785774231, + "num_tokens": 643228559.0, + "step": 24857 + }, + { + "epoch": 2.729848451570393, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2823023796081543, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7479594349861145, + "num_tokens": 643254623.0, + "step": 24858 + }, + { + "epoch": 2.729958269273007, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3354787826538086, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7435681819915771, + "num_tokens": 643280531.0, + "step": 24859 + }, + { + "epoch": 2.7300680869756206, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6987879276275635, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7392259240150452, + "num_tokens": 643300152.0, + "step": 24860 + }, + { + "epoch": 2.730177904678234, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.232696771621704, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7303593158721924, + "num_tokens": 643327754.0, + "step": 24861 + }, + { + "epoch": 2.7302877223808477, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5286965370178223, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7380576133728027, + "num_tokens": 643352108.0, + "step": 24862 + }, + { + "epoch": 2.7303975400834615, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.214113473892212, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7276122570037842, + "num_tokens": 643380193.0, + "step": 24863 + }, + { + "epoch": 2.7305073577860752, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.216919183731079, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7128845453262329, + "num_tokens": 643409522.0, + "step": 24864 + }, + { + "epoch": 2.730617175488689, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.237264394760132, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7338166832923889, + "num_tokens": 643437342.0, + "step": 24865 + }, + { + "epoch": 2.7307269931913023, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3477160930633545, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7267911434173584, + "num_tokens": 643463951.0, + "step": 24866 + }, + { + "epoch": 2.730836810893916, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.42954158782959, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7525661587715149, + "num_tokens": 643487187.0, + "step": 24867 + }, + { + "epoch": 2.73094662859653, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6012282371520996, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7236299514770508, + "num_tokens": 643508957.0, + "step": 24868 + }, + { + "epoch": 2.7310564462991436, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.345348358154297, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7286176681518555, + "num_tokens": 643534542.0, + "step": 24869 + }, + { + "epoch": 2.7311662640017573, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4716954231262207, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7321901321411133, + "num_tokens": 643560555.0, + "step": 24870 + }, + { + "epoch": 2.7312760817043706, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.0145721435546875, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7238758206367493, + "num_tokens": 643594435.0, + "step": 24871 + }, + { + "epoch": 2.7313858994069844, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1862285137176514, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7097244262695312, + "num_tokens": 643623464.0, + "step": 24872 + }, + { + "epoch": 2.731495717109598, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.306671619415283, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7260217666625977, + "num_tokens": 643651590.0, + "step": 24873 + }, + { + "epoch": 2.7316055348122115, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.527512311935425, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7323693633079529, + "num_tokens": 643679139.0, + "step": 24874 + }, + { + "epoch": 2.7317153525148252, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4814248085021973, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7204294800758362, + "num_tokens": 643704449.0, + "step": 24875 + }, + { + "epoch": 2.731825170217439, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.566352605819702, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7391749620437622, + "num_tokens": 643727376.0, + "step": 24876 + }, + { + "epoch": 2.7319349879200527, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.450169086456299, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7245964407920837, + "num_tokens": 643751934.0, + "step": 24877 + }, + { + "epoch": 2.7320448056226665, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.328228235244751, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7294370532035828, + "num_tokens": 643778941.0, + "step": 24878 + }, + { + "epoch": 2.73215462332528, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3428595066070557, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7319658994674683, + "num_tokens": 643805336.0, + "step": 24879 + }, + { + "epoch": 2.7322644410278936, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.16313099861145, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7007821798324585, + "num_tokens": 643834946.0, + "step": 24880 + }, + { + "epoch": 2.7323742587305073, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.161780595779419, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7216751575469971, + "num_tokens": 643864791.0, + "step": 24881 + }, + { + "epoch": 2.732484076433121, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.196220636367798, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.719470739364624, + "num_tokens": 643896644.0, + "step": 24882 + }, + { + "epoch": 2.732593894135735, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.350635290145874, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7391195297241211, + "num_tokens": 643924896.0, + "step": 24883 + }, + { + "epoch": 2.732703711838348, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1745493412017822, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7420270442962646, + "num_tokens": 643955376.0, + "step": 24884 + }, + { + "epoch": 2.732813529540962, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1352710723876953, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7443969249725342, + "num_tokens": 643984504.0, + "step": 24885 + }, + { + "epoch": 2.7329233472435757, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4285168647766113, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7270227670669556, + "num_tokens": 644010345.0, + "step": 24886 + }, + { + "epoch": 2.7330331649461894, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4755797386169434, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.733500599861145, + "num_tokens": 644034153.0, + "step": 24887 + }, + { + "epoch": 2.733142982648803, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3332033157348633, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7336658239364624, + "num_tokens": 644059787.0, + "step": 24888 + }, + { + "epoch": 2.7332528003514165, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4853031635284424, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7460002899169922, + "num_tokens": 644082650.0, + "step": 24889 + }, + { + "epoch": 2.7333626180540302, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.417396068572998, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7485922574996948, + "num_tokens": 644106614.0, + "step": 24890 + }, + { + "epoch": 2.733472435756644, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5030953884124756, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.722150444984436, + "num_tokens": 644130761.0, + "step": 24891 + }, + { + "epoch": 2.7335822534592578, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4884254932403564, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7467954754829407, + "num_tokens": 644154838.0, + "step": 24892 + }, + { + "epoch": 2.7336920711618715, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.481668710708618, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7416459918022156, + "num_tokens": 644179150.0, + "step": 24893 + }, + { + "epoch": 2.733801888864485, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2179555892944336, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7390555143356323, + "num_tokens": 644206990.0, + "step": 24894 + }, + { + "epoch": 2.7339117065670986, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4867756366729736, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.711235523223877, + "num_tokens": 644233162.0, + "step": 24895 + }, + { + "epoch": 2.7340215242697123, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4947943687438965, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7080729007720947, + "num_tokens": 644256968.0, + "step": 24896 + }, + { + "epoch": 2.734131341972326, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2696595191955566, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.72437983751297, + "num_tokens": 644284581.0, + "step": 24897 + }, + { + "epoch": 2.73424115967494, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5965511798858643, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7324726581573486, + "num_tokens": 644308226.0, + "step": 24898 + }, + { + "epoch": 2.734350977377553, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.15628719329834, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.713804304599762, + "num_tokens": 644340434.0, + "step": 24899 + }, + { + "epoch": 2.734460795080167, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2486679553985596, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7336998581886292, + "num_tokens": 644371815.0, + "step": 24900 + }, + { + "epoch": 2.7345706127827807, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3406786918640137, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.728106677532196, + "num_tokens": 644396875.0, + "step": 24901 + }, + { + "epoch": 2.734680430485394, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.290262460708618, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7236621975898743, + "num_tokens": 644423665.0, + "step": 24902 + }, + { + "epoch": 2.7347902481880078, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.149986505508423, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7147259712219238, + "num_tokens": 644455080.0, + "step": 24903 + }, + { + "epoch": 2.7349000658906215, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4821596145629883, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7457518577575684, + "num_tokens": 644478301.0, + "step": 24904 + }, + { + "epoch": 2.7350098835932353, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.607729434967041, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7381026744842529, + "num_tokens": 644500203.0, + "step": 24905 + }, + { + "epoch": 2.735119701295849, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.272033214569092, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7206599116325378, + "num_tokens": 644528204.0, + "step": 24906 + }, + { + "epoch": 2.7352295189984623, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.8967134952545166, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7496117353439331, + "num_tokens": 644547240.0, + "step": 24907 + }, + { + "epoch": 2.735339336701076, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.4920010566711426, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7397753000259399, + "num_tokens": 644570855.0, + "step": 24908 + }, + { + "epoch": 2.73544915440369, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.470026731491089, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7096115946769714, + "num_tokens": 644595127.0, + "step": 24909 + }, + { + "epoch": 2.7355589721063036, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.1806180477142334, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7336821556091309, + "num_tokens": 644625702.0, + "step": 24910 + }, + { + "epoch": 2.7356687898089174, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6088716983795166, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7173479199409485, + "num_tokens": 644649169.0, + "step": 24911 + }, + { + "epoch": 2.7357786075115307, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6937811374664307, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.746268630027771, + "num_tokens": 644670190.0, + "step": 24912 + }, + { + "epoch": 2.7358884252141444, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.3323915004730225, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7287588715553284, + "num_tokens": 644694420.0, + "step": 24913 + }, + { + "epoch": 2.735998242916758, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.5151257514953613, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7219350934028625, + "num_tokens": 644717165.0, + "step": 24914 + }, + { + "epoch": 2.736108060619372, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2521865367889404, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7420182824134827, + "num_tokens": 644745186.0, + "step": 24915 + }, + { + "epoch": 2.7362178783219857, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2873623371124268, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.718885600566864, + "num_tokens": 644773772.0, + "step": 24916 + }, + { + "epoch": 2.736327696024599, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.421505928039551, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7174170613288879, + "num_tokens": 644800001.0, + "step": 24917 + }, + { + "epoch": 2.736437513727213, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.405665636062622, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7221584320068359, + "num_tokens": 644825060.0, + "step": 24918 + }, + { + "epoch": 2.7365473314298265, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.514842987060547, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7339601516723633, + "num_tokens": 644848268.0, + "step": 24919 + }, + { + "epoch": 2.7366571491324403, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.696455717086792, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7393059730529785, + "num_tokens": 644868479.0, + "step": 24920 + }, + { + "epoch": 2.736766966835054, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 1.8746535778045654, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.727331280708313, + "num_tokens": 644904886.0, + "step": 24921 + }, + { + "epoch": 2.7368767845376674, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 1.9824943542480469, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7118152379989624, + "num_tokens": 644942507.0, + "step": 24922 + }, + { + "epoch": 2.736986602240281, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.6976258754730225, + "learning_rate": 1e-06, + "loss": 0.7568, + "mean_token_accuracy": 0.7604634165763855, + "num_tokens": 644960996.0, + "step": 24923 + }, + { + "epoch": 2.737096419942895, + "ewc_loss": 0.00021839141845703125, + "grad_norm": 2.2205212116241455, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7322176098823547, + "num_tokens": 644991209.0, + "step": 24924 + }, + { + "epoch": 2.737206237645508, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.257580280303955, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7117912769317627, + "num_tokens": 645021253.0, + "step": 24925 + }, + { + "epoch": 2.7373160553481224, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4473440647125244, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7478539943695068, + "num_tokens": 645045743.0, + "step": 24926 + }, + { + "epoch": 2.7374258730507357, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4202897548675537, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7283190488815308, + "num_tokens": 645070358.0, + "step": 24927 + }, + { + "epoch": 2.7375356907533495, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.433757781982422, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7489660978317261, + "num_tokens": 645093569.0, + "step": 24928 + }, + { + "epoch": 2.737645508455963, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1940383911132812, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7182794809341431, + "num_tokens": 645123189.0, + "step": 24929 + }, + { + "epoch": 2.7377553261585765, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.310786247253418, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7420948147773743, + "num_tokens": 645148340.0, + "step": 24930 + }, + { + "epoch": 2.7378651438611903, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.653449535369873, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7506327629089355, + "num_tokens": 645170493.0, + "step": 24931 + }, + { + "epoch": 2.737974961563804, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.418245792388916, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7200641632080078, + "num_tokens": 645196556.0, + "step": 24932 + }, + { + "epoch": 2.738084779266418, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.474815845489502, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7261340022087097, + "num_tokens": 645222041.0, + "step": 24933 + }, + { + "epoch": 2.7381945969690316, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5399696826934814, + "learning_rate": 1e-06, + "loss": 0.7739, + "mean_token_accuracy": 0.7483805418014526, + "num_tokens": 645243118.0, + "step": 24934 + }, + { + "epoch": 2.738304414671645, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 3.0093436241149902, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7349415421485901, + "num_tokens": 645261831.0, + "step": 24935 + }, + { + "epoch": 2.7384142323742586, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.3540968894958496, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7136422395706177, + "num_tokens": 645289555.0, + "step": 24936 + }, + { + "epoch": 2.7385240500768724, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.618433713912964, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7420313358306885, + "num_tokens": 645311314.0, + "step": 24937 + }, + { + "epoch": 2.738633867779486, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.293801784515381, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7121227979660034, + "num_tokens": 645341909.0, + "step": 24938 + }, + { + "epoch": 2.7387436854821, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5811703205108643, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7148460745811462, + "num_tokens": 645366844.0, + "step": 24939 + }, + { + "epoch": 2.738853503184713, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.628340005874634, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7250852584838867, + "num_tokens": 645389199.0, + "step": 24940 + }, + { + "epoch": 2.738963320887327, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.111111640930176, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7230280637741089, + "num_tokens": 645423038.0, + "step": 24941 + }, + { + "epoch": 2.7390731385899407, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2302439212799072, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7443864345550537, + "num_tokens": 645451731.0, + "step": 24942 + }, + { + "epoch": 2.7391829562925545, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2480108737945557, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.729549765586853, + "num_tokens": 645478334.0, + "step": 24943 + }, + { + "epoch": 2.7392927739951682, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.446488380432129, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7272990942001343, + "num_tokens": 645502011.0, + "step": 24944 + }, + { + "epoch": 2.7394025916977816, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4281437397003174, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7294277548789978, + "num_tokens": 645526537.0, + "step": 24945 + }, + { + "epoch": 2.7395124094003953, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.144904136657715, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.76274573802948, + "num_tokens": 645555576.0, + "step": 24946 + }, + { + "epoch": 2.739622227103009, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.589442253112793, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7352492213249207, + "num_tokens": 645579414.0, + "step": 24947 + }, + { + "epoch": 2.739732044805623, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.7038373947143555, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7427046298980713, + "num_tokens": 645600487.0, + "step": 24948 + }, + { + "epoch": 2.7398418625082366, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2634620666503906, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7224365472793579, + "num_tokens": 645627786.0, + "step": 24949 + }, + { + "epoch": 2.73995168021085, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.627040147781372, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7409430742263794, + "num_tokens": 645650574.0, + "step": 24950 + }, + { + "epoch": 2.7400614979134637, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.0714428424835205, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7039316892623901, + "num_tokens": 645687087.0, + "step": 24951 + }, + { + "epoch": 2.7401713156160774, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.396383762359619, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7360759973526001, + "num_tokens": 645710254.0, + "step": 24952 + }, + { + "epoch": 2.7402811333186907, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.179286241531372, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7195816040039062, + "num_tokens": 645740072.0, + "step": 24953 + }, + { + "epoch": 2.7403909510213045, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.987560749053955, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.741212785243988, + "num_tokens": 645758717.0, + "step": 24954 + }, + { + "epoch": 2.7405007687239182, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.327603340148926, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7180545330047607, + "num_tokens": 645783376.0, + "step": 24955 + }, + { + "epoch": 2.740610586426532, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.397406816482544, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7174824476242065, + "num_tokens": 645810540.0, + "step": 24956 + }, + { + "epoch": 2.7407204041291457, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.454011917114258, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.729031503200531, + "num_tokens": 645836400.0, + "step": 24957 + }, + { + "epoch": 2.740830221831759, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2017295360565186, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7122336626052856, + "num_tokens": 645865217.0, + "step": 24958 + }, + { + "epoch": 2.740940039534373, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.0826942920684814, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7146555185317993, + "num_tokens": 645897124.0, + "step": 24959 + }, + { + "epoch": 2.7410498572369866, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5659146308898926, + "learning_rate": 1e-06, + "loss": 0.7614, + "mean_token_accuracy": 0.7564207315444946, + "num_tokens": 645919008.0, + "step": 24960 + }, + { + "epoch": 2.7411596749396003, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.393573045730591, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7269222736358643, + "num_tokens": 645944119.0, + "step": 24961 + }, + { + "epoch": 2.741269492642214, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1733505725860596, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7105100154876709, + "num_tokens": 645976846.0, + "step": 24962 + }, + { + "epoch": 2.7413793103448274, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4797282218933105, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7383548021316528, + "num_tokens": 646000486.0, + "step": 24963 + }, + { + "epoch": 2.741489128047441, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.33296275138855, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.709392249584198, + "num_tokens": 646028394.0, + "step": 24964 + }, + { + "epoch": 2.741598945750055, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1128013134002686, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7158476114273071, + "num_tokens": 646058019.0, + "step": 24965 + }, + { + "epoch": 2.7417087634526687, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4707469940185547, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7186869382858276, + "num_tokens": 646082838.0, + "step": 24966 + }, + { + "epoch": 2.7418185811552824, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2016448974609375, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7416211366653442, + "num_tokens": 646112665.0, + "step": 24967 + }, + { + "epoch": 2.7419283988578957, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1159796714782715, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7171022891998291, + "num_tokens": 646143820.0, + "step": 24968 + }, + { + "epoch": 2.7420382165605095, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.323972225189209, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7185673713684082, + "num_tokens": 646170636.0, + "step": 24969 + }, + { + "epoch": 2.7421480342631233, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.6042287349700928, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7206594347953796, + "num_tokens": 646192243.0, + "step": 24970 + }, + { + "epoch": 2.742257851965737, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.158909797668457, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7081631422042847, + "num_tokens": 646224106.0, + "step": 24971 + }, + { + "epoch": 2.7423676696683508, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.263132095336914, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7167468070983887, + "num_tokens": 646251003.0, + "step": 24972 + }, + { + "epoch": 2.742477487370964, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.311124563217163, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7098729610443115, + "num_tokens": 646278161.0, + "step": 24973 + }, + { + "epoch": 2.742587305073578, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.539158582687378, + "learning_rate": 1e-06, + "loss": 0.7501, + "mean_token_accuracy": 0.7576694488525391, + "num_tokens": 646300023.0, + "step": 24974 + }, + { + "epoch": 2.7426971227761916, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1302103996276855, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7167712450027466, + "num_tokens": 646330269.0, + "step": 24975 + }, + { + "epoch": 2.742806940478805, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.237517833709717, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7163856029510498, + "num_tokens": 646359177.0, + "step": 24976 + }, + { + "epoch": 2.742916758181419, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.7715795040130615, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7361730933189392, + "num_tokens": 646378780.0, + "step": 24977 + }, + { + "epoch": 2.7430265758840324, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.215272903442383, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7366831302642822, + "num_tokens": 646408380.0, + "step": 24978 + }, + { + "epoch": 2.743136393586646, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.246720790863037, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7258267998695374, + "num_tokens": 646436526.0, + "step": 24979 + }, + { + "epoch": 2.74324621128926, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.686588764190674, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.74619460105896, + "num_tokens": 646457486.0, + "step": 24980 + }, + { + "epoch": 2.7433560289918733, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.6152143478393555, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7352071404457092, + "num_tokens": 646478215.0, + "step": 24981 + }, + { + "epoch": 2.743465846694487, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.3406155109405518, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7307928800582886, + "num_tokens": 646502529.0, + "step": 24982 + }, + { + "epoch": 2.7435756643971008, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2080109119415283, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7199507355690002, + "num_tokens": 646532780.0, + "step": 24983 + }, + { + "epoch": 2.7436854820997145, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.553424835205078, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7438379526138306, + "num_tokens": 646555900.0, + "step": 24984 + }, + { + "epoch": 2.7437952998023283, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 1.9691914319992065, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7037291526794434, + "num_tokens": 646590148.0, + "step": 24985 + }, + { + "epoch": 2.7439051175049416, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4529001712799072, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7469996213912964, + "num_tokens": 646615531.0, + "step": 24986 + }, + { + "epoch": 2.7440149352075554, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.3492374420166016, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7420080900192261, + "num_tokens": 646640800.0, + "step": 24987 + }, + { + "epoch": 2.744124752910169, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.6061575412750244, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7307178974151611, + "num_tokens": 646664533.0, + "step": 24988 + }, + { + "epoch": 2.744234570612783, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5338189601898193, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7464816570281982, + "num_tokens": 646687570.0, + "step": 24989 + }, + { + "epoch": 2.7443443883153966, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.3408470153808594, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7155846953392029, + "num_tokens": 646713050.0, + "step": 24990 + }, + { + "epoch": 2.74445420601801, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5210065841674805, + "learning_rate": 1e-06, + "loss": 0.8026, + "mean_token_accuracy": 0.7449765205383301, + "num_tokens": 646736710.0, + "step": 24991 + }, + { + "epoch": 2.7445640237206237, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.506892442703247, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7364005446434021, + "num_tokens": 646760872.0, + "step": 24992 + }, + { + "epoch": 2.7446738414232374, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4652493000030518, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7503006458282471, + "num_tokens": 646783717.0, + "step": 24993 + }, + { + "epoch": 2.744783659125851, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.30781626701355, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7209664583206177, + "num_tokens": 646810412.0, + "step": 24994 + }, + { + "epoch": 2.744893476828465, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.642287254333496, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7242977619171143, + "num_tokens": 646833771.0, + "step": 24995 + }, + { + "epoch": 2.7450032945310783, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4085886478424072, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.735533595085144, + "num_tokens": 646858231.0, + "step": 24996 + }, + { + "epoch": 2.745113112233692, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.751657724380493, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7305161952972412, + "num_tokens": 646879355.0, + "step": 24997 + }, + { + "epoch": 2.745222929936306, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2172420024871826, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6983787417411804, + "num_tokens": 646910891.0, + "step": 24998 + }, + { + "epoch": 2.7453327476389195, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2838361263275146, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7402235269546509, + "num_tokens": 646938572.0, + "step": 24999 + }, + { + "epoch": 2.7454425653415333, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2924156188964844, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7417481541633606, + "num_tokens": 646964803.0, + "step": 25000 + }, + { + "epoch": 2.7455523830441466, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.069087505340576, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7253843545913696, + "num_tokens": 646997019.0, + "step": 25001 + }, + { + "epoch": 2.7456622007467604, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5157630443573, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7455252408981323, + "num_tokens": 647020212.0, + "step": 25002 + }, + { + "epoch": 2.745772018449374, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1892409324645996, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7162784934043884, + "num_tokens": 647049845.0, + "step": 25003 + }, + { + "epoch": 2.7458818361519874, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2907660007476807, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.712794303894043, + "num_tokens": 647078205.0, + "step": 25004 + }, + { + "epoch": 2.745991653854601, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 1.9708491563796997, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7155247330665588, + "num_tokens": 647113634.0, + "step": 25005 + }, + { + "epoch": 2.746101471557215, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.6514487266540527, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7293730974197388, + "num_tokens": 647136875.0, + "step": 25006 + }, + { + "epoch": 2.7462112892598287, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.367666482925415, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7337266802787781, + "num_tokens": 647162932.0, + "step": 25007 + }, + { + "epoch": 2.7463211069624425, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.2569098472595215, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7139772772789001, + "num_tokens": 647190846.0, + "step": 25008 + }, + { + "epoch": 2.746430924665056, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.1999928951263428, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.723518431186676, + "num_tokens": 647219098.0, + "step": 25009 + }, + { + "epoch": 2.7465407423676695, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.437485456466675, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7255637049674988, + "num_tokens": 647243920.0, + "step": 25010 + }, + { + "epoch": 2.7466505600702833, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.5074286460876465, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7297457456588745, + "num_tokens": 647268324.0, + "step": 25011 + }, + { + "epoch": 2.746760377772897, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4890594482421875, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7339097857475281, + "num_tokens": 647291575.0, + "step": 25012 + }, + { + "epoch": 2.746870195475511, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.4151086807250977, + "learning_rate": 1e-06, + "loss": 0.769, + "mean_token_accuracy": 0.7547144889831543, + "num_tokens": 647314862.0, + "step": 25013 + }, + { + "epoch": 2.746980013178124, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.3536064624786377, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7252015471458435, + "num_tokens": 647338860.0, + "step": 25014 + }, + { + "epoch": 2.747089830880738, + "ewc_loss": 0.0002193450927734375, + "grad_norm": 2.173525810241699, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7268363237380981, + "num_tokens": 647367882.0, + "step": 25015 + }, + { + "epoch": 2.7471996485833516, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.314195394515991, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7111378908157349, + "num_tokens": 647394841.0, + "step": 25016 + }, + { + "epoch": 2.7473094662859654, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.334864616394043, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7313876748085022, + "num_tokens": 647419836.0, + "step": 25017 + }, + { + "epoch": 2.747419283988579, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4929394721984863, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7359153032302856, + "num_tokens": 647445445.0, + "step": 25018 + }, + { + "epoch": 2.7475291016911925, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.143998384475708, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7371196746826172, + "num_tokens": 647476707.0, + "step": 25019 + }, + { + "epoch": 2.7476389193938062, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 3.2363462448120117, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.742800772190094, + "num_tokens": 647493810.0, + "step": 25020 + }, + { + "epoch": 2.74774873709642, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2470381259918213, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7441890239715576, + "num_tokens": 647521427.0, + "step": 25021 + }, + { + "epoch": 2.7478585547990337, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2575767040252686, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7275429964065552, + "num_tokens": 647549213.0, + "step": 25022 + }, + { + "epoch": 2.7479683725016475, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.1001205444335938, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6905292272567749, + "num_tokens": 647581008.0, + "step": 25023 + }, + { + "epoch": 2.748078190204261, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3512117862701416, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7389536499977112, + "num_tokens": 647605790.0, + "step": 25024 + }, + { + "epoch": 2.7481880079068746, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.284824848175049, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7325581312179565, + "num_tokens": 647631816.0, + "step": 25025 + }, + { + "epoch": 2.7482978256094883, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.5287623405456543, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7441130876541138, + "num_tokens": 647656111.0, + "step": 25026 + }, + { + "epoch": 2.7484076433121016, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.6640541553497314, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7333043813705444, + "num_tokens": 647677258.0, + "step": 25027 + }, + { + "epoch": 2.748517461014716, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.8459789752960205, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7339305281639099, + "num_tokens": 647697016.0, + "step": 25028 + }, + { + "epoch": 2.748627278717329, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.371636152267456, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7245927453041077, + "num_tokens": 647723095.0, + "step": 25029 + }, + { + "epoch": 2.748737096419943, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4260761737823486, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7199058532714844, + "num_tokens": 647747672.0, + "step": 25030 + }, + { + "epoch": 2.7488469141225567, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.400432825088501, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7287675142288208, + "num_tokens": 647774007.0, + "step": 25031 + }, + { + "epoch": 2.74895673182517, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.268098831176758, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.721428632736206, + "num_tokens": 647804272.0, + "step": 25032 + }, + { + "epoch": 2.7490665495277837, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.5575010776519775, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7409561276435852, + "num_tokens": 647827428.0, + "step": 25033 + }, + { + "epoch": 2.7491763672303975, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.625818967819214, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7327229976654053, + "num_tokens": 647849150.0, + "step": 25034 + }, + { + "epoch": 2.7492861849330112, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.279118537902832, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7456738948822021, + "num_tokens": 647876113.0, + "step": 25035 + }, + { + "epoch": 2.749396002635625, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2835421562194824, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7173513174057007, + "num_tokens": 647903846.0, + "step": 25036 + }, + { + "epoch": 2.7495058203382383, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.006852388381958, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7053475379943848, + "num_tokens": 647938368.0, + "step": 25037 + }, + { + "epoch": 2.749615638040852, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.0740973949432373, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7195562124252319, + "num_tokens": 647970228.0, + "step": 25038 + }, + { + "epoch": 2.749725455743466, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.160299777984619, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.720446765422821, + "num_tokens": 647999975.0, + "step": 25039 + }, + { + "epoch": 2.7498352734460796, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.434267520904541, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7323555946350098, + "num_tokens": 648025289.0, + "step": 25040 + }, + { + "epoch": 2.7499450911486933, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.192323684692383, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7323712110519409, + "num_tokens": 648056141.0, + "step": 25041 + }, + { + "epoch": 2.7500549088513067, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.1363580226898193, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7318142056465149, + "num_tokens": 648087984.0, + "step": 25042 + }, + { + "epoch": 2.7501647265539204, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.1986265182495117, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.7426508665084839, + "num_tokens": 648118362.0, + "step": 25043 + }, + { + "epoch": 2.750274544256534, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3357486724853516, + "learning_rate": 1e-06, + "loss": 0.7708, + "mean_token_accuracy": 0.7559453845024109, + "num_tokens": 648142816.0, + "step": 25044 + }, + { + "epoch": 2.750384361959148, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.238762140274048, + "learning_rate": 1e-06, + "loss": 0.7783, + "mean_token_accuracy": 0.7520829439163208, + "num_tokens": 648170513.0, + "step": 25045 + }, + { + "epoch": 2.7504941796617617, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3551268577575684, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7370617389678955, + "num_tokens": 648195166.0, + "step": 25046 + }, + { + "epoch": 2.750603997364375, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3103296756744385, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7263232469558716, + "num_tokens": 648222767.0, + "step": 25047 + }, + { + "epoch": 2.7507138150669888, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.0119519233703613, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7259741425514221, + "num_tokens": 648255722.0, + "step": 25048 + }, + { + "epoch": 2.7508236327696025, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4339256286621094, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.745320737361908, + "num_tokens": 648280043.0, + "step": 25049 + }, + { + "epoch": 2.7509334504722163, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3019678592681885, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7042919397354126, + "num_tokens": 648307814.0, + "step": 25050 + }, + { + "epoch": 2.75104326817483, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2867610454559326, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7167292833328247, + "num_tokens": 648336217.0, + "step": 25051 + }, + { + "epoch": 2.7511530858774433, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.246109962463379, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7463949918746948, + "num_tokens": 648362643.0, + "step": 25052 + }, + { + "epoch": 2.751262903580057, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2984731197357178, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7170932292938232, + "num_tokens": 648389974.0, + "step": 25053 + }, + { + "epoch": 2.751372721282671, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2700154781341553, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.708672046661377, + "num_tokens": 648418638.0, + "step": 25054 + }, + { + "epoch": 2.751482538985284, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.546696424484253, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.733873724937439, + "num_tokens": 648440188.0, + "step": 25055 + }, + { + "epoch": 2.7515923566878984, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4313249588012695, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7513431310653687, + "num_tokens": 648463041.0, + "step": 25056 + }, + { + "epoch": 2.7517021743905117, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3158440589904785, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7431814670562744, + "num_tokens": 648489706.0, + "step": 25057 + }, + { + "epoch": 2.7518119920931254, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4522547721862793, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7248474359512329, + "num_tokens": 648513222.0, + "step": 25058 + }, + { + "epoch": 2.751921809795739, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.5631227493286133, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.743013858795166, + "num_tokens": 648538625.0, + "step": 25059 + }, + { + "epoch": 2.7520316274983525, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4871256351470947, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.734010636806488, + "num_tokens": 648562289.0, + "step": 25060 + }, + { + "epoch": 2.7521414452009663, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.0700342655181885, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7145907878875732, + "num_tokens": 648595217.0, + "step": 25061 + }, + { + "epoch": 2.75225126290358, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3841381072998047, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7285643219947815, + "num_tokens": 648620006.0, + "step": 25062 + }, + { + "epoch": 2.7523610806061938, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.1950340270996094, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6930670738220215, + "num_tokens": 648648475.0, + "step": 25063 + }, + { + "epoch": 2.7524708983088075, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4164576530456543, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7315718531608582, + "num_tokens": 648672852.0, + "step": 25064 + }, + { + "epoch": 2.752580716011421, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3981118202209473, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7559570670127869, + "num_tokens": 648698109.0, + "step": 25065 + }, + { + "epoch": 2.7526905337140346, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.267117977142334, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7242288589477539, + "num_tokens": 648725701.0, + "step": 25066 + }, + { + "epoch": 2.7528003514166484, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4451115131378174, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7567055821418762, + "num_tokens": 648749283.0, + "step": 25067 + }, + { + "epoch": 2.752910169119262, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4335997104644775, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.7596257925033569, + "num_tokens": 648771991.0, + "step": 25068 + }, + { + "epoch": 2.753019986821876, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4515485763549805, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7422856092453003, + "num_tokens": 648795677.0, + "step": 25069 + }, + { + "epoch": 2.753129804524489, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.508221387863159, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7323477268218994, + "num_tokens": 648818931.0, + "step": 25070 + }, + { + "epoch": 2.753239622227103, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3520865440368652, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7149276733398438, + "num_tokens": 648847331.0, + "step": 25071 + }, + { + "epoch": 2.7533494399297167, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.664306640625, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7229777574539185, + "num_tokens": 648868817.0, + "step": 25072 + }, + { + "epoch": 2.7534592576323305, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.382613182067871, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7222360372543335, + "num_tokens": 648895614.0, + "step": 25073 + }, + { + "epoch": 2.753569075334944, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3398447036743164, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7347534894943237, + "num_tokens": 648921781.0, + "step": 25074 + }, + { + "epoch": 2.7536788930375575, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.203026294708252, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7118711471557617, + "num_tokens": 648952615.0, + "step": 25075 + }, + { + "epoch": 2.7537887107401713, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.3832035064697266, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7412364482879639, + "num_tokens": 648976681.0, + "step": 25076 + }, + { + "epoch": 2.753898528442785, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.493056297302246, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7380052804946899, + "num_tokens": 649001451.0, + "step": 25077 + }, + { + "epoch": 2.754008346145399, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2445497512817383, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7090782523155212, + "num_tokens": 649030563.0, + "step": 25078 + }, + { + "epoch": 2.7541181638480126, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.6185801029205322, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7284657955169678, + "num_tokens": 649052981.0, + "step": 25079 + }, + { + "epoch": 2.754227981550626, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.258434295654297, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7306313514709473, + "num_tokens": 649081091.0, + "step": 25080 + }, + { + "epoch": 2.7543377992532396, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4068026542663574, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7327041625976562, + "num_tokens": 649105045.0, + "step": 25081 + }, + { + "epoch": 2.7544476169558534, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.2294318675994873, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7379270195960999, + "num_tokens": 649133027.0, + "step": 25082 + }, + { + "epoch": 2.7545574346584667, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.7674427032470703, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7440198063850403, + "num_tokens": 649151761.0, + "step": 25083 + }, + { + "epoch": 2.7546672523610805, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.0835797786712646, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7258439660072327, + "num_tokens": 649183703.0, + "step": 25084 + }, + { + "epoch": 2.754777070063694, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.421034812927246, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7229448556900024, + "num_tokens": 649209334.0, + "step": 25085 + }, + { + "epoch": 2.754886887766308, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.291321039199829, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7490772604942322, + "num_tokens": 649235028.0, + "step": 25086 + }, + { + "epoch": 2.7549967054689217, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4718332290649414, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7316828966140747, + "num_tokens": 649258176.0, + "step": 25087 + }, + { + "epoch": 2.755106523171535, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.47318172454834, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7342323660850525, + "num_tokens": 649281539.0, + "step": 25088 + }, + { + "epoch": 2.755216340874149, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.0775742530822754, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.6991400718688965, + "num_tokens": 649312403.0, + "step": 25089 + }, + { + "epoch": 2.7553261585767626, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.4249656200408936, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.711066722869873, + "num_tokens": 649341352.0, + "step": 25090 + }, + { + "epoch": 2.7554359762793763, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.570172071456909, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7370278239250183, + "num_tokens": 649363178.0, + "step": 25091 + }, + { + "epoch": 2.75554579398199, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.866177558898926, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7511075735092163, + "num_tokens": 649381050.0, + "step": 25092 + }, + { + "epoch": 2.7556556116846034, + "ewc_loss": 0.00022029876708984375, + "grad_norm": 2.500985622406006, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.733233630657196, + "num_tokens": 649405317.0, + "step": 25093 + }, + { + "epoch": 2.755765429387217, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4156413078308105, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7368621826171875, + "num_tokens": 649430816.0, + "step": 25094 + }, + { + "epoch": 2.755875247089831, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.376415491104126, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7544196248054504, + "num_tokens": 649456318.0, + "step": 25095 + }, + { + "epoch": 2.7559850647924446, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5275678634643555, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7308069467544556, + "num_tokens": 649480182.0, + "step": 25096 + }, + { + "epoch": 2.7560948824950584, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.551152229309082, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7365927696228027, + "num_tokens": 649504133.0, + "step": 25097 + }, + { + "epoch": 2.7562047001976717, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2024848461151123, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7261976003646851, + "num_tokens": 649533342.0, + "step": 25098 + }, + { + "epoch": 2.7563145179002855, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.113034248352051, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7085580229759216, + "num_tokens": 649567140.0, + "step": 25099 + }, + { + "epoch": 2.7564243356028992, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3325607776641846, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7346125841140747, + "num_tokens": 649591643.0, + "step": 25100 + }, + { + "epoch": 2.756534153305513, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.085671901702881, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7121019959449768, + "num_tokens": 649624483.0, + "step": 25101 + }, + { + "epoch": 2.7566439710081267, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.289602518081665, + "learning_rate": 1e-06, + "loss": 0.7181, + "mean_token_accuracy": 0.7712885737419128, + "num_tokens": 649650361.0, + "step": 25102 + }, + { + "epoch": 2.75675378871074, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6547162532806396, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7194595336914062, + "num_tokens": 649670929.0, + "step": 25103 + }, + { + "epoch": 2.756863606413354, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.004570484161377, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7278441190719604, + "num_tokens": 649704762.0, + "step": 25104 + }, + { + "epoch": 2.7569734241159676, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.551555871963501, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7107235193252563, + "num_tokens": 649728888.0, + "step": 25105 + }, + { + "epoch": 2.757083241818581, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2228775024414062, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6990530490875244, + "num_tokens": 649756537.0, + "step": 25106 + }, + { + "epoch": 2.757193059521195, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.370634078979492, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.729595959186554, + "num_tokens": 649784537.0, + "step": 25107 + }, + { + "epoch": 2.7573028772238084, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.279799699783325, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7062660455703735, + "num_tokens": 649812101.0, + "step": 25108 + }, + { + "epoch": 2.757412694926422, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.0538060665130615, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7408345937728882, + "num_tokens": 649840318.0, + "step": 25109 + }, + { + "epoch": 2.757522512629036, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.402601957321167, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7407531142234802, + "num_tokens": 649865001.0, + "step": 25110 + }, + { + "epoch": 2.7576323303316492, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4524857997894287, + "learning_rate": 1e-06, + "loss": 0.7896, + "mean_token_accuracy": 0.7471065521240234, + "num_tokens": 649889271.0, + "step": 25111 + }, + { + "epoch": 2.757742148034263, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2982723712921143, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7205634713172913, + "num_tokens": 649916816.0, + "step": 25112 + }, + { + "epoch": 2.7578519657368767, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4625747203826904, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7141435742378235, + "num_tokens": 649941801.0, + "step": 25113 + }, + { + "epoch": 2.7579617834394905, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.541592597961426, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7329564094543457, + "num_tokens": 649964154.0, + "step": 25114 + }, + { + "epoch": 2.7580716011421043, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1768693923950195, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7193409204483032, + "num_tokens": 649994431.0, + "step": 25115 + }, + { + "epoch": 2.7581814188447176, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2242519855499268, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7370517253875732, + "num_tokens": 650023704.0, + "step": 25116 + }, + { + "epoch": 2.7582912365473313, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5727999210357666, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7275546789169312, + "num_tokens": 650046884.0, + "step": 25117 + }, + { + "epoch": 2.758401054249945, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3622000217437744, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7479594945907593, + "num_tokens": 650071453.0, + "step": 25118 + }, + { + "epoch": 2.758510871952559, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.301884651184082, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7471102476119995, + "num_tokens": 650096453.0, + "step": 25119 + }, + { + "epoch": 2.7586206896551726, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4789798259735107, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7275246977806091, + "num_tokens": 650120391.0, + "step": 25120 + }, + { + "epoch": 2.758730507357786, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.507505178451538, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7228755950927734, + "num_tokens": 650144385.0, + "step": 25121 + }, + { + "epoch": 2.7588403250603997, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.458087682723999, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7113816142082214, + "num_tokens": 650169422.0, + "step": 25122 + }, + { + "epoch": 2.7589501427630134, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3646552562713623, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7245298624038696, + "num_tokens": 650194483.0, + "step": 25123 + }, + { + "epoch": 2.759059960465627, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4524030685424805, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7219300866127014, + "num_tokens": 650220021.0, + "step": 25124 + }, + { + "epoch": 2.759169778168241, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.272937297821045, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7256866693496704, + "num_tokens": 650246245.0, + "step": 25125 + }, + { + "epoch": 2.7592795958708543, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3821067810058594, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7181289792060852, + "num_tokens": 650272698.0, + "step": 25126 + }, + { + "epoch": 2.759389413573468, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.347256660461426, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7338789105415344, + "num_tokens": 650297928.0, + "step": 25127 + }, + { + "epoch": 2.7594992312760818, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.457834005355835, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7200653553009033, + "num_tokens": 650323221.0, + "step": 25128 + }, + { + "epoch": 2.7596090489786955, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.0652711391448975, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7416564226150513, + "num_tokens": 650354418.0, + "step": 25129 + }, + { + "epoch": 2.7597188666813093, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.47586727142334, + "learning_rate": 1e-06, + "loss": 0.7496, + "mean_token_accuracy": 0.7560688257217407, + "num_tokens": 650376365.0, + "step": 25130 + }, + { + "epoch": 2.7598286843839226, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.435802459716797, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7142834067344666, + "num_tokens": 650400045.0, + "step": 25131 + }, + { + "epoch": 2.7599385020865363, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.9986305236816406, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7369143962860107, + "num_tokens": 650418615.0, + "step": 25132 + }, + { + "epoch": 2.76004831978915, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2360451221466064, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7514408826828003, + "num_tokens": 650443488.0, + "step": 25133 + }, + { + "epoch": 2.7601581374917634, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2261736392974854, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7215200662612915, + "num_tokens": 650473019.0, + "step": 25134 + }, + { + "epoch": 2.760267955194377, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.28118634223938, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7403935194015503, + "num_tokens": 650499245.0, + "step": 25135 + }, + { + "epoch": 2.760377772896991, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4883358478546143, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7316806316375732, + "num_tokens": 650522784.0, + "step": 25136 + }, + { + "epoch": 2.7604875905996047, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5786213874816895, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7359263896942139, + "num_tokens": 650543810.0, + "step": 25137 + }, + { + "epoch": 2.7605974083022184, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5074193477630615, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.741166353225708, + "num_tokens": 650565417.0, + "step": 25138 + }, + { + "epoch": 2.7607072260048318, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.189521074295044, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7322725057601929, + "num_tokens": 650596977.0, + "step": 25139 + }, + { + "epoch": 2.7608170437074455, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.442204475402832, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7210103273391724, + "num_tokens": 650621860.0, + "step": 25140 + }, + { + "epoch": 2.7609268614100593, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.426239013671875, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7271639108657837, + "num_tokens": 650646498.0, + "step": 25141 + }, + { + "epoch": 2.761036679112673, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.85294246673584, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.7515162229537964, + "num_tokens": 650664077.0, + "step": 25142 + }, + { + "epoch": 2.761146496815287, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.67348051071167, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7236055731773376, + "num_tokens": 650686151.0, + "step": 25143 + }, + { + "epoch": 2.7612563145179, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.380565643310547, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7035791873931885, + "num_tokens": 650714203.0, + "step": 25144 + }, + { + "epoch": 2.761366132220514, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.715620279312134, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7501310110092163, + "num_tokens": 650734054.0, + "step": 25145 + }, + { + "epoch": 2.7614759499231276, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4324121475219727, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7109867334365845, + "num_tokens": 650761726.0, + "step": 25146 + }, + { + "epoch": 2.7615857676257414, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.590994358062744, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7344568967819214, + "num_tokens": 650784206.0, + "step": 25147 + }, + { + "epoch": 2.761695585328355, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5390827655792236, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7204679250717163, + "num_tokens": 650807736.0, + "step": 25148 + }, + { + "epoch": 2.7618054030309684, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4456229209899902, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7366070747375488, + "num_tokens": 650830542.0, + "step": 25149 + }, + { + "epoch": 2.761915220733582, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2551207542419434, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7370829582214355, + "num_tokens": 650859543.0, + "step": 25150 + }, + { + "epoch": 2.762025038436196, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5099167823791504, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7436295747756958, + "num_tokens": 650881841.0, + "step": 25151 + }, + { + "epoch": 2.7621348561388097, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.403736114501953, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7438758015632629, + "num_tokens": 650906005.0, + "step": 25152 + }, + { + "epoch": 2.7622446738414235, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.185333490371704, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7332123517990112, + "num_tokens": 650933378.0, + "step": 25153 + }, + { + "epoch": 2.762354491544037, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4059035778045654, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7135322093963623, + "num_tokens": 650960849.0, + "step": 25154 + }, + { + "epoch": 2.7624643092466505, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.0114402770996094, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.725969672203064, + "num_tokens": 650990245.0, + "step": 25155 + }, + { + "epoch": 2.7625741269492643, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.8445112705230713, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.74167400598526, + "num_tokens": 651007513.0, + "step": 25156 + }, + { + "epoch": 2.7626839446518776, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.445336103439331, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.733399510383606, + "num_tokens": 651033184.0, + "step": 25157 + }, + { + "epoch": 2.762793762354492, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2416129112243652, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7282043099403381, + "num_tokens": 651061072.0, + "step": 25158 + }, + { + "epoch": 2.762903580057105, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.458348035812378, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7011160254478455, + "num_tokens": 651086402.0, + "step": 25159 + }, + { + "epoch": 2.763013397759719, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4022786617279053, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7263155579566956, + "num_tokens": 651110355.0, + "step": 25160 + }, + { + "epoch": 2.7631232154623326, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4006869792938232, + "learning_rate": 1e-06, + "loss": 0.6901, + "mean_token_accuracy": 0.7853120565414429, + "num_tokens": 651132937.0, + "step": 25161 + }, + { + "epoch": 2.763233033164946, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2976245880126953, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7293559312820435, + "num_tokens": 651159254.0, + "step": 25162 + }, + { + "epoch": 2.7633428508675597, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4081032276153564, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7182433009147644, + "num_tokens": 651183771.0, + "step": 25163 + }, + { + "epoch": 2.7634526685701735, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3773131370544434, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7308658361434937, + "num_tokens": 651210647.0, + "step": 25164 + }, + { + "epoch": 2.763562486272787, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6098885536193848, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7276781797409058, + "num_tokens": 651232450.0, + "step": 25165 + }, + { + "epoch": 2.763672303975401, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.302872657775879, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7111449837684631, + "num_tokens": 651262363.0, + "step": 25166 + }, + { + "epoch": 2.7637821216780143, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.209655523300171, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7239980101585388, + "num_tokens": 651294423.0, + "step": 25167 + }, + { + "epoch": 2.763891939380628, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.438758611679077, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7200825214385986, + "num_tokens": 651320244.0, + "step": 25168 + }, + { + "epoch": 2.764001757083242, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4437267780303955, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7270735502243042, + "num_tokens": 651345228.0, + "step": 25169 + }, + { + "epoch": 2.7641115747858556, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3728883266448975, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7390176653862, + "num_tokens": 651370252.0, + "step": 25170 + }, + { + "epoch": 2.7642213924884693, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.12908673286438, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7373591661453247, + "num_tokens": 651399536.0, + "step": 25171 + }, + { + "epoch": 2.7643312101910826, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3042120933532715, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.709730863571167, + "num_tokens": 651427551.0, + "step": 25172 + }, + { + "epoch": 2.7644410278936964, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2523648738861084, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7228840589523315, + "num_tokens": 651456709.0, + "step": 25173 + }, + { + "epoch": 2.76455084559631, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4636218547821045, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7287230491638184, + "num_tokens": 651483211.0, + "step": 25174 + }, + { + "epoch": 2.764660663298924, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.7823879718780518, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7433614134788513, + "num_tokens": 651503020.0, + "step": 25175 + }, + { + "epoch": 2.7647704810015377, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4021363258361816, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7279624342918396, + "num_tokens": 651527231.0, + "step": 25176 + }, + { + "epoch": 2.764880298704151, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3134801387786865, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7443612813949585, + "num_tokens": 651554015.0, + "step": 25177 + }, + { + "epoch": 2.7649901164067647, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.370007038116455, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7306122779846191, + "num_tokens": 651580250.0, + "step": 25178 + }, + { + "epoch": 2.7650999341093785, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.225895404815674, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7356160879135132, + "num_tokens": 651606831.0, + "step": 25179 + }, + { + "epoch": 2.7652097518119922, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4639365673065186, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7329141497612, + "num_tokens": 651629974.0, + "step": 25180 + }, + { + "epoch": 2.765319569514606, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.261007070541382, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7314961552619934, + "num_tokens": 651657057.0, + "step": 25181 + }, + { + "epoch": 2.7654293872172193, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3699212074279785, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7443134784698486, + "num_tokens": 651682385.0, + "step": 25182 + }, + { + "epoch": 2.765539204919833, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4061636924743652, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7244476079940796, + "num_tokens": 651706546.0, + "step": 25183 + }, + { + "epoch": 2.765649022622447, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1198315620422363, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7469011545181274, + "num_tokens": 651734169.0, + "step": 25184 + }, + { + "epoch": 2.76575884032506, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3444128036499023, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7327888011932373, + "num_tokens": 651760196.0, + "step": 25185 + }, + { + "epoch": 2.765868658027674, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.182028293609619, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7171968221664429, + "num_tokens": 651791156.0, + "step": 25186 + }, + { + "epoch": 2.7659784757302877, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.7773518562316895, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7500518560409546, + "num_tokens": 651809139.0, + "step": 25187 + }, + { + "epoch": 2.7660882934329014, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6425397396087646, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.7424401044845581, + "num_tokens": 651829919.0, + "step": 25188 + }, + { + "epoch": 2.766198111135515, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.259326457977295, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7129266858100891, + "num_tokens": 651859312.0, + "step": 25189 + }, + { + "epoch": 2.7663079288381285, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1703808307647705, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7281211614608765, + "num_tokens": 651889055.0, + "step": 25190 + }, + { + "epoch": 2.7664177465407422, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.798631191253662, + "learning_rate": 1e-06, + "loss": 0.7083, + "mean_token_accuracy": 0.7664992213249207, + "num_tokens": 651906880.0, + "step": 25191 + }, + { + "epoch": 2.766527564243356, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4737207889556885, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7425392866134644, + "num_tokens": 651929096.0, + "step": 25192 + }, + { + "epoch": 2.7666373819459698, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.7405319213867188, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7154651880264282, + "num_tokens": 651951758.0, + "step": 25193 + }, + { + "epoch": 2.7667471996485835, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3395872116088867, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7259109020233154, + "num_tokens": 651976696.0, + "step": 25194 + }, + { + "epoch": 2.766857017351197, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.25179123878479, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.743877649307251, + "num_tokens": 652004851.0, + "step": 25195 + }, + { + "epoch": 2.7669668350538106, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.593816041946411, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7288240194320679, + "num_tokens": 652027235.0, + "step": 25196 + }, + { + "epoch": 2.7670766527564243, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.266808032989502, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7235433459281921, + "num_tokens": 652053645.0, + "step": 25197 + }, + { + "epoch": 2.767186470459038, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3072774410247803, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7196600437164307, + "num_tokens": 652080875.0, + "step": 25198 + }, + { + "epoch": 2.767296288161652, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4087276458740234, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.719404935836792, + "num_tokens": 652105946.0, + "step": 25199 + }, + { + "epoch": 2.767406105864265, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.36718487739563, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.729597806930542, + "num_tokens": 652132432.0, + "step": 25200 + }, + { + "epoch": 2.767515923566879, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.005892753601074, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7010543346405029, + "num_tokens": 652168571.0, + "step": 25201 + }, + { + "epoch": 2.7676257412694927, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.494622230529785, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7344719171524048, + "num_tokens": 652192561.0, + "step": 25202 + }, + { + "epoch": 2.7677355589721064, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.531242847442627, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7291804552078247, + "num_tokens": 652216199.0, + "step": 25203 + }, + { + "epoch": 2.76784537667472, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5968334674835205, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7164747714996338, + "num_tokens": 652240440.0, + "step": 25204 + }, + { + "epoch": 2.7679551943773335, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5283913612365723, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7239298820495605, + "num_tokens": 652264582.0, + "step": 25205 + }, + { + "epoch": 2.7680650120799473, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2795188426971436, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7504929304122925, + "num_tokens": 652291585.0, + "step": 25206 + }, + { + "epoch": 2.768174829782561, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.184187889099121, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7080734968185425, + "num_tokens": 652323196.0, + "step": 25207 + }, + { + "epoch": 2.7682846474851748, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.180013656616211, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7020914554595947, + "num_tokens": 652355899.0, + "step": 25208 + }, + { + "epoch": 2.7683944651877885, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4136412143707275, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7031494379043579, + "num_tokens": 652381810.0, + "step": 25209 + }, + { + "epoch": 2.768504282890402, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.60062837600708, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7497231960296631, + "num_tokens": 652403246.0, + "step": 25210 + }, + { + "epoch": 2.7686141005930156, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4540841579437256, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7146899700164795, + "num_tokens": 652427886.0, + "step": 25211 + }, + { + "epoch": 2.7687239182956294, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2925450801849365, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7348730564117432, + "num_tokens": 652454926.0, + "step": 25212 + }, + { + "epoch": 2.7688337359982427, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3110032081604004, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7451728582382202, + "num_tokens": 652481219.0, + "step": 25213 + }, + { + "epoch": 2.7689435537008564, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.216726541519165, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7170847058296204, + "num_tokens": 652507642.0, + "step": 25214 + }, + { + "epoch": 2.76905337140347, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.535757303237915, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7554062604904175, + "num_tokens": 652530668.0, + "step": 25215 + }, + { + "epoch": 2.769163189106084, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.59045672416687, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.736785888671875, + "num_tokens": 652552515.0, + "step": 25216 + }, + { + "epoch": 2.7692730068086977, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.619091749191284, + "learning_rate": 1e-06, + "loss": 0.7846, + "mean_token_accuracy": 0.7490233182907104, + "num_tokens": 652572944.0, + "step": 25217 + }, + { + "epoch": 2.769382824511311, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1760830879211426, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7215309143066406, + "num_tokens": 652601639.0, + "step": 25218 + }, + { + "epoch": 2.7694926422139248, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3829615116119385, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.715445876121521, + "num_tokens": 652626592.0, + "step": 25219 + }, + { + "epoch": 2.7696024599165385, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.27927827835083, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7020146250724792, + "num_tokens": 652655322.0, + "step": 25220 + }, + { + "epoch": 2.7697122776191523, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.562269687652588, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7460829019546509, + "num_tokens": 652679234.0, + "step": 25221 + }, + { + "epoch": 2.769822095321766, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.78324556350708, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7347338795661926, + "num_tokens": 652699212.0, + "step": 25222 + }, + { + "epoch": 2.7699319130243794, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.279280662536621, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7371242642402649, + "num_tokens": 652726972.0, + "step": 25223 + }, + { + "epoch": 2.770041730726993, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5268638134002686, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7495648264884949, + "num_tokens": 652748924.0, + "step": 25224 + }, + { + "epoch": 2.770151548429607, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.56105637550354, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7254352569580078, + "num_tokens": 652772025.0, + "step": 25225 + }, + { + "epoch": 2.7702613661322206, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.346938133239746, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7468396425247192, + "num_tokens": 652798692.0, + "step": 25226 + }, + { + "epoch": 2.7703711838348344, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1481285095214844, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7259807586669922, + "num_tokens": 652830187.0, + "step": 25227 + }, + { + "epoch": 2.7704810015374477, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.454505443572998, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7389139533042908, + "num_tokens": 652853772.0, + "step": 25228 + }, + { + "epoch": 2.7705908192400615, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.489225387573242, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7275869250297546, + "num_tokens": 652879406.0, + "step": 25229 + }, + { + "epoch": 2.770700636942675, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3969056606292725, + "learning_rate": 1e-06, + "loss": 0.7712, + "mean_token_accuracy": 0.7469659447669983, + "num_tokens": 652904322.0, + "step": 25230 + }, + { + "epoch": 2.770810454645289, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.20924973487854, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7324600219726562, + "num_tokens": 652932569.0, + "step": 25231 + }, + { + "epoch": 2.7709202723479027, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.368396759033203, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.728242039680481, + "num_tokens": 652957935.0, + "step": 25232 + }, + { + "epoch": 2.771030090050516, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3405330181121826, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.722771406173706, + "num_tokens": 652983410.0, + "step": 25233 + }, + { + "epoch": 2.77113990775313, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2800137996673584, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7197177410125732, + "num_tokens": 653009975.0, + "step": 25234 + }, + { + "epoch": 2.7712497254557436, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2256503105163574, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7439982891082764, + "num_tokens": 653036431.0, + "step": 25235 + }, + { + "epoch": 2.771359543158357, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.550649642944336, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7448313236236572, + "num_tokens": 653058546.0, + "step": 25236 + }, + { + "epoch": 2.771469360860971, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3208742141723633, + "learning_rate": 1e-06, + "loss": 0.7586, + "mean_token_accuracy": 0.758050799369812, + "num_tokens": 653083335.0, + "step": 25237 + }, + { + "epoch": 2.7715791785635844, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2565464973449707, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7429386377334595, + "num_tokens": 653110027.0, + "step": 25238 + }, + { + "epoch": 2.771688996266198, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3467636108398438, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7344201803207397, + "num_tokens": 653136265.0, + "step": 25239 + }, + { + "epoch": 2.771798813968812, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.482477903366089, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7254558205604553, + "num_tokens": 653159842.0, + "step": 25240 + }, + { + "epoch": 2.771908631671425, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.431920289993286, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7302071452140808, + "num_tokens": 653186815.0, + "step": 25241 + }, + { + "epoch": 2.772018449374039, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4676015377044678, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7176938652992249, + "num_tokens": 653211266.0, + "step": 25242 + }, + { + "epoch": 2.7721282670766527, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.282400608062744, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7231833934783936, + "num_tokens": 653240510.0, + "step": 25243 + }, + { + "epoch": 2.7722380847792665, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.633227586746216, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7147040963172913, + "num_tokens": 653261176.0, + "step": 25244 + }, + { + "epoch": 2.7723479024818802, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2293150424957275, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7072476148605347, + "num_tokens": 653290730.0, + "step": 25245 + }, + { + "epoch": 2.7724577201844935, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.410369634628296, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7170741558074951, + "num_tokens": 653316082.0, + "step": 25246 + }, + { + "epoch": 2.7725675378871073, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3390085697174072, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7426764965057373, + "num_tokens": 653341875.0, + "step": 25247 + }, + { + "epoch": 2.772677355589721, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.45847225189209, + "learning_rate": 1e-06, + "loss": 0.7785, + "mean_token_accuracy": 0.7527198791503906, + "num_tokens": 653366545.0, + "step": 25248 + }, + { + "epoch": 2.772787173292335, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.479280471801758, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7242239713668823, + "num_tokens": 653391297.0, + "step": 25249 + }, + { + "epoch": 2.7728969909949486, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.433711051940918, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.737006425857544, + "num_tokens": 653414622.0, + "step": 25250 + }, + { + "epoch": 2.773006808697562, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4481701850891113, + "learning_rate": 1e-06, + "loss": 0.7922, + "mean_token_accuracy": 0.7507448196411133, + "num_tokens": 653438389.0, + "step": 25251 + }, + { + "epoch": 2.7731166264001756, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.452233076095581, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7238677740097046, + "num_tokens": 653462007.0, + "step": 25252 + }, + { + "epoch": 2.7732264441027894, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1524252891540527, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7209247946739197, + "num_tokens": 653491402.0, + "step": 25253 + }, + { + "epoch": 2.773336261805403, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2450830936431885, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7100451588630676, + "num_tokens": 653521194.0, + "step": 25254 + }, + { + "epoch": 2.773446079508017, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4325757026672363, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7194675207138062, + "num_tokens": 653546587.0, + "step": 25255 + }, + { + "epoch": 2.7735558972106302, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3659379482269287, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7235387563705444, + "num_tokens": 653572844.0, + "step": 25256 + }, + { + "epoch": 2.773665714913244, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.0961060523986816, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7170358300209045, + "num_tokens": 653604279.0, + "step": 25257 + }, + { + "epoch": 2.7737755326158577, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2295758724212646, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7191734910011292, + "num_tokens": 653630499.0, + "step": 25258 + }, + { + "epoch": 2.7738853503184715, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4885356426239014, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7322986721992493, + "num_tokens": 653652860.0, + "step": 25259 + }, + { + "epoch": 2.7739951680210853, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.483356475830078, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.720933735370636, + "num_tokens": 653678067.0, + "step": 25260 + }, + { + "epoch": 2.7741049857236986, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3000638484954834, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7254869937896729, + "num_tokens": 653705602.0, + "step": 25261 + }, + { + "epoch": 2.7742148034263123, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.548931360244751, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7526225447654724, + "num_tokens": 653728381.0, + "step": 25262 + }, + { + "epoch": 2.774324621128926, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.487046241760254, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.726465106010437, + "num_tokens": 653752289.0, + "step": 25263 + }, + { + "epoch": 2.7744344388315394, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.508047580718994, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7413825988769531, + "num_tokens": 653776572.0, + "step": 25264 + }, + { + "epoch": 2.774544256534153, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4020142555236816, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7388145923614502, + "num_tokens": 653801743.0, + "step": 25265 + }, + { + "epoch": 2.774654074236767, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.397421360015869, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.715954065322876, + "num_tokens": 653828938.0, + "step": 25266 + }, + { + "epoch": 2.7747638919393807, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.301208257675171, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7294387221336365, + "num_tokens": 653855618.0, + "step": 25267 + }, + { + "epoch": 2.7748737096419944, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.507366418838501, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7445923686027527, + "num_tokens": 653878047.0, + "step": 25268 + }, + { + "epoch": 2.7749835273446077, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.084935188293457, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7035171985626221, + "num_tokens": 653910661.0, + "step": 25269 + }, + { + "epoch": 2.7750933450472215, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6302075386047363, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7445740699768066, + "num_tokens": 653931833.0, + "step": 25270 + }, + { + "epoch": 2.7752031627498353, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3783152103424072, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.721436619758606, + "num_tokens": 653956526.0, + "step": 25271 + }, + { + "epoch": 2.775312980452449, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.367462158203125, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7228633165359497, + "num_tokens": 653981615.0, + "step": 25272 + }, + { + "epoch": 2.7754227981550628, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.212902069091797, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7366626262664795, + "num_tokens": 654009960.0, + "step": 25273 + }, + { + "epoch": 2.775532615857676, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5760879516601562, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7300388216972351, + "num_tokens": 654032444.0, + "step": 25274 + }, + { + "epoch": 2.77564243356029, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.9077694416046143, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7314821481704712, + "num_tokens": 654051166.0, + "step": 25275 + }, + { + "epoch": 2.7757522512629036, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.444913148880005, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7131596207618713, + "num_tokens": 654076902.0, + "step": 25276 + }, + { + "epoch": 2.7758620689655173, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.739631414413452, + "learning_rate": 1e-06, + "loss": 0.7846, + "mean_token_accuracy": 0.7566287517547607, + "num_tokens": 654096938.0, + "step": 25277 + }, + { + "epoch": 2.775971886668131, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.29569935798645, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7344348430633545, + "num_tokens": 654122767.0, + "step": 25278 + }, + { + "epoch": 2.7760817043707444, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5685737133026123, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7281753420829773, + "num_tokens": 654145677.0, + "step": 25279 + }, + { + "epoch": 2.776191522073358, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.394578695297241, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7343653440475464, + "num_tokens": 654169535.0, + "step": 25280 + }, + { + "epoch": 2.776301339775972, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.391688108444214, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.7601872086524963, + "num_tokens": 654192822.0, + "step": 25281 + }, + { + "epoch": 2.7764111574785857, + "ewc_loss": 0.00022125244140625, + "grad_norm": 1.9962882995605469, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7225906252861023, + "num_tokens": 654225419.0, + "step": 25282 + }, + { + "epoch": 2.7765209751811994, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3407888412475586, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7330639958381653, + "num_tokens": 654249585.0, + "step": 25283 + }, + { + "epoch": 2.7766307928838128, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.4528591632843018, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7306329011917114, + "num_tokens": 654273537.0, + "step": 25284 + }, + { + "epoch": 2.7767406105864265, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1724157333374023, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7210602760314941, + "num_tokens": 654303510.0, + "step": 25285 + }, + { + "epoch": 2.7768504282890403, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.099623918533325, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7103468179702759, + "num_tokens": 654333327.0, + "step": 25286 + }, + { + "epoch": 2.7769602459916536, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.703422784805298, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7335226535797119, + "num_tokens": 654353996.0, + "step": 25287 + }, + { + "epoch": 2.777070063694268, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.72800612449646, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7482022047042847, + "num_tokens": 654374873.0, + "step": 25288 + }, + { + "epoch": 2.777179881396881, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5156033039093018, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7435800433158875, + "num_tokens": 654397303.0, + "step": 25289 + }, + { + "epoch": 2.777289699099495, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5018086433410645, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7155022621154785, + "num_tokens": 654421715.0, + "step": 25290 + }, + { + "epoch": 2.7773995168021086, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.36222243309021, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7230198979377747, + "num_tokens": 654446136.0, + "step": 25291 + }, + { + "epoch": 2.777509334504722, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.418836832046509, + "learning_rate": 1e-06, + "loss": 0.747, + "mean_token_accuracy": 0.7579522728919983, + "num_tokens": 654469152.0, + "step": 25292 + }, + { + "epoch": 2.7776191522073357, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2376842498779297, + "learning_rate": 1e-06, + "loss": 0.7797, + "mean_token_accuracy": 0.7639300227165222, + "num_tokens": 654495377.0, + "step": 25293 + }, + { + "epoch": 2.7777289699099494, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6141586303710938, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7282938361167908, + "num_tokens": 654516281.0, + "step": 25294 + }, + { + "epoch": 2.777838787612563, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.483755350112915, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7363013625144958, + "num_tokens": 654540311.0, + "step": 25295 + }, + { + "epoch": 2.777948605315177, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.6142210960388184, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7316924333572388, + "num_tokens": 654563102.0, + "step": 25296 + }, + { + "epoch": 2.7780584230177903, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.2323520183563232, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7151161432266235, + "num_tokens": 654592510.0, + "step": 25297 + }, + { + "epoch": 2.778168240720404, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.450047254562378, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7398133277893066, + "num_tokens": 654615826.0, + "step": 25298 + }, + { + "epoch": 2.778278058423018, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.5034773349761963, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7212502956390381, + "num_tokens": 654641014.0, + "step": 25299 + }, + { + "epoch": 2.7783878761256315, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.255711317062378, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7420369386672974, + "num_tokens": 654669484.0, + "step": 25300 + }, + { + "epoch": 2.7784976938282453, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.517238140106201, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7285711169242859, + "num_tokens": 654691943.0, + "step": 25301 + }, + { + "epoch": 2.7786075115308586, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.1041150093078613, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7385707497596741, + "num_tokens": 654724098.0, + "step": 25302 + }, + { + "epoch": 2.7787173292334724, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.3275115489959717, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7251719236373901, + "num_tokens": 654751649.0, + "step": 25303 + }, + { + "epoch": 2.778827146936086, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.724092960357666, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7426319122314453, + "num_tokens": 654774086.0, + "step": 25304 + }, + { + "epoch": 2.7789369646387, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.314758062362671, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7107238173484802, + "num_tokens": 654800126.0, + "step": 25305 + }, + { + "epoch": 2.7790467823413136, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.279540777206421, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7186930179595947, + "num_tokens": 654828206.0, + "step": 25306 + }, + { + "epoch": 2.779156600043927, + "ewc_loss": 0.00022125244140625, + "grad_norm": 2.67299747467041, + "learning_rate": 1e-06, + "loss": 0.7123, + "mean_token_accuracy": 0.7722257971763611, + "num_tokens": 654847164.0, + "step": 25307 + }, + { + "epoch": 2.7792664177465407, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.6331217288970947, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.73881596326828, + "num_tokens": 654869336.0, + "step": 25308 + }, + { + "epoch": 2.7793762354491545, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.1942293643951416, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7392070293426514, + "num_tokens": 654897779.0, + "step": 25309 + }, + { + "epoch": 2.779486053151768, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.7425293922424316, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7336824536323547, + "num_tokens": 654917220.0, + "step": 25310 + }, + { + "epoch": 2.779595870854382, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.67226505279541, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7274847030639648, + "num_tokens": 654939692.0, + "step": 25311 + }, + { + "epoch": 2.7797056885569953, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.340209722518921, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7035344839096069, + "num_tokens": 654965842.0, + "step": 25312 + }, + { + "epoch": 2.779815506259609, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.300527334213257, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7163342833518982, + "num_tokens": 654993830.0, + "step": 25313 + }, + { + "epoch": 2.779925323962223, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.4681010246276855, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7358524203300476, + "num_tokens": 655015958.0, + "step": 25314 + }, + { + "epoch": 2.780035141664836, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.2576205730438232, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7476537227630615, + "num_tokens": 655044444.0, + "step": 25315 + }, + { + "epoch": 2.78014495936745, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.2575178146362305, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7385904788970947, + "num_tokens": 655071755.0, + "step": 25316 + }, + { + "epoch": 2.7802547770700636, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.3265693187713623, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.6999499797821045, + "num_tokens": 655099814.0, + "step": 25317 + }, + { + "epoch": 2.7803645947726774, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.271085262298584, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7434388399124146, + "num_tokens": 655129052.0, + "step": 25318 + }, + { + "epoch": 2.780474412475291, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.369645833969116, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7455556988716125, + "num_tokens": 655153983.0, + "step": 25319 + }, + { + "epoch": 2.7805842301779045, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.3561577796936035, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7226867079734802, + "num_tokens": 655180614.0, + "step": 25320 + }, + { + "epoch": 2.780694047880518, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.2523200511932373, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7166386246681213, + "num_tokens": 655209629.0, + "step": 25321 + }, + { + "epoch": 2.780803865583132, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.388185977935791, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7415779829025269, + "num_tokens": 655234356.0, + "step": 25322 + }, + { + "epoch": 2.7809136832857457, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.2363622188568115, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7281537055969238, + "num_tokens": 655263774.0, + "step": 25323 + }, + { + "epoch": 2.7810235009883595, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.3955862522125244, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7354961037635803, + "num_tokens": 655290867.0, + "step": 25324 + }, + { + "epoch": 2.781133318690973, + "ewc_loss": 0.00022220611572265625, + "grad_norm": 2.2372941970825195, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7186543941497803, + "num_tokens": 655319273.0, + "step": 25325 + }, + { + "epoch": 2.7812431363935866, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.371556043624878, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7463620901107788, + "num_tokens": 655344099.0, + "step": 25326 + }, + { + "epoch": 2.7813529540962003, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.1095950603485107, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7055176496505737, + "num_tokens": 655376303.0, + "step": 25327 + }, + { + "epoch": 2.781462771798814, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3667168617248535, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7355111837387085, + "num_tokens": 655404148.0, + "step": 25328 + }, + { + "epoch": 2.781572589501428, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.6297249794006348, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7221498489379883, + "num_tokens": 655425556.0, + "step": 25329 + }, + { + "epoch": 2.781682407204041, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4727587699890137, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7529958486557007, + "num_tokens": 655448683.0, + "step": 25330 + }, + { + "epoch": 2.781792224906655, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2510197162628174, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7353758215904236, + "num_tokens": 655475223.0, + "step": 25331 + }, + { + "epoch": 2.7819020426092687, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.1835501194000244, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7394251823425293, + "num_tokens": 655503829.0, + "step": 25332 + }, + { + "epoch": 2.7820118603118824, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.288364887237549, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7308845520019531, + "num_tokens": 655530470.0, + "step": 25333 + }, + { + "epoch": 2.782121678014496, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.5994789600372314, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7373725771903992, + "num_tokens": 655553107.0, + "step": 25334 + }, + { + "epoch": 2.7822314957171095, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.384509563446045, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7329239845275879, + "num_tokens": 655579577.0, + "step": 25335 + }, + { + "epoch": 2.7823413134197232, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.431467294692993, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7444335222244263, + "num_tokens": 655604063.0, + "step": 25336 + }, + { + "epoch": 2.782451131122337, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4165337085723877, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7446386814117432, + "num_tokens": 655629811.0, + "step": 25337 + }, + { + "epoch": 2.7825609488249503, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3825435638427734, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7124927043914795, + "num_tokens": 655658096.0, + "step": 25338 + }, + { + "epoch": 2.7826707665275645, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3638360500335693, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7195532321929932, + "num_tokens": 655686339.0, + "step": 25339 + }, + { + "epoch": 2.782780584230178, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.579514980316162, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.717282772064209, + "num_tokens": 655710433.0, + "step": 25340 + }, + { + "epoch": 2.7828904019327916, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.247479200363159, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7266985177993774, + "num_tokens": 655738374.0, + "step": 25341 + }, + { + "epoch": 2.7830002196354053, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4362847805023193, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7295730113983154, + "num_tokens": 655763612.0, + "step": 25342 + }, + { + "epoch": 2.7831100373380186, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.208052158355713, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.717145562171936, + "num_tokens": 655794128.0, + "step": 25343 + }, + { + "epoch": 2.7832198550406324, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.5634422302246094, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7379018068313599, + "num_tokens": 655817587.0, + "step": 25344 + }, + { + "epoch": 2.783329672743246, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.6261868476867676, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7327345609664917, + "num_tokens": 655840349.0, + "step": 25345 + }, + { + "epoch": 2.78343949044586, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4720678329467773, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7390924096107483, + "num_tokens": 655863924.0, + "step": 25346 + }, + { + "epoch": 2.7835493081484737, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.633183479309082, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7335410118103027, + "num_tokens": 655886611.0, + "step": 25347 + }, + { + "epoch": 2.783659125851087, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2785043716430664, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7326834201812744, + "num_tokens": 655914869.0, + "step": 25348 + }, + { + "epoch": 2.7837689435537007, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3218958377838135, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7420375347137451, + "num_tokens": 655941675.0, + "step": 25349 + }, + { + "epoch": 2.7838787612563145, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.142808198928833, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7073616981506348, + "num_tokens": 655971798.0, + "step": 25350 + }, + { + "epoch": 2.7839885789589283, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3544094562530518, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7514026165008545, + "num_tokens": 655995942.0, + "step": 25351 + }, + { + "epoch": 2.784098396661542, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.05043888092041, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7099663019180298, + "num_tokens": 656030686.0, + "step": 25352 + }, + { + "epoch": 2.7842082143641553, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2568581104278564, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7265493869781494, + "num_tokens": 656060557.0, + "step": 25353 + }, + { + "epoch": 2.784318032066769, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.29080867767334, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7474662065505981, + "num_tokens": 656086004.0, + "step": 25354 + }, + { + "epoch": 2.784427849769383, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.418220281600952, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7200725674629211, + "num_tokens": 656111463.0, + "step": 25355 + }, + { + "epoch": 2.7845376674719966, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.563178062438965, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7294321060180664, + "num_tokens": 656136462.0, + "step": 25356 + }, + { + "epoch": 2.7846474851746104, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4513227939605713, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7324153780937195, + "num_tokens": 656160109.0, + "step": 25357 + }, + { + "epoch": 2.7847573028772237, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3545491695404053, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7249257564544678, + "num_tokens": 656184727.0, + "step": 25358 + }, + { + "epoch": 2.7848671205798374, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.7925238609313965, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7368324398994446, + "num_tokens": 656204424.0, + "step": 25359 + }, + { + "epoch": 2.784976938282451, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3101799488067627, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7211109399795532, + "num_tokens": 656231785.0, + "step": 25360 + }, + { + "epoch": 2.785086755985065, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.5299603939056396, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.729348361492157, + "num_tokens": 656255960.0, + "step": 25361 + }, + { + "epoch": 2.7851965736876787, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 1.9723787307739258, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7031112909317017, + "num_tokens": 656289877.0, + "step": 25362 + }, + { + "epoch": 2.785306391390292, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.596503973007202, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7185835838317871, + "num_tokens": 656313410.0, + "step": 25363 + }, + { + "epoch": 2.7854162090929058, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4097256660461426, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7125762701034546, + "num_tokens": 656338139.0, + "step": 25364 + }, + { + "epoch": 2.7855260267955195, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.3906171321868896, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7194598913192749, + "num_tokens": 656365724.0, + "step": 25365 + }, + { + "epoch": 2.785635844498133, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2830405235290527, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7316076755523682, + "num_tokens": 656397010.0, + "step": 25366 + }, + { + "epoch": 2.7857456622007466, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2443573474884033, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7002065181732178, + "num_tokens": 656426045.0, + "step": 25367 + }, + { + "epoch": 2.7858554799033604, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.742314338684082, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7493345141410828, + "num_tokens": 656445907.0, + "step": 25368 + }, + { + "epoch": 2.785965297605974, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.381190299987793, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7199139595031738, + "num_tokens": 656471183.0, + "step": 25369 + }, + { + "epoch": 2.786075115308588, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.240122079849243, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7277183532714844, + "num_tokens": 656499016.0, + "step": 25370 + }, + { + "epoch": 2.786184933011201, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.5136067867279053, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7273542881011963, + "num_tokens": 656523225.0, + "step": 25371 + }, + { + "epoch": 2.786294750713815, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.445773124694824, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.7072054147720337, + "num_tokens": 656548239.0, + "step": 25372 + }, + { + "epoch": 2.7864045684164287, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2504310607910156, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7352882623672485, + "num_tokens": 656574403.0, + "step": 25373 + }, + { + "epoch": 2.7865143861190425, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.488300085067749, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7435047626495361, + "num_tokens": 656597325.0, + "step": 25374 + }, + { + "epoch": 2.786624203821656, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.219442844390869, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7281984686851501, + "num_tokens": 656624767.0, + "step": 25375 + }, + { + "epoch": 2.7867340215242695, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4370691776275635, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7292839884757996, + "num_tokens": 656650865.0, + "step": 25376 + }, + { + "epoch": 2.7868438392268833, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.2637367248535156, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7251787185668945, + "num_tokens": 656679434.0, + "step": 25377 + }, + { + "epoch": 2.786953656929497, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.0992987155914307, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7372989654541016, + "num_tokens": 656707461.0, + "step": 25378 + }, + { + "epoch": 2.787063474632111, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4913442134857178, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7362568378448486, + "num_tokens": 656731067.0, + "step": 25379 + }, + { + "epoch": 2.7871732923347246, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4617271423339844, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7068857550621033, + "num_tokens": 656756016.0, + "step": 25380 + }, + { + "epoch": 2.787283110037338, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5937869548797607, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7477097511291504, + "num_tokens": 656776780.0, + "step": 25381 + }, + { + "epoch": 2.7873929277399516, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4025375843048096, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7231807112693787, + "num_tokens": 656802819.0, + "step": 25382 + }, + { + "epoch": 2.7875027454425654, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.602154493331909, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7321483492851257, + "num_tokens": 656825854.0, + "step": 25383 + }, + { + "epoch": 2.787612563145179, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.8185713291168213, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7448427677154541, + "num_tokens": 656846330.0, + "step": 25384 + }, + { + "epoch": 2.787722380847793, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.606600284576416, + "learning_rate": 1e-06, + "loss": 0.8274, + "mean_token_accuracy": 0.7414078712463379, + "num_tokens": 656870508.0, + "step": 25385 + }, + { + "epoch": 2.787832198550406, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.4888622760772705, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7479848265647888, + "num_tokens": 656894746.0, + "step": 25386 + }, + { + "epoch": 2.78794201625302, + "ewc_loss": 0.0002231597900390625, + "grad_norm": 2.1560416221618652, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.739414393901825, + "num_tokens": 656924634.0, + "step": 25387 + }, + { + "epoch": 2.7880518339556337, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1732945442199707, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7403091788291931, + "num_tokens": 656954425.0, + "step": 25388 + }, + { + "epoch": 2.7881616516582475, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3870084285736084, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7215134501457214, + "num_tokens": 656981932.0, + "step": 25389 + }, + { + "epoch": 2.7882714693608612, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2415544986724854, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6954634189605713, + "num_tokens": 657014208.0, + "step": 25390 + }, + { + "epoch": 2.7883812870634745, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.0595881938934326, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.725104033946991, + "num_tokens": 657043290.0, + "step": 25391 + }, + { + "epoch": 2.7884911047660883, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.6577041149139404, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7420375347137451, + "num_tokens": 657064427.0, + "step": 25392 + }, + { + "epoch": 2.788600922468702, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3501946926116943, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7305126190185547, + "num_tokens": 657089835.0, + "step": 25393 + }, + { + "epoch": 2.7887107401713154, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2754099369049072, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7309507727622986, + "num_tokens": 657117918.0, + "step": 25394 + }, + { + "epoch": 2.788820557873929, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.134068489074707, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.6980981230735779, + "num_tokens": 657150714.0, + "step": 25395 + }, + { + "epoch": 2.788930375576543, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3713223934173584, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7299972772598267, + "num_tokens": 657175174.0, + "step": 25396 + }, + { + "epoch": 2.7890401932791566, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.405792713165283, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7189139127731323, + "num_tokens": 657202885.0, + "step": 25397 + }, + { + "epoch": 2.7891500109817704, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1825969219207764, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7253855466842651, + "num_tokens": 657231845.0, + "step": 25398 + }, + { + "epoch": 2.7892598286843837, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1775295734405518, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.728553056716919, + "num_tokens": 657261026.0, + "step": 25399 + }, + { + "epoch": 2.7893696463869975, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.210426092147827, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.715785562992096, + "num_tokens": 657291041.0, + "step": 25400 + }, + { + "epoch": 2.7894794640896112, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1871461868286133, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7306121587753296, + "num_tokens": 657319078.0, + "step": 25401 + }, + { + "epoch": 2.789589281792225, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.829888343811035, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.7611596584320068, + "num_tokens": 657338584.0, + "step": 25402 + }, + { + "epoch": 2.7896990994948387, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.6178338527679443, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7236245274543762, + "num_tokens": 657361728.0, + "step": 25403 + }, + { + "epoch": 2.789808917197452, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5848748683929443, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7412802577018738, + "num_tokens": 657384977.0, + "step": 25404 + }, + { + "epoch": 2.789918734900066, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.144896984100342, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7108465433120728, + "num_tokens": 657415883.0, + "step": 25405 + }, + { + "epoch": 2.7900285526026796, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.510718584060669, + "learning_rate": 1e-06, + "loss": 0.7404, + "mean_token_accuracy": 0.7672860622406006, + "num_tokens": 657438089.0, + "step": 25406 + }, + { + "epoch": 2.7901383703052933, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.162945032119751, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.719162106513977, + "num_tokens": 657467505.0, + "step": 25407 + }, + { + "epoch": 2.790248188007907, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.274415969848633, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7126154899597168, + "num_tokens": 657496368.0, + "step": 25408 + }, + { + "epoch": 2.7903580057105204, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.0817854404449463, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7292656898498535, + "num_tokens": 657529212.0, + "step": 25409 + }, + { + "epoch": 2.790467823413134, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2236149311065674, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7313483357429504, + "num_tokens": 657556847.0, + "step": 25410 + }, + { + "epoch": 2.790577641115748, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1696112155914307, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.710233747959137, + "num_tokens": 657587255.0, + "step": 25411 + }, + { + "epoch": 2.7906874588183617, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.761209011077881, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.7540916204452515, + "num_tokens": 657607807.0, + "step": 25412 + }, + { + "epoch": 2.7907972765209754, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2983217239379883, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7558238506317139, + "num_tokens": 657632519.0, + "step": 25413 + }, + { + "epoch": 2.7909070942235887, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3090708255767822, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7018821239471436, + "num_tokens": 657660615.0, + "step": 25414 + }, + { + "epoch": 2.7910169119262025, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3356831073760986, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.710469126701355, + "num_tokens": 657688699.0, + "step": 25415 + }, + { + "epoch": 2.7911267296288162, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.238018751144409, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7363894581794739, + "num_tokens": 657717203.0, + "step": 25416 + }, + { + "epoch": 2.7912365473314296, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.383991003036499, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7267041206359863, + "num_tokens": 657743949.0, + "step": 25417 + }, + { + "epoch": 2.7913463650340438, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.0942280292510986, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7209764122962952, + "num_tokens": 657774713.0, + "step": 25418 + }, + { + "epoch": 2.791456182736657, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.380106210708618, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7538278698921204, + "num_tokens": 657799743.0, + "step": 25419 + }, + { + "epoch": 2.791566000439271, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.151278495788574, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7437820434570312, + "num_tokens": 657829342.0, + "step": 25420 + }, + { + "epoch": 2.7916758181418846, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3238205909729004, + "learning_rate": 1e-06, + "loss": 0.7648, + "mean_token_accuracy": 0.7586928606033325, + "num_tokens": 657854566.0, + "step": 25421 + }, + { + "epoch": 2.791785635844498, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5197434425354004, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7431926727294922, + "num_tokens": 657879366.0, + "step": 25422 + }, + { + "epoch": 2.7918954535471117, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3317458629608154, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7305160760879517, + "num_tokens": 657905414.0, + "step": 25423 + }, + { + "epoch": 2.7920052712497254, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3488478660583496, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7305822372436523, + "num_tokens": 657931013.0, + "step": 25424 + }, + { + "epoch": 2.792115088952339, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.247918128967285, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7073136568069458, + "num_tokens": 657959293.0, + "step": 25425 + }, + { + "epoch": 2.792224906654953, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.279309034347534, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7109614014625549, + "num_tokens": 657985708.0, + "step": 25426 + }, + { + "epoch": 2.7923347243575662, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3154122829437256, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7287366390228271, + "num_tokens": 658013079.0, + "step": 25427 + }, + { + "epoch": 2.79244454206018, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2103636264801025, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7186769247055054, + "num_tokens": 658042567.0, + "step": 25428 + }, + { + "epoch": 2.7925543597627938, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.163203001022339, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7335180640220642, + "num_tokens": 658072626.0, + "step": 25429 + }, + { + "epoch": 2.7926641774654075, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1203114986419678, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7145701050758362, + "num_tokens": 658102681.0, + "step": 25430 + }, + { + "epoch": 2.7927739951680213, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.494253635406494, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7509549856185913, + "num_tokens": 658125645.0, + "step": 25431 + }, + { + "epoch": 2.7928838128706346, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.541977643966675, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7373708486557007, + "num_tokens": 658148908.0, + "step": 25432 + }, + { + "epoch": 2.7929936305732483, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.482067108154297, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7211794853210449, + "num_tokens": 658171824.0, + "step": 25433 + }, + { + "epoch": 2.793103448275862, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4034481048583984, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7425479888916016, + "num_tokens": 658196709.0, + "step": 25434 + }, + { + "epoch": 2.793213265978476, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.8252298831939697, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7514713406562805, + "num_tokens": 658216649.0, + "step": 25435 + }, + { + "epoch": 2.7933230836810896, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3681821823120117, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7492109537124634, + "num_tokens": 658241009.0, + "step": 25436 + }, + { + "epoch": 2.793432901383703, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.399090528488159, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.721898078918457, + "num_tokens": 658267432.0, + "step": 25437 + }, + { + "epoch": 2.7935427190863167, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3166751861572266, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.729368269443512, + "num_tokens": 658294948.0, + "step": 25438 + }, + { + "epoch": 2.7936525367889304, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4359993934631348, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7468342781066895, + "num_tokens": 658319781.0, + "step": 25439 + }, + { + "epoch": 2.793762354491544, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2877745628356934, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7222926616668701, + "num_tokens": 658347955.0, + "step": 25440 + }, + { + "epoch": 2.793872172194158, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.6261916160583496, + "learning_rate": 1e-06, + "loss": 0.7638, + "mean_token_accuracy": 0.7572134733200073, + "num_tokens": 658370227.0, + "step": 25441 + }, + { + "epoch": 2.7939819898967713, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.251804828643799, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.732793927192688, + "num_tokens": 658399478.0, + "step": 25442 + }, + { + "epoch": 2.794091807599385, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.7619755268096924, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7351863980293274, + "num_tokens": 658419334.0, + "step": 25443 + }, + { + "epoch": 2.794201625301999, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.214974880218506, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7432816028594971, + "num_tokens": 658446821.0, + "step": 25444 + }, + { + "epoch": 2.794311443004612, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3228511810302734, + "learning_rate": 1e-06, + "loss": 0.7655, + "mean_token_accuracy": 0.7578932046890259, + "num_tokens": 658472890.0, + "step": 25445 + }, + { + "epoch": 2.794421260707226, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1610639095306396, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.725045919418335, + "num_tokens": 658501453.0, + "step": 25446 + }, + { + "epoch": 2.7945310784098396, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3847429752349854, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7492027282714844, + "num_tokens": 658525789.0, + "step": 25447 + }, + { + "epoch": 2.7946408961124534, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.603456497192383, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7328377962112427, + "num_tokens": 658547999.0, + "step": 25448 + }, + { + "epoch": 2.794750713815067, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.74396014213562, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.7534515857696533, + "num_tokens": 658567804.0, + "step": 25449 + }, + { + "epoch": 2.7948605315176804, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.260869264602661, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.724929690361023, + "num_tokens": 658595068.0, + "step": 25450 + }, + { + "epoch": 2.794970349220294, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5265417098999023, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7278692126274109, + "num_tokens": 658621032.0, + "step": 25451 + }, + { + "epoch": 2.795080166922908, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5464260578155518, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7375853657722473, + "num_tokens": 658641357.0, + "step": 25452 + }, + { + "epoch": 2.7951899846255217, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2956137657165527, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7096068263053894, + "num_tokens": 658671439.0, + "step": 25453 + }, + { + "epoch": 2.7952998023281355, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4037561416625977, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7318453788757324, + "num_tokens": 658696053.0, + "step": 25454 + }, + { + "epoch": 2.7954096200307488, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3509860038757324, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7446486353874207, + "num_tokens": 658721063.0, + "step": 25455 + }, + { + "epoch": 2.7955194377333625, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2930054664611816, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7496158480644226, + "num_tokens": 658746630.0, + "step": 25456 + }, + { + "epoch": 2.7956292554359763, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3035333156585693, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7403412461280823, + "num_tokens": 658771703.0, + "step": 25457 + }, + { + "epoch": 2.79573907313859, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.159501791000366, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.737526535987854, + "num_tokens": 658800801.0, + "step": 25458 + }, + { + "epoch": 2.795848890841204, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1265571117401123, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7213684916496277, + "num_tokens": 658832820.0, + "step": 25459 + }, + { + "epoch": 2.795958708543817, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.080300807952881, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.699790358543396, + "num_tokens": 658868258.0, + "step": 25460 + }, + { + "epoch": 2.796068526246431, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.481452465057373, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7303067445755005, + "num_tokens": 658891424.0, + "step": 25461 + }, + { + "epoch": 2.7961783439490446, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.224585771560669, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7053214311599731, + "num_tokens": 658919905.0, + "step": 25462 + }, + { + "epoch": 2.7962881616516584, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4620416164398193, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.729364275932312, + "num_tokens": 658945164.0, + "step": 25463 + }, + { + "epoch": 2.796397979354272, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4356801509857178, + "learning_rate": 1e-06, + "loss": 0.756, + "mean_token_accuracy": 0.7546815872192383, + "num_tokens": 658967595.0, + "step": 25464 + }, + { + "epoch": 2.7965077970568855, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.096580982208252, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7297959327697754, + "num_tokens": 658999509.0, + "step": 25465 + }, + { + "epoch": 2.796617614759499, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.267878293991089, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7255947589874268, + "num_tokens": 659028209.0, + "step": 25466 + }, + { + "epoch": 2.796727432462113, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.298022985458374, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7314680814743042, + "num_tokens": 659054733.0, + "step": 25467 + }, + { + "epoch": 2.7968372501647263, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.276790142059326, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6948399543762207, + "num_tokens": 659086566.0, + "step": 25468 + }, + { + "epoch": 2.7969470678673405, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3357622623443604, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7138864994049072, + "num_tokens": 659113076.0, + "step": 25469 + }, + { + "epoch": 2.797056885569954, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.266979455947876, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7438576817512512, + "num_tokens": 659140510.0, + "step": 25470 + }, + { + "epoch": 2.7971667032725676, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.190478563308716, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7293668985366821, + "num_tokens": 659171065.0, + "step": 25471 + }, + { + "epoch": 2.7972765209751813, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3250739574432373, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7134273052215576, + "num_tokens": 659198950.0, + "step": 25472 + }, + { + "epoch": 2.7973863386777946, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.1643831729888916, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7112822532653809, + "num_tokens": 659228794.0, + "step": 25473 + }, + { + "epoch": 2.7974961563804084, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.603403091430664, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7242552042007446, + "num_tokens": 659250785.0, + "step": 25474 + }, + { + "epoch": 2.797605974083022, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.542989730834961, + "learning_rate": 1e-06, + "loss": 0.7181, + "mean_token_accuracy": 0.7677811980247498, + "num_tokens": 659273650.0, + "step": 25475 + }, + { + "epoch": 2.797715791785636, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.365934133529663, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7513967752456665, + "num_tokens": 659297636.0, + "step": 25476 + }, + { + "epoch": 2.7978256094882497, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2866885662078857, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7288919687271118, + "num_tokens": 659325961.0, + "step": 25477 + }, + { + "epoch": 2.797935427190863, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.2091658115386963, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7121385931968689, + "num_tokens": 659355236.0, + "step": 25478 + }, + { + "epoch": 2.7980452448934767, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4130728244781494, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.74601149559021, + "num_tokens": 659380549.0, + "step": 25479 + }, + { + "epoch": 2.7981550625960905, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3827261924743652, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.751469612121582, + "num_tokens": 659404170.0, + "step": 25480 + }, + { + "epoch": 2.7982648802987042, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.3892197608947754, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7562688589096069, + "num_tokens": 659429358.0, + "step": 25481 + }, + { + "epoch": 2.798374698001318, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.499025821685791, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7356098890304565, + "num_tokens": 659452045.0, + "step": 25482 + }, + { + "epoch": 2.7984845157039313, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.082077980041504, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.723442554473877, + "num_tokens": 659482183.0, + "step": 25483 + }, + { + "epoch": 2.798594333406545, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.441084146499634, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7217262387275696, + "num_tokens": 659506878.0, + "step": 25484 + }, + { + "epoch": 2.798704151109159, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.298990488052368, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7270042896270752, + "num_tokens": 659534891.0, + "step": 25485 + }, + { + "epoch": 2.7988139688117726, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.296938180923462, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7371294498443604, + "num_tokens": 659563039.0, + "step": 25486 + }, + { + "epoch": 2.7989237865143863, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.432441473007202, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7182011604309082, + "num_tokens": 659587458.0, + "step": 25487 + }, + { + "epoch": 2.7990336042169996, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.4485878944396973, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7365384101867676, + "num_tokens": 659610684.0, + "step": 25488 + }, + { + "epoch": 2.7991434219196134, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.5593645572662354, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.7541043758392334, + "num_tokens": 659631616.0, + "step": 25489 + }, + { + "epoch": 2.799253239622227, + "ewc_loss": 0.00022411346435546875, + "grad_norm": 2.343508005142212, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7206753492355347, + "num_tokens": 659659811.0, + "step": 25490 + }, + { + "epoch": 2.799363057324841, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6099724769592285, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.7486037015914917, + "num_tokens": 659681866.0, + "step": 25491 + }, + { + "epoch": 2.7994728750274547, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2395145893096924, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.727748453617096, + "num_tokens": 659708529.0, + "step": 25492 + }, + { + "epoch": 2.799582692730068, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2311224937438965, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7341877818107605, + "num_tokens": 659737239.0, + "step": 25493 + }, + { + "epoch": 2.7996925104326817, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6335737705230713, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7393507361412048, + "num_tokens": 659758319.0, + "step": 25494 + }, + { + "epoch": 2.7998023281352955, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5370192527770996, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7277048826217651, + "num_tokens": 659781992.0, + "step": 25495 + }, + { + "epoch": 2.799912145837909, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.726799249649048, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7290127277374268, + "num_tokens": 659803204.0, + "step": 25496 + }, + { + "epoch": 2.8000219635405226, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.463629961013794, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.727593183517456, + "num_tokens": 659827306.0, + "step": 25497 + }, + { + "epoch": 2.8001317812431363, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6711695194244385, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7339773178100586, + "num_tokens": 659847819.0, + "step": 25498 + }, + { + "epoch": 2.80024159894575, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.510549306869507, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.723068356513977, + "num_tokens": 659871600.0, + "step": 25499 + }, + { + "epoch": 2.800351416648364, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1956207752227783, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7380356788635254, + "num_tokens": 659900805.0, + "step": 25500 + }, + { + "epoch": 2.800461234350977, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1820058822631836, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7234575748443604, + "num_tokens": 659929084.0, + "step": 25501 + }, + { + "epoch": 2.800571052053591, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1088783740997314, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7163317203521729, + "num_tokens": 659961327.0, + "step": 25502 + }, + { + "epoch": 2.8006808697562047, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4786903858184814, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7100980281829834, + "num_tokens": 659986817.0, + "step": 25503 + }, + { + "epoch": 2.8007906874588184, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7638227939605713, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.720525324344635, + "num_tokens": 660008085.0, + "step": 25504 + }, + { + "epoch": 2.800900505161432, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3424441814422607, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7447553277015686, + "num_tokens": 660034970.0, + "step": 25505 + }, + { + "epoch": 2.8010103228640455, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6451828479766846, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7423737049102783, + "num_tokens": 660056318.0, + "step": 25506 + }, + { + "epoch": 2.8011201405666593, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4696755409240723, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7490089535713196, + "num_tokens": 660079209.0, + "step": 25507 + }, + { + "epoch": 2.801229958269273, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3056015968322754, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7069936990737915, + "num_tokens": 660107530.0, + "step": 25508 + }, + { + "epoch": 2.8013397759718868, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.506403923034668, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.737747311592102, + "num_tokens": 660131087.0, + "step": 25509 + }, + { + "epoch": 2.8014495936745005, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1797847747802734, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7137256860733032, + "num_tokens": 660162306.0, + "step": 25510 + }, + { + "epoch": 2.801559411377114, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4177370071411133, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7350389957427979, + "num_tokens": 660186485.0, + "step": 25511 + }, + { + "epoch": 2.8016692290797276, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.328702211380005, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.725838303565979, + "num_tokens": 660210493.0, + "step": 25512 + }, + { + "epoch": 2.8017790467823414, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.446967124938965, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7154281139373779, + "num_tokens": 660233809.0, + "step": 25513 + }, + { + "epoch": 2.801888864484955, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2567124366760254, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7322297096252441, + "num_tokens": 660262261.0, + "step": 25514 + }, + { + "epoch": 2.801998682187569, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5506184101104736, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7419818639755249, + "num_tokens": 660285282.0, + "step": 25515 + }, + { + "epoch": 2.802108499890182, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4095821380615234, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7357675433158875, + "num_tokens": 660312026.0, + "step": 25516 + }, + { + "epoch": 2.802218317592796, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.521341323852539, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.725934624671936, + "num_tokens": 660336797.0, + "step": 25517 + }, + { + "epoch": 2.8023281352954097, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.350370407104492, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7337805032730103, + "num_tokens": 660363648.0, + "step": 25518 + }, + { + "epoch": 2.802437952998023, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.56072998046875, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7318757772445679, + "num_tokens": 660388352.0, + "step": 25519 + }, + { + "epoch": 2.802547770700637, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5097086429595947, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7292732000350952, + "num_tokens": 660412677.0, + "step": 25520 + }, + { + "epoch": 2.8026575884032505, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1832571029663086, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7038669586181641, + "num_tokens": 660442384.0, + "step": 25521 + }, + { + "epoch": 2.8027674061058643, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.144196033477783, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7452934980392456, + "num_tokens": 660471277.0, + "step": 25522 + }, + { + "epoch": 2.802877223808478, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2870728969573975, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7291017174720764, + "num_tokens": 660499776.0, + "step": 25523 + }, + { + "epoch": 2.8029870415110913, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.166138172149658, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7201000452041626, + "num_tokens": 660529542.0, + "step": 25524 + }, + { + "epoch": 2.803096859213705, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.318404197692871, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7242569923400879, + "num_tokens": 660557255.0, + "step": 25525 + }, + { + "epoch": 2.803206676916319, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.65559458732605, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7414069175720215, + "num_tokens": 660577448.0, + "step": 25526 + }, + { + "epoch": 2.8033164946189326, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0704281330108643, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7210423350334167, + "num_tokens": 660611802.0, + "step": 25527 + }, + { + "epoch": 2.8034263123215464, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.795485734939575, + "learning_rate": 1e-06, + "loss": 0.7435, + "mean_token_accuracy": 0.7605161070823669, + "num_tokens": 660632030.0, + "step": 25528 + }, + { + "epoch": 2.8035361300241597, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3899991512298584, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7166367769241333, + "num_tokens": 660656917.0, + "step": 25529 + }, + { + "epoch": 2.8036459477267734, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.690530776977539, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7518883943557739, + "num_tokens": 660677591.0, + "step": 25530 + }, + { + "epoch": 2.803755765429387, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3454763889312744, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7470420598983765, + "num_tokens": 660704066.0, + "step": 25531 + }, + { + "epoch": 2.803865583132001, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.052182912826538, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7179973125457764, + "num_tokens": 660736158.0, + "step": 25532 + }, + { + "epoch": 2.8039754008346147, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.222321033477783, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7392957210540771, + "num_tokens": 660764931.0, + "step": 25533 + }, + { + "epoch": 2.804085218537228, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3525233268737793, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7196556925773621, + "num_tokens": 660791878.0, + "step": 25534 + }, + { + "epoch": 2.804195036239842, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.351487874984741, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7401142120361328, + "num_tokens": 660818303.0, + "step": 25535 + }, + { + "epoch": 2.8043048539424555, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1319637298583984, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7257407903671265, + "num_tokens": 660847686.0, + "step": 25536 + }, + { + "epoch": 2.8044146716450693, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.272002935409546, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7292852401733398, + "num_tokens": 660874618.0, + "step": 25537 + }, + { + "epoch": 2.804524489347683, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.646054983139038, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7510687112808228, + "num_tokens": 660895920.0, + "step": 25538 + }, + { + "epoch": 2.8046343070502964, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.734956979751587, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7343682646751404, + "num_tokens": 660918930.0, + "step": 25539 + }, + { + "epoch": 2.80474412475291, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.251995801925659, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7528412342071533, + "num_tokens": 660946257.0, + "step": 25540 + }, + { + "epoch": 2.804853942455524, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4667694568634033, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7205771207809448, + "num_tokens": 660970851.0, + "step": 25541 + }, + { + "epoch": 2.8049637601581376, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1953930854797363, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6998106837272644, + "num_tokens": 661001842.0, + "step": 25542 + }, + { + "epoch": 2.8050735778607514, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.142693519592285, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7226318120956421, + "num_tokens": 661033093.0, + "step": 25543 + }, + { + "epoch": 2.8051833955633647, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.584195137023926, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.740035891532898, + "num_tokens": 661055272.0, + "step": 25544 + }, + { + "epoch": 2.8052932132659785, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.387383222579956, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7320103049278259, + "num_tokens": 661080432.0, + "step": 25545 + }, + { + "epoch": 2.8054030309685922, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4815316200256348, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7403950095176697, + "num_tokens": 661105508.0, + "step": 25546 + }, + { + "epoch": 2.8055128486712055, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2263171672821045, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7280324697494507, + "num_tokens": 661135946.0, + "step": 25547 + }, + { + "epoch": 2.8056226663738193, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5423521995544434, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7268866896629333, + "num_tokens": 661159454.0, + "step": 25548 + }, + { + "epoch": 2.805732484076433, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4685676097869873, + "learning_rate": 1e-06, + "loss": 0.7672, + "mean_token_accuracy": 0.7634664177894592, + "num_tokens": 661182410.0, + "step": 25549 + }, + { + "epoch": 2.805842301779047, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.9157497882843018, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7454395294189453, + "num_tokens": 661201229.0, + "step": 25550 + }, + { + "epoch": 2.8059521194816606, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.315303087234497, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7093987464904785, + "num_tokens": 661229079.0, + "step": 25551 + }, + { + "epoch": 2.806061937184274, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4875240325927734, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7340952157974243, + "num_tokens": 661252597.0, + "step": 25552 + }, + { + "epoch": 2.8061717548868876, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3290297985076904, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.70721435546875, + "num_tokens": 661280899.0, + "step": 25553 + }, + { + "epoch": 2.8062815725895014, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.322049140930176, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.730982780456543, + "num_tokens": 661309258.0, + "step": 25554 + }, + { + "epoch": 2.806391390292115, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4598748683929443, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7109885215759277, + "num_tokens": 661334630.0, + "step": 25555 + }, + { + "epoch": 2.806501207994729, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3242905139923096, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.736000657081604, + "num_tokens": 661361697.0, + "step": 25556 + }, + { + "epoch": 2.806611025697342, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.432119131088257, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7354974746704102, + "num_tokens": 661387248.0, + "step": 25557 + }, + { + "epoch": 2.806720843399956, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.421802520751953, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7258085608482361, + "num_tokens": 661412647.0, + "step": 25558 + }, + { + "epoch": 2.8068306611025697, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.313617706298828, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7278503775596619, + "num_tokens": 661438448.0, + "step": 25559 + }, + { + "epoch": 2.8069404788051835, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4637084007263184, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7338577508926392, + "num_tokens": 661463863.0, + "step": 25560 + }, + { + "epoch": 2.8070502965077972, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2164323329925537, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7239093780517578, + "num_tokens": 661492452.0, + "step": 25561 + }, + { + "epoch": 2.8071601142104106, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6522228717803955, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7395592927932739, + "num_tokens": 661514392.0, + "step": 25562 + }, + { + "epoch": 2.8072699319130243, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.206718921661377, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.741724967956543, + "num_tokens": 661540750.0, + "step": 25563 + }, + { + "epoch": 2.807379749615638, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4000625610351562, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7099087834358215, + "num_tokens": 661568038.0, + "step": 25564 + }, + { + "epoch": 2.807489567318252, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.263026475906372, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7352065443992615, + "num_tokens": 661594930.0, + "step": 25565 + }, + { + "epoch": 2.8075993850208656, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.209136486053467, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.718695878982544, + "num_tokens": 661623558.0, + "step": 25566 + }, + { + "epoch": 2.807709202723479, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2244904041290283, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7308511137962341, + "num_tokens": 661652206.0, + "step": 25567 + }, + { + "epoch": 2.8078190204260927, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1499404907226562, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7176638245582581, + "num_tokens": 661683089.0, + "step": 25568 + }, + { + "epoch": 2.8079288381287064, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3726303577423096, + "learning_rate": 1e-06, + "loss": 0.7812, + "mean_token_accuracy": 0.7528125643730164, + "num_tokens": 661707272.0, + "step": 25569 + }, + { + "epoch": 2.80803865583132, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4965453147888184, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7421459555625916, + "num_tokens": 661730598.0, + "step": 25570 + }, + { + "epoch": 2.808148473533934, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.546844959259033, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7452083826065063, + "num_tokens": 661753754.0, + "step": 25571 + }, + { + "epoch": 2.8082582912365472, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3910703659057617, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7360124588012695, + "num_tokens": 661779277.0, + "step": 25572 + }, + { + "epoch": 2.808368108939161, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.479959011077881, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.738277792930603, + "num_tokens": 661804152.0, + "step": 25573 + }, + { + "epoch": 2.8084779266417748, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6072206497192383, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7167507410049438, + "num_tokens": 661826690.0, + "step": 25574 + }, + { + "epoch": 2.808587744344388, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.247481107711792, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7164175510406494, + "num_tokens": 661858275.0, + "step": 25575 + }, + { + "epoch": 2.808697562047002, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.410080671310425, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7188669443130493, + "num_tokens": 661885693.0, + "step": 25576 + }, + { + "epoch": 2.8088073797496156, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.37941312789917, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7249202728271484, + "num_tokens": 661910365.0, + "step": 25577 + }, + { + "epoch": 2.8089171974522293, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6652719974517822, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7466021776199341, + "num_tokens": 661930340.0, + "step": 25578 + }, + { + "epoch": 2.809027015154843, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.704024076461792, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7335599064826965, + "num_tokens": 661950143.0, + "step": 25579 + }, + { + "epoch": 2.8091368328574564, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7118301391601562, + "learning_rate": 1e-06, + "loss": 0.8036, + "mean_token_accuracy": 0.7468996047973633, + "num_tokens": 661972186.0, + "step": 25580 + }, + { + "epoch": 2.80924665056007, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.259554624557495, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7368954420089722, + "num_tokens": 661998612.0, + "step": 25581 + }, + { + "epoch": 2.809356468262684, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3470795154571533, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7352474331855774, + "num_tokens": 662027212.0, + "step": 25582 + }, + { + "epoch": 2.8094662859652977, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4977269172668457, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7381294965744019, + "num_tokens": 662052504.0, + "step": 25583 + }, + { + "epoch": 2.8095761036679114, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4949042797088623, + "learning_rate": 1e-06, + "loss": 0.7812, + "mean_token_accuracy": 0.7492793202400208, + "num_tokens": 662075237.0, + "step": 25584 + }, + { + "epoch": 2.8096859213705248, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3818137645721436, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7247664928436279, + "num_tokens": 662100237.0, + "step": 25585 + }, + { + "epoch": 2.8097957390731385, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.31437349319458, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7585512399673462, + "num_tokens": 662125680.0, + "step": 25586 + }, + { + "epoch": 2.8099055567757523, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.729552745819092, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7266251444816589, + "num_tokens": 662146244.0, + "step": 25587 + }, + { + "epoch": 2.810015374478366, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1773836612701416, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7346821427345276, + "num_tokens": 662177330.0, + "step": 25588 + }, + { + "epoch": 2.81012519218098, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6834723949432373, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7526909112930298, + "num_tokens": 662197208.0, + "step": 25589 + }, + { + "epoch": 2.810235009883593, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.8133294582366943, + "learning_rate": 1e-06, + "loss": 0.7269, + "mean_token_accuracy": 0.7664252519607544, + "num_tokens": 662216227.0, + "step": 25590 + }, + { + "epoch": 2.810344827586207, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4411356449127197, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.723027229309082, + "num_tokens": 662241684.0, + "step": 25591 + }, + { + "epoch": 2.8104546452888206, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6844515800476074, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7392863035202026, + "num_tokens": 662262461.0, + "step": 25592 + }, + { + "epoch": 2.8105644629914344, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3940184116363525, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7387954592704773, + "num_tokens": 662287571.0, + "step": 25593 + }, + { + "epoch": 2.810674280694048, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1650922298431396, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7271444797515869, + "num_tokens": 662317521.0, + "step": 25594 + }, + { + "epoch": 2.8107840983966614, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.19311785697937, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7093943953514099, + "num_tokens": 662346423.0, + "step": 25595 + }, + { + "epoch": 2.810893916099275, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6917057037353516, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7247148752212524, + "num_tokens": 662367305.0, + "step": 25596 + }, + { + "epoch": 2.811003733801889, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3813188076019287, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7291531562805176, + "num_tokens": 662392026.0, + "step": 25597 + }, + { + "epoch": 2.8111135515045023, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5028092861175537, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7463717460632324, + "num_tokens": 662417973.0, + "step": 25598 + }, + { + "epoch": 2.8112233692071165, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.097503185272217, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7345194220542908, + "num_tokens": 662449961.0, + "step": 25599 + }, + { + "epoch": 2.8113331869097298, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.348508358001709, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7258922457695007, + "num_tokens": 662474543.0, + "step": 25600 + }, + { + "epoch": 2.8114430046123435, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.335374593734741, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7130364179611206, + "num_tokens": 662503803.0, + "step": 25601 + }, + { + "epoch": 2.8115528223149573, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0285236835479736, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7331082820892334, + "num_tokens": 662538210.0, + "step": 25602 + }, + { + "epoch": 2.8116626400175706, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.627751111984253, + "learning_rate": 1e-06, + "loss": 0.7563, + "mean_token_accuracy": 0.758065938949585, + "num_tokens": 662559687.0, + "step": 25603 + }, + { + "epoch": 2.8117724577201844, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7001659870147705, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.749508798122406, + "num_tokens": 662583448.0, + "step": 25604 + }, + { + "epoch": 2.811882275422798, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.388425350189209, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7250427603721619, + "num_tokens": 662608705.0, + "step": 25605 + }, + { + "epoch": 2.811992093125412, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.644040822982788, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7087999582290649, + "num_tokens": 662630715.0, + "step": 25606 + }, + { + "epoch": 2.8121019108280256, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.305785655975342, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7203031778335571, + "num_tokens": 662657962.0, + "step": 25607 + }, + { + "epoch": 2.812211728530639, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4161789417266846, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7348752021789551, + "num_tokens": 662684548.0, + "step": 25608 + }, + { + "epoch": 2.8123215462332527, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3663828372955322, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.744202196598053, + "num_tokens": 662710245.0, + "step": 25609 + }, + { + "epoch": 2.8124313639358665, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4595673084259033, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7371272444725037, + "num_tokens": 662735937.0, + "step": 25610 + }, + { + "epoch": 2.81254118163848, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7655904293060303, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7361143827438354, + "num_tokens": 662755845.0, + "step": 25611 + }, + { + "epoch": 2.812650999341094, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.507277011871338, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7244884371757507, + "num_tokens": 662781038.0, + "step": 25612 + }, + { + "epoch": 2.8127608170437073, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.497427463531494, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7308716773986816, + "num_tokens": 662805811.0, + "step": 25613 + }, + { + "epoch": 2.812870634746321, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6693010330200195, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7458469271659851, + "num_tokens": 662826943.0, + "step": 25614 + }, + { + "epoch": 2.812980452448935, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.197695255279541, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7242065668106079, + "num_tokens": 662854053.0, + "step": 25615 + }, + { + "epoch": 2.8130902701515486, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2841897010803223, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7172330617904663, + "num_tokens": 662882986.0, + "step": 25616 + }, + { + "epoch": 2.8132000878541623, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2499263286590576, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7324981093406677, + "num_tokens": 662913476.0, + "step": 25617 + }, + { + "epoch": 2.8133099055567756, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3597264289855957, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7396297454833984, + "num_tokens": 662940969.0, + "step": 25618 + }, + { + "epoch": 2.8134197232593894, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1861677169799805, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7480520009994507, + "num_tokens": 662969481.0, + "step": 25619 + }, + { + "epoch": 2.813529540962003, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.791189432144165, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7207746505737305, + "num_tokens": 662990900.0, + "step": 25620 + }, + { + "epoch": 2.813639358664617, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.11030650138855, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.6986139416694641, + "num_tokens": 663025348.0, + "step": 25621 + }, + { + "epoch": 2.8137491763672307, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1603708267211914, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7267683148384094, + "num_tokens": 663055529.0, + "step": 25622 + }, + { + "epoch": 2.813858994069844, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3008711338043213, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7424612045288086, + "num_tokens": 663082171.0, + "step": 25623 + }, + { + "epoch": 2.8139688117724577, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.68931245803833, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7424262762069702, + "num_tokens": 663102817.0, + "step": 25624 + }, + { + "epoch": 2.8140786294750715, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3915016651153564, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7226080894470215, + "num_tokens": 663126762.0, + "step": 25625 + }, + { + "epoch": 2.814188447177685, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.581413507461548, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7349013686180115, + "num_tokens": 663149265.0, + "step": 25626 + }, + { + "epoch": 2.8142982648802986, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.667708396911621, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7472106218338013, + "num_tokens": 663171640.0, + "step": 25627 + }, + { + "epoch": 2.8144080825829123, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0172977447509766, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7408071756362915, + "num_tokens": 663205859.0, + "step": 25628 + }, + { + "epoch": 2.814517900285526, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.471264600753784, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.7516062259674072, + "num_tokens": 663229362.0, + "step": 25629 + }, + { + "epoch": 2.81462771798814, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4861178398132324, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7498169541358948, + "num_tokens": 663251463.0, + "step": 25630 + }, + { + "epoch": 2.814737535690753, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.475888252258301, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7257755994796753, + "num_tokens": 663275502.0, + "step": 25631 + }, + { + "epoch": 2.814847353393367, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.144439935684204, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7254894971847534, + "num_tokens": 663305986.0, + "step": 25632 + }, + { + "epoch": 2.8149571710959806, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3864641189575195, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7531408667564392, + "num_tokens": 663330428.0, + "step": 25633 + }, + { + "epoch": 2.8150669887985944, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.403628349304199, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7301937341690063, + "num_tokens": 663356398.0, + "step": 25634 + }, + { + "epoch": 2.815176806501208, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7470290660858154, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7309050559997559, + "num_tokens": 663376647.0, + "step": 25635 + }, + { + "epoch": 2.8152866242038215, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5674402713775635, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7385314106941223, + "num_tokens": 663399079.0, + "step": 25636 + }, + { + "epoch": 2.8153964419064352, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.357304573059082, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7417868971824646, + "num_tokens": 663424783.0, + "step": 25637 + }, + { + "epoch": 2.815506259609049, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.084132432937622, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7208194732666016, + "num_tokens": 663456579.0, + "step": 25638 + }, + { + "epoch": 2.8156160773116627, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.411344051361084, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.720906674861908, + "num_tokens": 663481581.0, + "step": 25639 + }, + { + "epoch": 2.8157258950142765, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4829297065734863, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7350687384605408, + "num_tokens": 663504593.0, + "step": 25640 + }, + { + "epoch": 2.81583571271689, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.335815668106079, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.739414632320404, + "num_tokens": 663529251.0, + "step": 25641 + }, + { + "epoch": 2.8159455304195036, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.419644832611084, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7514728903770447, + "num_tokens": 663553956.0, + "step": 25642 + }, + { + "epoch": 2.8160553481221173, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.536165475845337, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7385663986206055, + "num_tokens": 663575866.0, + "step": 25643 + }, + { + "epoch": 2.816165165824731, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0808358192443848, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7224867939949036, + "num_tokens": 663609903.0, + "step": 25644 + }, + { + "epoch": 2.816274983527345, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3144757747650146, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7282670736312866, + "num_tokens": 663637331.0, + "step": 25645 + }, + { + "epoch": 2.816384801229958, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3109073638916016, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7380577921867371, + "num_tokens": 663667723.0, + "step": 25646 + }, + { + "epoch": 2.816494618932572, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.437528371810913, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7216728925704956, + "num_tokens": 663691961.0, + "step": 25647 + }, + { + "epoch": 2.8166044366351857, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.808117389678955, + "learning_rate": 1e-06, + "loss": 0.7755, + "mean_token_accuracy": 0.7569916248321533, + "num_tokens": 663712363.0, + "step": 25648 + }, + { + "epoch": 2.816714254337799, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0837206840515137, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7164474129676819, + "num_tokens": 663744232.0, + "step": 25649 + }, + { + "epoch": 2.816824072040413, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.272266387939453, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7283341288566589, + "num_tokens": 663771402.0, + "step": 25650 + }, + { + "epoch": 2.8169338897430265, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.263000726699829, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7301446199417114, + "num_tokens": 663798841.0, + "step": 25651 + }, + { + "epoch": 2.8170437074456403, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2675576210021973, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7242388725280762, + "num_tokens": 663825270.0, + "step": 25652 + }, + { + "epoch": 2.817153525148254, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0253024101257324, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7158563137054443, + "num_tokens": 663856581.0, + "step": 25653 + }, + { + "epoch": 2.8172633428508673, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2747325897216797, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7547454833984375, + "num_tokens": 663882578.0, + "step": 25654 + }, + { + "epoch": 2.817373160553481, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3106961250305176, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.725445568561554, + "num_tokens": 663908507.0, + "step": 25655 + }, + { + "epoch": 2.817482978256095, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.329197645187378, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7419196367263794, + "num_tokens": 663934012.0, + "step": 25656 + }, + { + "epoch": 2.8175927959587086, + "ewc_loss": 0.000225067138671875, + "grad_norm": 1.9532314538955688, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7172572612762451, + "num_tokens": 663971194.0, + "step": 25657 + }, + { + "epoch": 2.8177026136613224, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3518590927124023, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.7432630062103271, + "num_tokens": 663996352.0, + "step": 25658 + }, + { + "epoch": 2.8178124313639357, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6695849895477295, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7227557301521301, + "num_tokens": 664016776.0, + "step": 25659 + }, + { + "epoch": 2.8179222490665494, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3439066410064697, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7557457089424133, + "num_tokens": 664042536.0, + "step": 25660 + }, + { + "epoch": 2.818032066769163, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.262791633605957, + "learning_rate": 1e-06, + "loss": 0.7851, + "mean_token_accuracy": 0.7552261352539062, + "num_tokens": 664070228.0, + "step": 25661 + }, + { + "epoch": 2.818141884471777, + "ewc_loss": 0.00022602081298828125, + "grad_norm": 2.51444411277771, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7443997263908386, + "num_tokens": 664093475.0, + "step": 25662 + }, + { + "epoch": 2.8182517021743907, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6616461277008057, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7318035364151001, + "num_tokens": 664114673.0, + "step": 25663 + }, + { + "epoch": 2.818361519877004, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4815399646759033, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7320548892021179, + "num_tokens": 664139000.0, + "step": 25664 + }, + { + "epoch": 2.8184713375796178, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.236818552017212, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7188963890075684, + "num_tokens": 664169452.0, + "step": 25665 + }, + { + "epoch": 2.8185811552822315, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6060726642608643, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.739153265953064, + "num_tokens": 664190664.0, + "step": 25666 + }, + { + "epoch": 2.8186909729848453, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.457871437072754, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7291172742843628, + "num_tokens": 664214414.0, + "step": 25667 + }, + { + "epoch": 2.818800790687459, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.231678009033203, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7299433946609497, + "num_tokens": 664243671.0, + "step": 25668 + }, + { + "epoch": 2.8189106083900723, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2631514072418213, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7193595170974731, + "num_tokens": 664272682.0, + "step": 25669 + }, + { + "epoch": 2.819020426092686, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4178829193115234, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7316046953201294, + "num_tokens": 664297167.0, + "step": 25670 + }, + { + "epoch": 2.8191302437953, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.176372766494751, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7180305123329163, + "num_tokens": 664327436.0, + "step": 25671 + }, + { + "epoch": 2.8192400614979136, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.300650119781494, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7287476658821106, + "num_tokens": 664355794.0, + "step": 25672 + }, + { + "epoch": 2.8193498792005274, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.260225296020508, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7070091366767883, + "num_tokens": 664385123.0, + "step": 25673 + }, + { + "epoch": 2.8194596969031407, + "ewc_loss": 0.000225067138671875, + "grad_norm": 3.3319759368896484, + "learning_rate": 1e-06, + "loss": 0.7223, + "mean_token_accuracy": 0.7669956684112549, + "num_tokens": 664399166.0, + "step": 25674 + }, + { + "epoch": 2.8195695146057544, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.261150598526001, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7201843857765198, + "num_tokens": 664426734.0, + "step": 25675 + }, + { + "epoch": 2.819679332308368, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4520761966705322, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7494868040084839, + "num_tokens": 664449383.0, + "step": 25676 + }, + { + "epoch": 2.8197891500109815, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1984734535217285, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7362381815910339, + "num_tokens": 664478943.0, + "step": 25677 + }, + { + "epoch": 2.8198989677135953, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4981813430786133, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7234445214271545, + "num_tokens": 664504458.0, + "step": 25678 + }, + { + "epoch": 2.820008785416209, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0791149139404297, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7182289958000183, + "num_tokens": 664535890.0, + "step": 25679 + }, + { + "epoch": 2.820118603118823, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5726866722106934, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7375538349151611, + "num_tokens": 664558345.0, + "step": 25680 + }, + { + "epoch": 2.8202284208214365, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5079431533813477, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7347725033760071, + "num_tokens": 664580299.0, + "step": 25681 + }, + { + "epoch": 2.82033823852405, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.831056594848633, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7383217811584473, + "num_tokens": 664599077.0, + "step": 25682 + }, + { + "epoch": 2.8204480562266636, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4837756156921387, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7311899662017822, + "num_tokens": 664624121.0, + "step": 25683 + }, + { + "epoch": 2.8205578739292774, + "ewc_loss": 0.000225067138671875, + "grad_norm": 1.9664143323898315, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7304636836051941, + "num_tokens": 664660951.0, + "step": 25684 + }, + { + "epoch": 2.820667691631891, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.522118091583252, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7547412514686584, + "num_tokens": 664684410.0, + "step": 25685 + }, + { + "epoch": 2.820777509334505, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5453784465789795, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7322412729263306, + "num_tokens": 664706851.0, + "step": 25686 + }, + { + "epoch": 2.820887327037118, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4267048835754395, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7265880703926086, + "num_tokens": 664731863.0, + "step": 25687 + }, + { + "epoch": 2.820997144739732, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.73755145072937, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7381935119628906, + "num_tokens": 664752228.0, + "step": 25688 + }, + { + "epoch": 2.8211069624423457, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5288498401641846, + "learning_rate": 1e-06, + "loss": 0.7303, + "mean_token_accuracy": 0.7611498832702637, + "num_tokens": 664773933.0, + "step": 25689 + }, + { + "epoch": 2.8212167801449595, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2052066326141357, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7285732626914978, + "num_tokens": 664802609.0, + "step": 25690 + }, + { + "epoch": 2.8213265978475732, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.8960440158843994, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7292636632919312, + "num_tokens": 664821200.0, + "step": 25691 + }, + { + "epoch": 2.8214364155501865, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.353895425796509, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7183786630630493, + "num_tokens": 664848426.0, + "step": 25692 + }, + { + "epoch": 2.8215462332528003, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.212601900100708, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7325645685195923, + "num_tokens": 664874057.0, + "step": 25693 + }, + { + "epoch": 2.821656050955414, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6262266635894775, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7346910238265991, + "num_tokens": 664897031.0, + "step": 25694 + }, + { + "epoch": 2.821765868658028, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4219231605529785, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7005599141120911, + "num_tokens": 664922382.0, + "step": 25695 + }, + { + "epoch": 2.8218756863606416, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.402930736541748, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7314873337745667, + "num_tokens": 664946332.0, + "step": 25696 + }, + { + "epoch": 2.821985504063255, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4157655239105225, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7313488125801086, + "num_tokens": 664970614.0, + "step": 25697 + }, + { + "epoch": 2.8220953217658686, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.288698196411133, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7259138822555542, + "num_tokens": 664997642.0, + "step": 25698 + }, + { + "epoch": 2.8222051394684824, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.179590940475464, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7435715198516846, + "num_tokens": 665025423.0, + "step": 25699 + }, + { + "epoch": 2.8223149571710957, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.278184175491333, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7247967720031738, + "num_tokens": 665053357.0, + "step": 25700 + }, + { + "epoch": 2.82242477487371, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.498429298400879, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7353554368019104, + "num_tokens": 665078191.0, + "step": 25701 + }, + { + "epoch": 2.822534592576323, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5896592140197754, + "learning_rate": 1e-06, + "loss": 0.7623, + "mean_token_accuracy": 0.7564437389373779, + "num_tokens": 665100123.0, + "step": 25702 + }, + { + "epoch": 2.822644410278937, + "ewc_loss": 0.00022602081298828125, + "grad_norm": 2.1711385250091553, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6962286829948425, + "num_tokens": 665130091.0, + "step": 25703 + }, + { + "epoch": 2.8227542279815507, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.9605307579040527, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7371507883071899, + "num_tokens": 665148247.0, + "step": 25704 + }, + { + "epoch": 2.822864045684164, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.0789647102355957, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7112412452697754, + "num_tokens": 665178671.0, + "step": 25705 + }, + { + "epoch": 2.822973863386778, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1922359466552734, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7329039573669434, + "num_tokens": 665206942.0, + "step": 25706 + }, + { + "epoch": 2.8230836810893916, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.773505449295044, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7270605564117432, + "num_tokens": 665228016.0, + "step": 25707 + }, + { + "epoch": 2.8231934987920053, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.445875406265259, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7255697250366211, + "num_tokens": 665252564.0, + "step": 25708 + }, + { + "epoch": 2.823303316494619, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.392932176589966, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7152149081230164, + "num_tokens": 665279318.0, + "step": 25709 + }, + { + "epoch": 2.8234131341972324, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2606287002563477, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.741500735282898, + "num_tokens": 665305386.0, + "step": 25710 + }, + { + "epoch": 2.823522951899846, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6491799354553223, + "learning_rate": 1e-06, + "loss": 0.7074, + "mean_token_accuracy": 0.7722804546356201, + "num_tokens": 665327383.0, + "step": 25711 + }, + { + "epoch": 2.82363276960246, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1639554500579834, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7158771753311157, + "num_tokens": 665356505.0, + "step": 25712 + }, + { + "epoch": 2.8237425873050737, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4119131565093994, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7158857583999634, + "num_tokens": 665382325.0, + "step": 25713 + }, + { + "epoch": 2.8238524050076874, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2460999488830566, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7424972653388977, + "num_tokens": 665408629.0, + "step": 25714 + }, + { + "epoch": 2.8239622227103007, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.29251766204834, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7230091094970703, + "num_tokens": 665436323.0, + "step": 25715 + }, + { + "epoch": 2.8240720404129145, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2323014736175537, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.7712880969047546, + "num_tokens": 665463777.0, + "step": 25716 + }, + { + "epoch": 2.8241818581155282, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.6301026344299316, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.734933078289032, + "num_tokens": 665485836.0, + "step": 25717 + }, + { + "epoch": 2.824291675818142, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.431959390640259, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7233167886734009, + "num_tokens": 665510241.0, + "step": 25718 + }, + { + "epoch": 2.8244014935207558, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2649710178375244, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7113595008850098, + "num_tokens": 665539111.0, + "step": 25719 + }, + { + "epoch": 2.824511311223369, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1501851081848145, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7102857232093811, + "num_tokens": 665572159.0, + "step": 25720 + }, + { + "epoch": 2.824621128925983, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.2776286602020264, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7179570198059082, + "num_tokens": 665600417.0, + "step": 25721 + }, + { + "epoch": 2.8247309466285966, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.5044920444488525, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.723478376865387, + "num_tokens": 665625111.0, + "step": 25722 + }, + { + "epoch": 2.8248407643312103, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.9527225494384766, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7347654104232788, + "num_tokens": 665645082.0, + "step": 25723 + }, + { + "epoch": 2.824950582033824, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1517364978790283, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7167435884475708, + "num_tokens": 665675922.0, + "step": 25724 + }, + { + "epoch": 2.8250603997364374, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.388685941696167, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7445791959762573, + "num_tokens": 665701672.0, + "step": 25725 + }, + { + "epoch": 2.825170217439051, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.425767183303833, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7406643033027649, + "num_tokens": 665726620.0, + "step": 25726 + }, + { + "epoch": 2.825280035141665, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3432776927948, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.711463451385498, + "num_tokens": 665752886.0, + "step": 25727 + }, + { + "epoch": 2.8253898528442782, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.469533681869507, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7255123257637024, + "num_tokens": 665776049.0, + "step": 25728 + }, + { + "epoch": 2.825499670546892, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3843014240264893, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7348683476448059, + "num_tokens": 665801166.0, + "step": 25729 + }, + { + "epoch": 2.8256094882495058, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.235321521759033, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7096209526062012, + "num_tokens": 665829909.0, + "step": 25730 + }, + { + "epoch": 2.8257193059521195, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.7140238285064697, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7369133830070496, + "num_tokens": 665851708.0, + "step": 25731 + }, + { + "epoch": 2.8258291236547333, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4366142749786377, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7248413562774658, + "num_tokens": 665875189.0, + "step": 25732 + }, + { + "epoch": 2.8259389413573466, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4572231769561768, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7236552834510803, + "num_tokens": 665897904.0, + "step": 25733 + }, + { + "epoch": 2.8260487590599603, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4916625022888184, + "learning_rate": 1e-06, + "loss": 0.769, + "mean_token_accuracy": 0.7508759498596191, + "num_tokens": 665921091.0, + "step": 25734 + }, + { + "epoch": 2.826158576762574, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1856114864349365, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7467435002326965, + "num_tokens": 665948538.0, + "step": 25735 + }, + { + "epoch": 2.826268394465188, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.564840793609619, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7200031280517578, + "num_tokens": 665974966.0, + "step": 25736 + }, + { + "epoch": 2.8263782121678016, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.778961181640625, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7383033633232117, + "num_tokens": 665994760.0, + "step": 25737 + }, + { + "epoch": 2.826488029870415, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.315964937210083, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7268819808959961, + "num_tokens": 666021963.0, + "step": 25738 + }, + { + "epoch": 2.8265978475730287, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.138622283935547, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7324783205986023, + "num_tokens": 666050709.0, + "step": 25739 + }, + { + "epoch": 2.8267076652756424, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1141462326049805, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7180690169334412, + "num_tokens": 666080470.0, + "step": 25740 + }, + { + "epoch": 2.826817482978256, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.3314452171325684, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.722023069858551, + "num_tokens": 666105655.0, + "step": 25741 + }, + { + "epoch": 2.82692730068087, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.559542655944824, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7331937551498413, + "num_tokens": 666129224.0, + "step": 25742 + }, + { + "epoch": 2.8270371183834833, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.346890449523926, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7328909039497375, + "num_tokens": 666153247.0, + "step": 25743 + }, + { + "epoch": 2.827146936086097, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.639451026916504, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.730060338973999, + "num_tokens": 666175056.0, + "step": 25744 + }, + { + "epoch": 2.8272567537887108, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.1707355976104736, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.722861647605896, + "num_tokens": 666207477.0, + "step": 25745 + }, + { + "epoch": 2.8273665714913245, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.501538038253784, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7237491011619568, + "num_tokens": 666233368.0, + "step": 25746 + }, + { + "epoch": 2.8274763891939383, + "ewc_loss": 0.000225067138671875, + "grad_norm": 2.4579966068267822, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7295576333999634, + "num_tokens": 666258159.0, + "step": 25747 + }, + { + "epoch": 2.8275862068965516, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4157090187072754, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7251564264297485, + "num_tokens": 666284219.0, + "step": 25748 + }, + { + "epoch": 2.8276960245991654, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6421773433685303, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7364304065704346, + "num_tokens": 666305589.0, + "step": 25749 + }, + { + "epoch": 2.827805842301779, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.492133855819702, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7149158120155334, + "num_tokens": 666330272.0, + "step": 25750 + }, + { + "epoch": 2.827915660004393, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.37164044380188, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7338168621063232, + "num_tokens": 666356032.0, + "step": 25751 + }, + { + "epoch": 2.8280254777070066, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 3.1096243858337402, + "learning_rate": 1e-06, + "loss": 0.7695, + "mean_token_accuracy": 0.7524960041046143, + "num_tokens": 666372562.0, + "step": 25752 + }, + { + "epoch": 2.82813529540962, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.719099998474121, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7343807220458984, + "num_tokens": 666395419.0, + "step": 25753 + }, + { + "epoch": 2.8282451131122337, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2863423824310303, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7636904716491699, + "num_tokens": 666420506.0, + "step": 25754 + }, + { + "epoch": 2.8283549308148475, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.19686222076416, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7094703912734985, + "num_tokens": 666450363.0, + "step": 25755 + }, + { + "epoch": 2.8284647485174608, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.546321392059326, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7410070896148682, + "num_tokens": 666472731.0, + "step": 25756 + }, + { + "epoch": 2.8285745662200745, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.422914505004883, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7231879234313965, + "num_tokens": 666498792.0, + "step": 25757 + }, + { + "epoch": 2.8286843839226883, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3014819622039795, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7166226506233215, + "num_tokens": 666526752.0, + "step": 25758 + }, + { + "epoch": 2.828794201625302, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5796337127685547, + "learning_rate": 1e-06, + "loss": 0.7677, + "mean_token_accuracy": 0.7583819031715393, + "num_tokens": 666548945.0, + "step": 25759 + }, + { + "epoch": 2.828904019327916, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6529366970062256, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7211098670959473, + "num_tokens": 666569835.0, + "step": 25760 + }, + { + "epoch": 2.829013837030529, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.510913372039795, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7326493263244629, + "num_tokens": 666593958.0, + "step": 25761 + }, + { + "epoch": 2.829123654733143, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.48923397064209, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7241902351379395, + "num_tokens": 666618422.0, + "step": 25762 + }, + { + "epoch": 2.8292334724357566, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6932718753814697, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.7513524293899536, + "num_tokens": 666639019.0, + "step": 25763 + }, + { + "epoch": 2.8293432901383704, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.513016700744629, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7302912473678589, + "num_tokens": 666662132.0, + "step": 25764 + }, + { + "epoch": 2.829453107840984, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.8197834491729736, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7425978183746338, + "num_tokens": 666680745.0, + "step": 25765 + }, + { + "epoch": 2.8295629255435975, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2624168395996094, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7462265491485596, + "num_tokens": 666708013.0, + "step": 25766 + }, + { + "epoch": 2.829672743246211, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2522644996643066, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.6933917999267578, + "num_tokens": 666738869.0, + "step": 25767 + }, + { + "epoch": 2.829782560948825, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.328003168106079, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7348290085792542, + "num_tokens": 666764646.0, + "step": 25768 + }, + { + "epoch": 2.8298923786514387, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.310215950012207, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7129043340682983, + "num_tokens": 666792890.0, + "step": 25769 + }, + { + "epoch": 2.8300021963540525, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.371551275253296, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7422078847885132, + "num_tokens": 666817196.0, + "step": 25770 + }, + { + "epoch": 2.830112014056666, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6695990562438965, + "learning_rate": 1e-06, + "loss": 0.7245, + "mean_token_accuracy": 0.7695344686508179, + "num_tokens": 666837965.0, + "step": 25771 + }, + { + "epoch": 2.8302218317592795, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.471017599105835, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7242839336395264, + "num_tokens": 666861412.0, + "step": 25772 + }, + { + "epoch": 2.8303316494618933, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6829276084899902, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7246369123458862, + "num_tokens": 666883947.0, + "step": 25773 + }, + { + "epoch": 2.830441467164507, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.73521089553833, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7509315013885498, + "num_tokens": 666904341.0, + "step": 25774 + }, + { + "epoch": 2.830551284867121, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4284751415252686, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7372419238090515, + "num_tokens": 666926188.0, + "step": 25775 + }, + { + "epoch": 2.830661102569734, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5087339878082275, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7319227457046509, + "num_tokens": 666950234.0, + "step": 25776 + }, + { + "epoch": 2.830770920272348, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3357298374176025, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7237561345100403, + "num_tokens": 666976412.0, + "step": 25777 + }, + { + "epoch": 2.8308807379749616, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.211724042892456, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7041804790496826, + "num_tokens": 667005457.0, + "step": 25778 + }, + { + "epoch": 2.830990555677575, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.610318899154663, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7384555339813232, + "num_tokens": 667028380.0, + "step": 25779 + }, + { + "epoch": 2.831100373380189, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.046755075454712, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7106183171272278, + "num_tokens": 667059785.0, + "step": 25780 + }, + { + "epoch": 2.8312101910828025, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3488001823425293, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7414186596870422, + "num_tokens": 667083512.0, + "step": 25781 + }, + { + "epoch": 2.8313200087854162, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5548455715179443, + "learning_rate": 1e-06, + "loss": 0.7194, + "mean_token_accuracy": 0.7622871994972229, + "num_tokens": 667104526.0, + "step": 25782 + }, + { + "epoch": 2.83142982648803, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7394299507141113, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7199851274490356, + "num_tokens": 667125859.0, + "step": 25783 + }, + { + "epoch": 2.8315396441906433, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3872036933898926, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7335743308067322, + "num_tokens": 667151027.0, + "step": 25784 + }, + { + "epoch": 2.831649461893257, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 1.9235730171203613, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7480106353759766, + "num_tokens": 667186120.0, + "step": 25785 + }, + { + "epoch": 2.831759279595871, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7146718502044678, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7241579294204712, + "num_tokens": 667209178.0, + "step": 25786 + }, + { + "epoch": 2.8318690972984846, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5198822021484375, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7419599890708923, + "num_tokens": 667233824.0, + "step": 25787 + }, + { + "epoch": 2.8319789150010983, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6816160678863525, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7441476583480835, + "num_tokens": 667255284.0, + "step": 25788 + }, + { + "epoch": 2.8320887327037116, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.26591157913208, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.711888313293457, + "num_tokens": 667284434.0, + "step": 25789 + }, + { + "epoch": 2.8321985504063254, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4873766899108887, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7275106906890869, + "num_tokens": 667307963.0, + "step": 25790 + }, + { + "epoch": 2.832308368108939, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.685948133468628, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7328164577484131, + "num_tokens": 667330400.0, + "step": 25791 + }, + { + "epoch": 2.832418185811553, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.452765464782715, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7167142629623413, + "num_tokens": 667353480.0, + "step": 25792 + }, + { + "epoch": 2.8325280035141667, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.492048978805542, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7415092587471008, + "num_tokens": 667378113.0, + "step": 25793 + }, + { + "epoch": 2.83263782121678, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2203965187072754, + "learning_rate": 1e-06, + "loss": 0.7722, + "mean_token_accuracy": 0.7546395063400269, + "num_tokens": 667405448.0, + "step": 25794 + }, + { + "epoch": 2.8327476389193937, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3241145610809326, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.745794415473938, + "num_tokens": 667431979.0, + "step": 25795 + }, + { + "epoch": 2.8328574566220075, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.544104814529419, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7451931238174438, + "num_tokens": 667454702.0, + "step": 25796 + }, + { + "epoch": 2.8329672743246213, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.747774362564087, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7253031730651855, + "num_tokens": 667476501.0, + "step": 25797 + }, + { + "epoch": 2.833077092027235, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3637235164642334, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7273659706115723, + "num_tokens": 667502412.0, + "step": 25798 + }, + { + "epoch": 2.8331869097298483, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.783783197402954, + "learning_rate": 1e-06, + "loss": 0.7639, + "mean_token_accuracy": 0.7555861473083496, + "num_tokens": 667521563.0, + "step": 25799 + }, + { + "epoch": 2.833296727432462, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3552613258361816, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7300382852554321, + "num_tokens": 667548691.0, + "step": 25800 + }, + { + "epoch": 2.833406545135076, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.294433832168579, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7199972867965698, + "num_tokens": 667575700.0, + "step": 25801 + }, + { + "epoch": 2.8335163628376896, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4689016342163086, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7320916056632996, + "num_tokens": 667599917.0, + "step": 25802 + }, + { + "epoch": 2.8336261805403034, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.219871759414673, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.714708149433136, + "num_tokens": 667630484.0, + "step": 25803 + }, + { + "epoch": 2.8337359982429167, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.0552010536193848, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7062854766845703, + "num_tokens": 667662177.0, + "step": 25804 + }, + { + "epoch": 2.8338458159455304, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.403027057647705, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7145528197288513, + "num_tokens": 667687948.0, + "step": 25805 + }, + { + "epoch": 2.833955633648144, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4868147373199463, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7304412126541138, + "num_tokens": 667711620.0, + "step": 25806 + }, + { + "epoch": 2.8340654513507575, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.548198699951172, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7559152841567993, + "num_tokens": 667733656.0, + "step": 25807 + }, + { + "epoch": 2.8341752690533712, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.05975341796875, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7236741781234741, + "num_tokens": 667766584.0, + "step": 25808 + }, + { + "epoch": 2.834285086755985, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3146755695343018, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7139872908592224, + "num_tokens": 667795665.0, + "step": 25809 + }, + { + "epoch": 2.8343949044585988, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.523632049560547, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7339872121810913, + "num_tokens": 667820599.0, + "step": 25810 + }, + { + "epoch": 2.8345047221612125, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2322373390197754, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7367918491363525, + "num_tokens": 667849042.0, + "step": 25811 + }, + { + "epoch": 2.834614539863826, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2044119834899902, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7500604391098022, + "num_tokens": 667877786.0, + "step": 25812 + }, + { + "epoch": 2.8347243575664396, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7210259437561035, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7491104602813721, + "num_tokens": 667900474.0, + "step": 25813 + }, + { + "epoch": 2.8348341752690533, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.115072011947632, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7368960976600647, + "num_tokens": 667930778.0, + "step": 25814 + }, + { + "epoch": 2.834943992971667, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3361573219299316, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7517273426055908, + "num_tokens": 667956985.0, + "step": 25815 + }, + { + "epoch": 2.835053810674281, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6668715476989746, + "learning_rate": 1e-06, + "loss": 0.7634, + "mean_token_accuracy": 0.7592368125915527, + "num_tokens": 667976498.0, + "step": 25816 + }, + { + "epoch": 2.835163628376894, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5737407207489014, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7012841701507568, + "num_tokens": 667999562.0, + "step": 25817 + }, + { + "epoch": 2.835273446079508, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.441896677017212, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7279994487762451, + "num_tokens": 668024673.0, + "step": 25818 + }, + { + "epoch": 2.8353832637821217, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.759864091873169, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7489089369773865, + "num_tokens": 668044984.0, + "step": 25819 + }, + { + "epoch": 2.8354930814847354, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.937072515487671, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.748231053352356, + "num_tokens": 668062676.0, + "step": 25820 + }, + { + "epoch": 2.835602899187349, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.0448944568634033, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7143727540969849, + "num_tokens": 668095964.0, + "step": 25821 + }, + { + "epoch": 2.8357127168899625, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.40889048576355, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7289005517959595, + "num_tokens": 668120610.0, + "step": 25822 + }, + { + "epoch": 2.8358225345925763, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1725752353668213, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7070982456207275, + "num_tokens": 668154666.0, + "step": 25823 + }, + { + "epoch": 2.83593235229519, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5909087657928467, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7164075374603271, + "num_tokens": 668179455.0, + "step": 25824 + }, + { + "epoch": 2.836042169997804, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3887360095977783, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.744949996471405, + "num_tokens": 668204743.0, + "step": 25825 + }, + { + "epoch": 2.8361519877004175, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4077868461608887, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.730441153049469, + "num_tokens": 668229429.0, + "step": 25826 + }, + { + "epoch": 2.836261805403031, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4739744663238525, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7444783449172974, + "num_tokens": 668252642.0, + "step": 25827 + }, + { + "epoch": 2.8363716231056446, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4806647300720215, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7325930595397949, + "num_tokens": 668274927.0, + "step": 25828 + }, + { + "epoch": 2.8364814408082584, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.525251865386963, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7287701964378357, + "num_tokens": 668300529.0, + "step": 25829 + }, + { + "epoch": 2.8365912585108717, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5364813804626465, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7445789575576782, + "num_tokens": 668323105.0, + "step": 25830 + }, + { + "epoch": 2.836701076213486, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7721641063690186, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7347756624221802, + "num_tokens": 668343240.0, + "step": 25831 + }, + { + "epoch": 2.836810893916099, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4523909091949463, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7223246097564697, + "num_tokens": 668366235.0, + "step": 25832 + }, + { + "epoch": 2.836920711618713, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4526970386505127, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.710949182510376, + "num_tokens": 668390790.0, + "step": 25833 + }, + { + "epoch": 2.8370305293213267, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3743669986724854, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7278640866279602, + "num_tokens": 668417367.0, + "step": 25834 + }, + { + "epoch": 2.83714034702394, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.953587770462036, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7406826615333557, + "num_tokens": 668436434.0, + "step": 25835 + }, + { + "epoch": 2.837250164726554, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3933639526367188, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7445359230041504, + "num_tokens": 668462158.0, + "step": 25836 + }, + { + "epoch": 2.8373599824291675, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5142533779144287, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7261244058609009, + "num_tokens": 668484016.0, + "step": 25837 + }, + { + "epoch": 2.8374698001317813, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.418823003768921, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7300430536270142, + "num_tokens": 668509697.0, + "step": 25838 + }, + { + "epoch": 2.837579617834395, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.505167007446289, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7314060926437378, + "num_tokens": 668534839.0, + "step": 25839 + }, + { + "epoch": 2.8376894355370084, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5581650733947754, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7290598154067993, + "num_tokens": 668558902.0, + "step": 25840 + }, + { + "epoch": 2.837799253239622, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2602591514587402, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7157772183418274, + "num_tokens": 668589644.0, + "step": 25841 + }, + { + "epoch": 2.837909070942236, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.0511834621429443, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7183555364608765, + "num_tokens": 668623296.0, + "step": 25842 + }, + { + "epoch": 2.8380188886448496, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1231346130371094, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7136052846908569, + "num_tokens": 668652015.0, + "step": 25843 + }, + { + "epoch": 2.8381287063474634, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.250748634338379, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7265042066574097, + "num_tokens": 668681762.0, + "step": 25844 + }, + { + "epoch": 2.8382385240500767, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2521727085113525, + "learning_rate": 1e-06, + "loss": 0.7676, + "mean_token_accuracy": 0.7522991299629211, + "num_tokens": 668706177.0, + "step": 25845 + }, + { + "epoch": 2.8383483417526905, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.818460464477539, + "learning_rate": 1e-06, + "loss": 0.7429, + "mean_token_accuracy": 0.7606003880500793, + "num_tokens": 668724696.0, + "step": 25846 + }, + { + "epoch": 2.838458159455304, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4063560962677, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7418636083602905, + "num_tokens": 668748726.0, + "step": 25847 + }, + { + "epoch": 2.838567977157918, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5642447471618652, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7472564578056335, + "num_tokens": 668771087.0, + "step": 25848 + }, + { + "epoch": 2.8386777948605317, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4888973236083984, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7218414545059204, + "num_tokens": 668796119.0, + "step": 25849 + }, + { + "epoch": 2.838787612563145, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3899338245391846, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7314456105232239, + "num_tokens": 668821536.0, + "step": 25850 + }, + { + "epoch": 2.838897430265759, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3500070571899414, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7307777404785156, + "num_tokens": 668849118.0, + "step": 25851 + }, + { + "epoch": 2.8390072479683726, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.306644916534424, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7100611925125122, + "num_tokens": 668878656.0, + "step": 25852 + }, + { + "epoch": 2.8391170656709863, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7029616832733154, + "learning_rate": 1e-06, + "loss": 0.7365, + "mean_token_accuracy": 0.7653388381004333, + "num_tokens": 668897345.0, + "step": 25853 + }, + { + "epoch": 2.8392268833736, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.486982583999634, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7266857624053955, + "num_tokens": 668921684.0, + "step": 25854 + }, + { + "epoch": 2.8393367010762134, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2277514934539795, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7294647693634033, + "num_tokens": 668948081.0, + "step": 25855 + }, + { + "epoch": 2.839446518778827, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.136019468307495, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7029115557670593, + "num_tokens": 668978585.0, + "step": 25856 + }, + { + "epoch": 2.839556336481441, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1195499897003174, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7109013795852661, + "num_tokens": 669009970.0, + "step": 25857 + }, + { + "epoch": 2.839666154184054, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3987276554107666, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7224405407905579, + "num_tokens": 669036112.0, + "step": 25858 + }, + { + "epoch": 2.839775971886668, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3359851837158203, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7388456463813782, + "num_tokens": 669062067.0, + "step": 25859 + }, + { + "epoch": 2.8398857895892817, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.31945538520813, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7463504672050476, + "num_tokens": 669089372.0, + "step": 25860 + }, + { + "epoch": 2.8399956072918955, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3035614490509033, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7365936040878296, + "num_tokens": 669115387.0, + "step": 25861 + }, + { + "epoch": 2.8401054249945092, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.450272560119629, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7217598557472229, + "num_tokens": 669139887.0, + "step": 25862 + }, + { + "epoch": 2.8402152426971226, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2773807048797607, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7092430591583252, + "num_tokens": 669169332.0, + "step": 25863 + }, + { + "epoch": 2.8403250603997363, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.657421588897705, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7253795266151428, + "num_tokens": 669193611.0, + "step": 25864 + }, + { + "epoch": 2.84043487810235, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.45218825340271, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7503294944763184, + "num_tokens": 669217802.0, + "step": 25865 + }, + { + "epoch": 2.840544695804964, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5144121646881104, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7475686073303223, + "num_tokens": 669241755.0, + "step": 25866 + }, + { + "epoch": 2.8406545135075776, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.759672164916992, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7076195478439331, + "num_tokens": 669263832.0, + "step": 25867 + }, + { + "epoch": 2.840764331210191, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4915764331817627, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7395201921463013, + "num_tokens": 669286704.0, + "step": 25868 + }, + { + "epoch": 2.8408741489128047, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3131093978881836, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7318832874298096, + "num_tokens": 669313161.0, + "step": 25869 + }, + { + "epoch": 2.8409839666154184, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.773831367492676, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7188272476196289, + "num_tokens": 669334300.0, + "step": 25870 + }, + { + "epoch": 2.841093784318032, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.230520486831665, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7304215431213379, + "num_tokens": 669362793.0, + "step": 25871 + }, + { + "epoch": 2.841203602020646, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.235814332962036, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7293037176132202, + "num_tokens": 669391541.0, + "step": 25872 + }, + { + "epoch": 2.8413134197232592, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1670494079589844, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7368643283843994, + "num_tokens": 669418882.0, + "step": 25873 + }, + { + "epoch": 2.841423237425873, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6248319149017334, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7196488976478577, + "num_tokens": 669441562.0, + "step": 25874 + }, + { + "epoch": 2.8415330551284868, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2892332077026367, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7042880058288574, + "num_tokens": 669468471.0, + "step": 25875 + }, + { + "epoch": 2.8416428728311005, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4599008560180664, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7166221737861633, + "num_tokens": 669493183.0, + "step": 25876 + }, + { + "epoch": 2.8417526905337143, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6045615673065186, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7257137298583984, + "num_tokens": 669516414.0, + "step": 25877 + }, + { + "epoch": 2.8418625082363276, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3527379035949707, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7128377556800842, + "num_tokens": 669542733.0, + "step": 25878 + }, + { + "epoch": 2.8419723259389413, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5189049243927, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7430288791656494, + "num_tokens": 669566008.0, + "step": 25879 + }, + { + "epoch": 2.842082143641555, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5088038444519043, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.731209397315979, + "num_tokens": 669593634.0, + "step": 25880 + }, + { + "epoch": 2.8421919613441684, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2629501819610596, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7376111745834351, + "num_tokens": 669621603.0, + "step": 25881 + }, + { + "epoch": 2.8423017790467826, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4444057941436768, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7384181022644043, + "num_tokens": 669644770.0, + "step": 25882 + }, + { + "epoch": 2.842411596749396, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4573209285736084, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.734553337097168, + "num_tokens": 669669975.0, + "step": 25883 + }, + { + "epoch": 2.8425214144520097, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.515359878540039, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7056528329849243, + "num_tokens": 669693620.0, + "step": 25884 + }, + { + "epoch": 2.8426312321546234, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2613720893859863, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7248597741127014, + "num_tokens": 669720800.0, + "step": 25885 + }, + { + "epoch": 2.8427410498572367, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5807549953460693, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7514550685882568, + "num_tokens": 669743147.0, + "step": 25886 + }, + { + "epoch": 2.8428508675598505, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.246385097503662, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7230468988418579, + "num_tokens": 669773579.0, + "step": 25887 + }, + { + "epoch": 2.8429606852624643, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2816848754882812, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7214117050170898, + "num_tokens": 669801002.0, + "step": 25888 + }, + { + "epoch": 2.843070502965078, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4139087200164795, + "learning_rate": 1e-06, + "loss": 0.7905, + "mean_token_accuracy": 0.7501965761184692, + "num_tokens": 669823367.0, + "step": 25889 + }, + { + "epoch": 2.8431803206676918, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1179862022399902, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7282887697219849, + "num_tokens": 669853913.0, + "step": 25890 + }, + { + "epoch": 2.843290138370305, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1544108390808105, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7168594598770142, + "num_tokens": 669885033.0, + "step": 25891 + }, + { + "epoch": 2.843399956072919, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.230715751647949, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7235780358314514, + "num_tokens": 669915585.0, + "step": 25892 + }, + { + "epoch": 2.8435097737755326, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.425466537475586, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7242506742477417, + "num_tokens": 669940756.0, + "step": 25893 + }, + { + "epoch": 2.8436195914781464, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.686185121536255, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7421346306800842, + "num_tokens": 669961158.0, + "step": 25894 + }, + { + "epoch": 2.84372940918076, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.471503496170044, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7333105802536011, + "num_tokens": 669984561.0, + "step": 25895 + }, + { + "epoch": 2.8438392268833734, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4956326484680176, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7310481071472168, + "num_tokens": 670008493.0, + "step": 25896 + }, + { + "epoch": 2.843949044585987, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.33198618888855, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7156316637992859, + "num_tokens": 670033800.0, + "step": 25897 + }, + { + "epoch": 2.844058862288601, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2901976108551025, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7265148162841797, + "num_tokens": 670060141.0, + "step": 25898 + }, + { + "epoch": 2.8441686799912147, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3131723403930664, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7397252321243286, + "num_tokens": 670087456.0, + "step": 25899 + }, + { + "epoch": 2.8442784976938285, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1146318912506104, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7310330271720886, + "num_tokens": 670118073.0, + "step": 25900 + }, + { + "epoch": 2.8443883153964418, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.433074951171875, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7332589626312256, + "num_tokens": 670142476.0, + "step": 25901 + }, + { + "epoch": 2.8444981330990555, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3163387775421143, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7317339181900024, + "num_tokens": 670168641.0, + "step": 25902 + }, + { + "epoch": 2.8446079508016693, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.182286262512207, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7221550941467285, + "num_tokens": 670201852.0, + "step": 25903 + }, + { + "epoch": 2.844717768504283, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6363298892974854, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7462804317474365, + "num_tokens": 670224372.0, + "step": 25904 + }, + { + "epoch": 2.844827586206897, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.326470375061035, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7297638058662415, + "num_tokens": 670252213.0, + "step": 25905 + }, + { + "epoch": 2.84493740390951, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.563490390777588, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7136691212654114, + "num_tokens": 670280173.0, + "step": 25906 + }, + { + "epoch": 2.845047221612124, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1751739978790283, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7477872967720032, + "num_tokens": 670310822.0, + "step": 25907 + }, + { + "epoch": 2.8451570393147376, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1356379985809326, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7368097305297852, + "num_tokens": 670338498.0, + "step": 25908 + }, + { + "epoch": 2.845266857017351, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5891189575195312, + "learning_rate": 1e-06, + "loss": 0.7728, + "mean_token_accuracy": 0.7594259977340698, + "num_tokens": 670360003.0, + "step": 25909 + }, + { + "epoch": 2.8453766747199647, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.157193660736084, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.722052812576294, + "num_tokens": 670392819.0, + "step": 25910 + }, + { + "epoch": 2.8454864924225785, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.276848316192627, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7204457521438599, + "num_tokens": 670420464.0, + "step": 25911 + }, + { + "epoch": 2.845596310125192, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1572675704956055, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.706752359867096, + "num_tokens": 670451696.0, + "step": 25912 + }, + { + "epoch": 2.845706127827806, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4171149730682373, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7249555587768555, + "num_tokens": 670480806.0, + "step": 25913 + }, + { + "epoch": 2.8458159455304193, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.759974956512451, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7423522472381592, + "num_tokens": 670502025.0, + "step": 25914 + }, + { + "epoch": 2.845925763233033, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.153477907180786, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7264107465744019, + "num_tokens": 670533258.0, + "step": 25915 + }, + { + "epoch": 2.846035580935647, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4294278621673584, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7296180725097656, + "num_tokens": 670560018.0, + "step": 25916 + }, + { + "epoch": 2.8461453986382605, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.226844310760498, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.698111891746521, + "num_tokens": 670590028.0, + "step": 25917 + }, + { + "epoch": 2.8462552163408743, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.3825719356536865, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7458323240280151, + "num_tokens": 670615101.0, + "step": 25918 + }, + { + "epoch": 2.8463650340434876, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.139711618423462, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7459629774093628, + "num_tokens": 670646292.0, + "step": 25919 + }, + { + "epoch": 2.8464748517461014, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.5238256454467773, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7340388298034668, + "num_tokens": 670667824.0, + "step": 25920 + }, + { + "epoch": 2.846584669448715, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6295535564422607, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7288000583648682, + "num_tokens": 670691045.0, + "step": 25921 + }, + { + "epoch": 2.846694487151329, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.113273859024048, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7257882356643677, + "num_tokens": 670722992.0, + "step": 25922 + }, + { + "epoch": 2.8468043048539426, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1572189331054688, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7254800200462341, + "num_tokens": 670751400.0, + "step": 25923 + }, + { + "epoch": 2.846914122556556, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2034780979156494, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7085527181625366, + "num_tokens": 670780341.0, + "step": 25924 + }, + { + "epoch": 2.8470239402591697, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.171072006225586, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.722619891166687, + "num_tokens": 670809210.0, + "step": 25925 + }, + { + "epoch": 2.8471337579617835, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.22550892829895, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7230832576751709, + "num_tokens": 670837346.0, + "step": 25926 + }, + { + "epoch": 2.8472435756643972, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.0924248695373535, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7428809404373169, + "num_tokens": 670866039.0, + "step": 25927 + }, + { + "epoch": 2.847353393367011, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.50683856010437, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7278057932853699, + "num_tokens": 670889048.0, + "step": 25928 + }, + { + "epoch": 2.8474632110696243, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.221278190612793, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7042874097824097, + "num_tokens": 670918102.0, + "step": 25929 + }, + { + "epoch": 2.847573028772238, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3570475578308105, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7309746146202087, + "num_tokens": 670944479.0, + "step": 25930 + }, + { + "epoch": 2.847682846474852, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3455746173858643, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7330226302146912, + "num_tokens": 670972970.0, + "step": 25931 + }, + { + "epoch": 2.8477926641774656, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.588663101196289, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7194939851760864, + "num_tokens": 670997133.0, + "step": 25932 + }, + { + "epoch": 2.8479024818800793, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5122313499450684, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7255133390426636, + "num_tokens": 671023043.0, + "step": 25933 + }, + { + "epoch": 2.8480122995826926, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6468265056610107, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.724776566028595, + "num_tokens": 671046145.0, + "step": 25934 + }, + { + "epoch": 2.8481221172853064, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2792696952819824, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7070794701576233, + "num_tokens": 671073990.0, + "step": 25935 + }, + { + "epoch": 2.84823193498792, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4799444675445557, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7448174357414246, + "num_tokens": 671096700.0, + "step": 25936 + }, + { + "epoch": 2.8483417526905335, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.337061643600464, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7215214967727661, + "num_tokens": 671121925.0, + "step": 25937 + }, + { + "epoch": 2.8484515703931472, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.213092088699341, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.731031596660614, + "num_tokens": 671149683.0, + "step": 25938 + }, + { + "epoch": 2.848561388095761, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4318363666534424, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7300058007240295, + "num_tokens": 671174455.0, + "step": 25939 + }, + { + "epoch": 2.8486712057983747, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4668612480163574, + "learning_rate": 1e-06, + "loss": 0.7745, + "mean_token_accuracy": 0.7540561556816101, + "num_tokens": 671197544.0, + "step": 25940 + }, + { + "epoch": 2.8487810235009885, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.2182693481445312, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7125628590583801, + "num_tokens": 671227099.0, + "step": 25941 + }, + { + "epoch": 2.848890841203602, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.7801501750946045, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7516787648200989, + "num_tokens": 671247278.0, + "step": 25942 + }, + { + "epoch": 2.8490006589062156, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.759056806564331, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7240766286849976, + "num_tokens": 671270021.0, + "step": 25943 + }, + { + "epoch": 2.8491104766088293, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3304359912872314, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7447444200515747, + "num_tokens": 671297070.0, + "step": 25944 + }, + { + "epoch": 2.849220294311443, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1584532260894775, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.719014048576355, + "num_tokens": 671327388.0, + "step": 25945 + }, + { + "epoch": 2.849330112014057, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.173776626586914, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7258018255233765, + "num_tokens": 671357515.0, + "step": 25946 + }, + { + "epoch": 2.84943992971667, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.302239418029785, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7430416345596313, + "num_tokens": 671383917.0, + "step": 25947 + }, + { + "epoch": 2.849549747419284, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.6087093353271484, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7371315956115723, + "num_tokens": 671405249.0, + "step": 25948 + }, + { + "epoch": 2.8496595651218977, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.394798994064331, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7173775434494019, + "num_tokens": 671434360.0, + "step": 25949 + }, + { + "epoch": 2.8497693828245114, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.1861982345581055, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7266727685928345, + "num_tokens": 671463945.0, + "step": 25950 + }, + { + "epoch": 2.849879200527125, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.366260051727295, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7464118003845215, + "num_tokens": 671488614.0, + "step": 25951 + }, + { + "epoch": 2.8499890182297385, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4214820861816406, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7269442081451416, + "num_tokens": 671516910.0, + "step": 25952 + }, + { + "epoch": 2.8500988359323522, + "ewc_loss": 0.0002269744873046875, + "grad_norm": 2.4188733100891113, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7366250157356262, + "num_tokens": 671542002.0, + "step": 25953 + }, + { + "epoch": 2.850208653634966, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.7658350467681885, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7283065319061279, + "num_tokens": 671564639.0, + "step": 25954 + }, + { + "epoch": 2.8503184713375798, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.403078556060791, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7246569395065308, + "num_tokens": 671591875.0, + "step": 25955 + }, + { + "epoch": 2.8504282890401935, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.402156352996826, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7161057591438293, + "num_tokens": 671618230.0, + "step": 25956 + }, + { + "epoch": 2.850538106742807, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.139849901199341, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7313275337219238, + "num_tokens": 671650632.0, + "step": 25957 + }, + { + "epoch": 2.8506479244454206, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4938273429870605, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.720238447189331, + "num_tokens": 671674132.0, + "step": 25958 + }, + { + "epoch": 2.8507577421480343, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3676202297210693, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7164244651794434, + "num_tokens": 671701303.0, + "step": 25959 + }, + { + "epoch": 2.8508675598506477, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.218271255493164, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7157918810844421, + "num_tokens": 671731503.0, + "step": 25960 + }, + { + "epoch": 2.850977377553262, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3106625080108643, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7551279664039612, + "num_tokens": 671758936.0, + "step": 25961 + }, + { + "epoch": 2.851087195255875, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.209806442260742, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7188402414321899, + "num_tokens": 671789371.0, + "step": 25962 + }, + { + "epoch": 2.851197012958489, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4482040405273438, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7393350005149841, + "num_tokens": 671815168.0, + "step": 25963 + }, + { + "epoch": 2.8513068306611027, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.575549602508545, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.7586386203765869, + "num_tokens": 671835725.0, + "step": 25964 + }, + { + "epoch": 2.851416648363716, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1772115230560303, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7121831178665161, + "num_tokens": 671864361.0, + "step": 25965 + }, + { + "epoch": 2.8515264660663298, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.262188673019409, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7167446613311768, + "num_tokens": 671893204.0, + "step": 25966 + }, + { + "epoch": 2.8516362837689435, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5041956901550293, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7307360172271729, + "num_tokens": 671917250.0, + "step": 25967 + }, + { + "epoch": 2.8517461014715573, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1462039947509766, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7153621912002563, + "num_tokens": 671948757.0, + "step": 25968 + }, + { + "epoch": 2.851855919174171, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.616225004196167, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7318168878555298, + "num_tokens": 671971315.0, + "step": 25969 + }, + { + "epoch": 2.8519657368767843, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.334773540496826, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7040996551513672, + "num_tokens": 672001667.0, + "step": 25970 + }, + { + "epoch": 2.852075554579398, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6608633995056152, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7519508600234985, + "num_tokens": 672021785.0, + "step": 25971 + }, + { + "epoch": 2.852185372282012, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3025805950164795, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7332723140716553, + "num_tokens": 672047617.0, + "step": 25972 + }, + { + "epoch": 2.8522951899846256, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6637234687805176, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7250058054924011, + "num_tokens": 672068906.0, + "step": 25973 + }, + { + "epoch": 2.8524050076872394, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2124907970428467, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7210745215415955, + "num_tokens": 672096120.0, + "step": 25974 + }, + { + "epoch": 2.8525148253898527, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3272855281829834, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7414246797561646, + "num_tokens": 672123519.0, + "step": 25975 + }, + { + "epoch": 2.8526246430924664, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4049088954925537, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7323562502861023, + "num_tokens": 672148463.0, + "step": 25976 + }, + { + "epoch": 2.85273446079508, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.333070755004883, + "learning_rate": 1e-06, + "loss": 0.774, + "mean_token_accuracy": 0.7622457146644592, + "num_tokens": 672175014.0, + "step": 25977 + }, + { + "epoch": 2.852844278497694, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6275534629821777, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7322646975517273, + "num_tokens": 672196646.0, + "step": 25978 + }, + { + "epoch": 2.8529540962003077, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3890364170074463, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7304168939590454, + "num_tokens": 672223961.0, + "step": 25979 + }, + { + "epoch": 2.853063913902921, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2139546871185303, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.732074499130249, + "num_tokens": 672252712.0, + "step": 25980 + }, + { + "epoch": 2.853173731605535, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4211349487304688, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.728842556476593, + "num_tokens": 672276394.0, + "step": 25981 + }, + { + "epoch": 2.8532835493081485, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4101145267486572, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.751001238822937, + "num_tokens": 672300503.0, + "step": 25982 + }, + { + "epoch": 2.8533933670107623, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.630061149597168, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7334826588630676, + "num_tokens": 672322330.0, + "step": 25983 + }, + { + "epoch": 2.853503184713376, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5267155170440674, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.732178807258606, + "num_tokens": 672345273.0, + "step": 25984 + }, + { + "epoch": 2.8536130024159894, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.340991258621216, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7191420793533325, + "num_tokens": 672372616.0, + "step": 25985 + }, + { + "epoch": 2.853722820118603, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.405404567718506, + "learning_rate": 1e-06, + "loss": 0.7036, + "mean_token_accuracy": 0.7771965265274048, + "num_tokens": 672394953.0, + "step": 25986 + }, + { + "epoch": 2.853832637821217, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2158236503601074, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7216504216194153, + "num_tokens": 672423046.0, + "step": 25987 + }, + { + "epoch": 2.85394245552383, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.7310476303100586, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.743847131729126, + "num_tokens": 672443008.0, + "step": 25988 + }, + { + "epoch": 2.854052273226444, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.471337080001831, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7309917211532593, + "num_tokens": 672467994.0, + "step": 25989 + }, + { + "epoch": 2.8541620909290577, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.22139573097229, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7072444558143616, + "num_tokens": 672497637.0, + "step": 25990 + }, + { + "epoch": 2.8542719086316715, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.288896083831787, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7255607843399048, + "num_tokens": 672522773.0, + "step": 25991 + }, + { + "epoch": 2.854381726334285, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1651229858398438, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7299264669418335, + "num_tokens": 672550357.0, + "step": 25992 + }, + { + "epoch": 2.8544915440368985, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5467944145202637, + "learning_rate": 1e-06, + "loss": 0.772, + "mean_token_accuracy": 0.7569442391395569, + "num_tokens": 672572533.0, + "step": 25993 + }, + { + "epoch": 2.8546013617395123, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2522525787353516, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7300082445144653, + "num_tokens": 672599089.0, + "step": 25994 + }, + { + "epoch": 2.854711179442126, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3108415603637695, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7409292459487915, + "num_tokens": 672626113.0, + "step": 25995 + }, + { + "epoch": 2.85482099714474, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6532540321350098, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7251181602478027, + "num_tokens": 672651555.0, + "step": 25996 + }, + { + "epoch": 2.8549308148473536, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6244990825653076, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7232742309570312, + "num_tokens": 672674340.0, + "step": 25997 + }, + { + "epoch": 2.855040632549967, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2077219486236572, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7502816915512085, + "num_tokens": 672704680.0, + "step": 25998 + }, + { + "epoch": 2.8551504502525806, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.430560827255249, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7094835042953491, + "num_tokens": 672731569.0, + "step": 25999 + }, + { + "epoch": 2.8552602679551944, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.0890603065490723, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7072387933731079, + "num_tokens": 672763196.0, + "step": 26000 + }, + { + "epoch": 2.855370085657808, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.5777881145477295, + "learning_rate": 1e-06, + "loss": 0.7501, + "mean_token_accuracy": 0.7589623928070068, + "num_tokens": 672784965.0, + "step": 26001 + }, + { + "epoch": 2.855479903360422, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2022335529327393, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7011481523513794, + "num_tokens": 672816672.0, + "step": 26002 + }, + { + "epoch": 2.855589721063035, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.459092378616333, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7520067691802979, + "num_tokens": 672840746.0, + "step": 26003 + }, + { + "epoch": 2.855699538765649, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.390303134918213, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7441140413284302, + "num_tokens": 672866472.0, + "step": 26004 + }, + { + "epoch": 2.8558093564682627, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2708096504211426, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7258440256118774, + "num_tokens": 672894541.0, + "step": 26005 + }, + { + "epoch": 2.8559191741708765, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4293553829193115, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7355793714523315, + "num_tokens": 672919064.0, + "step": 26006 + }, + { + "epoch": 2.8560289918734902, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.2761988639831543, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.725898265838623, + "num_tokens": 672946396.0, + "step": 26007 + }, + { + "epoch": 2.8561388095761036, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1685361862182617, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7257145643234253, + "num_tokens": 672975469.0, + "step": 26008 + }, + { + "epoch": 2.8562486272787173, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.0683960914611816, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.748343825340271, + "num_tokens": 673005393.0, + "step": 26009 + }, + { + "epoch": 2.856358444981331, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.818300724029541, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7553676962852478, + "num_tokens": 673023441.0, + "step": 26010 + }, + { + "epoch": 2.8564682626839444, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3461809158325195, + "learning_rate": 1e-06, + "loss": 0.7604, + "mean_token_accuracy": 0.757574200630188, + "num_tokens": 673049411.0, + "step": 26011 + }, + { + "epoch": 2.8565780803865586, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.9107167720794678, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7340865135192871, + "num_tokens": 673071869.0, + "step": 26012 + }, + { + "epoch": 2.856687898089172, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.319077491760254, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7113674879074097, + "num_tokens": 673098439.0, + "step": 26013 + }, + { + "epoch": 2.8567977157917857, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.58744478225708, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7351311445236206, + "num_tokens": 673120897.0, + "step": 26014 + }, + { + "epoch": 2.8569075334943994, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.6520869731903076, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7347109317779541, + "num_tokens": 673141537.0, + "step": 26015 + }, + { + "epoch": 2.8570173511970127, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3222527503967285, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7301480174064636, + "num_tokens": 673169449.0, + "step": 26016 + }, + { + "epoch": 2.8571271688996265, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.187591552734375, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7359389066696167, + "num_tokens": 673197705.0, + "step": 26017 + }, + { + "epoch": 2.8572369866022402, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1850616931915283, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7141463160514832, + "num_tokens": 673227243.0, + "step": 26018 + }, + { + "epoch": 2.857346804304854, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.520883560180664, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7214691042900085, + "num_tokens": 673252030.0, + "step": 26019 + }, + { + "epoch": 2.8574566220074678, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1982481479644775, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7380984425544739, + "num_tokens": 673280208.0, + "step": 26020 + }, + { + "epoch": 2.857566439710081, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.0539190769195557, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7168313264846802, + "num_tokens": 673311114.0, + "step": 26021 + }, + { + "epoch": 2.857676257412695, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.279226541519165, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7337878346443176, + "num_tokens": 673336523.0, + "step": 26022 + }, + { + "epoch": 2.8577860751153086, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.356539487838745, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.719397783279419, + "num_tokens": 673363109.0, + "step": 26023 + }, + { + "epoch": 2.8578958928179223, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.3445019721984863, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7475559711456299, + "num_tokens": 673388866.0, + "step": 26024 + }, + { + "epoch": 2.858005710520536, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.577414035797119, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7321229577064514, + "num_tokens": 673413250.0, + "step": 26025 + }, + { + "epoch": 2.8581155282231494, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.7203056812286377, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7395498752593994, + "num_tokens": 673433658.0, + "step": 26026 + }, + { + "epoch": 2.858225345925763, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.534184217453003, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7471957206726074, + "num_tokens": 673453861.0, + "step": 26027 + }, + { + "epoch": 2.858335163628377, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.1620678901672363, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6894201636314392, + "num_tokens": 673484844.0, + "step": 26028 + }, + { + "epoch": 2.8584449813309907, + "ewc_loss": 0.0002288818359375, + "grad_norm": 2.4435606002807617, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7464798092842102, + "num_tokens": 673510058.0, + "step": 26029 + }, + { + "epoch": 2.8585547990336044, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2752161026000977, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7317274808883667, + "num_tokens": 673537680.0, + "step": 26030 + }, + { + "epoch": 2.8586646167362177, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4730706214904785, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7337222099304199, + "num_tokens": 673562118.0, + "step": 26031 + }, + { + "epoch": 2.8587744344388315, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3842360973358154, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7490123510360718, + "num_tokens": 673587339.0, + "step": 26032 + }, + { + "epoch": 2.8588842521414453, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4556851387023926, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7323135733604431, + "num_tokens": 673612062.0, + "step": 26033 + }, + { + "epoch": 2.858994069844059, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.219785690307617, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7362552285194397, + "num_tokens": 673639397.0, + "step": 26034 + }, + { + "epoch": 2.8591038875466728, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.103227376937866, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7378576993942261, + "num_tokens": 673672041.0, + "step": 26035 + }, + { + "epoch": 2.859213705249286, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.610250949859619, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7220484614372253, + "num_tokens": 673696084.0, + "step": 26036 + }, + { + "epoch": 2.8593235229519, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1492373943328857, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7217981815338135, + "num_tokens": 673725454.0, + "step": 26037 + }, + { + "epoch": 2.8594333406545136, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6444802284240723, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7388641834259033, + "num_tokens": 673746551.0, + "step": 26038 + }, + { + "epoch": 2.859543158357127, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.370255708694458, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7419968843460083, + "num_tokens": 673772122.0, + "step": 26039 + }, + { + "epoch": 2.8596529760597407, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3626487255096436, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7296332120895386, + "num_tokens": 673797623.0, + "step": 26040 + }, + { + "epoch": 2.8597627937623544, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.109565019607544, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7211618423461914, + "num_tokens": 673828902.0, + "step": 26041 + }, + { + "epoch": 2.859872611464968, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1633636951446533, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7279307246208191, + "num_tokens": 673859008.0, + "step": 26042 + }, + { + "epoch": 2.859982429167582, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4782674312591553, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7465661764144897, + "num_tokens": 673881917.0, + "step": 26043 + }, + { + "epoch": 2.8600922468701953, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.19399094581604, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7335450649261475, + "num_tokens": 673912750.0, + "step": 26044 + }, + { + "epoch": 2.860202064572809, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5167348384857178, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7290509939193726, + "num_tokens": 673934835.0, + "step": 26045 + }, + { + "epoch": 2.8603118822754228, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.231579303741455, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7274448275566101, + "num_tokens": 673963165.0, + "step": 26046 + }, + { + "epoch": 2.8604216999780365, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5693962574005127, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7421596050262451, + "num_tokens": 673984440.0, + "step": 26047 + }, + { + "epoch": 2.8605315176806503, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.293389081954956, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7446534633636475, + "num_tokens": 674009921.0, + "step": 26048 + }, + { + "epoch": 2.8606413353832636, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.496081829071045, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7215537428855896, + "num_tokens": 674034168.0, + "step": 26049 + }, + { + "epoch": 2.8607511530858774, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.819783926010132, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.748689591884613, + "num_tokens": 674054595.0, + "step": 26050 + }, + { + "epoch": 2.860860970788491, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.8722777366638184, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7321248054504395, + "num_tokens": 674073090.0, + "step": 26051 + }, + { + "epoch": 2.860970788491105, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.150094985961914, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7259243726730347, + "num_tokens": 674102260.0, + "step": 26052 + }, + { + "epoch": 2.8610806061937186, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6295883655548096, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7209108471870422, + "num_tokens": 674123703.0, + "step": 26053 + }, + { + "epoch": 2.861190423896332, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.295522689819336, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7141910195350647, + "num_tokens": 674151101.0, + "step": 26054 + }, + { + "epoch": 2.8613002415989457, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4464190006256104, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7182866334915161, + "num_tokens": 674176981.0, + "step": 26055 + }, + { + "epoch": 2.8614100593015594, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.337933301925659, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7146141529083252, + "num_tokens": 674202164.0, + "step": 26056 + }, + { + "epoch": 2.861519877004173, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.479393243789673, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7402926683425903, + "num_tokens": 674227273.0, + "step": 26057 + }, + { + "epoch": 2.861629694706787, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3899288177490234, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7404845952987671, + "num_tokens": 674251889.0, + "step": 26058 + }, + { + "epoch": 2.8617395124094003, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.199093818664551, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7065584659576416, + "num_tokens": 674282868.0, + "step": 26059 + }, + { + "epoch": 2.861849330112014, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.321970224380493, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7257824540138245, + "num_tokens": 674310603.0, + "step": 26060 + }, + { + "epoch": 2.861959147814628, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5141072273254395, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7326287031173706, + "num_tokens": 674333678.0, + "step": 26061 + }, + { + "epoch": 2.862068965517241, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.293125629425049, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7516783475875854, + "num_tokens": 674361458.0, + "step": 26062 + }, + { + "epoch": 2.8621787832198553, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.586092233657837, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7448427677154541, + "num_tokens": 674384351.0, + "step": 26063 + }, + { + "epoch": 2.8622886009224686, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.241103172302246, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7086647748947144, + "num_tokens": 674411895.0, + "step": 26064 + }, + { + "epoch": 2.8623984186250824, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.337388515472412, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7054111957550049, + "num_tokens": 674441370.0, + "step": 26065 + }, + { + "epoch": 2.862508236327696, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0622448921203613, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7097470164299011, + "num_tokens": 674473057.0, + "step": 26066 + }, + { + "epoch": 2.8626180540303094, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3186488151550293, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7346330881118774, + "num_tokens": 674500205.0, + "step": 26067 + }, + { + "epoch": 2.862727871732923, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.654747724533081, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7318646907806396, + "num_tokens": 674522500.0, + "step": 26068 + }, + { + "epoch": 2.862837689435537, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1917521953582764, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7127968072891235, + "num_tokens": 674553971.0, + "step": 26069 + }, + { + "epoch": 2.8629475071381507, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1271979808807373, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7143833637237549, + "num_tokens": 674583947.0, + "step": 26070 + }, + { + "epoch": 2.8630573248407645, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4340291023254395, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7221325635910034, + "num_tokens": 674609286.0, + "step": 26071 + }, + { + "epoch": 2.863167142543378, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.9417479038238525, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7333725094795227, + "num_tokens": 674627776.0, + "step": 26072 + }, + { + "epoch": 2.8632769602459915, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3392999172210693, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7341170310974121, + "num_tokens": 674652947.0, + "step": 26073 + }, + { + "epoch": 2.8633867779486053, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2421910762786865, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7218064069747925, + "num_tokens": 674680767.0, + "step": 26074 + }, + { + "epoch": 2.863496595651219, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3235251903533936, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7185943722724915, + "num_tokens": 674706645.0, + "step": 26075 + }, + { + "epoch": 2.863606413353833, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3290815353393555, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7255655527114868, + "num_tokens": 674735371.0, + "step": 26076 + }, + { + "epoch": 2.863716231056446, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3936421871185303, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7202176451683044, + "num_tokens": 674761652.0, + "step": 26077 + }, + { + "epoch": 2.86382604875906, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2161238193511963, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7399052977561951, + "num_tokens": 674790968.0, + "step": 26078 + }, + { + "epoch": 2.8639358664616736, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2167272567749023, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.7018983960151672, + "num_tokens": 674820329.0, + "step": 26079 + }, + { + "epoch": 2.8640456841642874, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4302453994750977, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7401081919670105, + "num_tokens": 674843544.0, + "step": 26080 + }, + { + "epoch": 2.864155501866901, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.599296808242798, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7533663511276245, + "num_tokens": 674866950.0, + "step": 26081 + }, + { + "epoch": 2.8642653195695145, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4162521362304688, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.702483594417572, + "num_tokens": 674892458.0, + "step": 26082 + }, + { + "epoch": 2.8643751372721282, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7346067428588867, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.733005166053772, + "num_tokens": 674912625.0, + "step": 26083 + }, + { + "epoch": 2.864484954974742, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.349414110183716, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7237120866775513, + "num_tokens": 674937917.0, + "step": 26084 + }, + { + "epoch": 2.8645947726773557, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3297619819641113, + "learning_rate": 1e-06, + "loss": 0.7906, + "mean_token_accuracy": 0.7491593956947327, + "num_tokens": 674963788.0, + "step": 26085 + }, + { + "epoch": 2.8647045903799695, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.592092752456665, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7444890737533569, + "num_tokens": 674985862.0, + "step": 26086 + }, + { + "epoch": 2.864814408082583, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6601316928863525, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7320297360420227, + "num_tokens": 675006808.0, + "step": 26087 + }, + { + "epoch": 2.8649242257851966, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.335038423538208, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7311984300613403, + "num_tokens": 675033623.0, + "step": 26088 + }, + { + "epoch": 2.8650340434878103, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.570664167404175, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7456746697425842, + "num_tokens": 675053083.0, + "step": 26089 + }, + { + "epoch": 2.8651438611904236, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2193520069122314, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7221935987472534, + "num_tokens": 675082485.0, + "step": 26090 + }, + { + "epoch": 2.8652536788930374, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7396080493927, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7131138443946838, + "num_tokens": 675105349.0, + "step": 26091 + }, + { + "epoch": 2.865363496595651, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4422943592071533, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.730866014957428, + "num_tokens": 675129468.0, + "step": 26092 + }, + { + "epoch": 2.865473314298265, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.215137004852295, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7295479774475098, + "num_tokens": 675158711.0, + "step": 26093 + }, + { + "epoch": 2.8655831320008787, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2391340732574463, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7550851106643677, + "num_tokens": 675186157.0, + "step": 26094 + }, + { + "epoch": 2.865692949703492, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2032535076141357, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7246168255805969, + "num_tokens": 675213161.0, + "step": 26095 + }, + { + "epoch": 2.8658027674061057, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2956185340881348, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7241548895835876, + "num_tokens": 675243061.0, + "step": 26096 + }, + { + "epoch": 2.8659125851087195, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4769861698150635, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7430604696273804, + "num_tokens": 675267805.0, + "step": 26097 + }, + { + "epoch": 2.8660224028113332, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6761317253112793, + "learning_rate": 1e-06, + "loss": 0.7905, + "mean_token_accuracy": 0.7502002716064453, + "num_tokens": 675288863.0, + "step": 26098 + }, + { + "epoch": 2.866132220513947, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.348135232925415, + "learning_rate": 1e-06, + "loss": 0.7511, + "mean_token_accuracy": 0.7608124017715454, + "num_tokens": 675312823.0, + "step": 26099 + }, + { + "epoch": 2.8662420382165603, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2478854656219482, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.72698974609375, + "num_tokens": 675343825.0, + "step": 26100 + }, + { + "epoch": 2.866351855919174, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.712476968765259, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7215718030929565, + "num_tokens": 675364877.0, + "step": 26101 + }, + { + "epoch": 2.866461673621788, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2965073585510254, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7317212224006653, + "num_tokens": 675390248.0, + "step": 26102 + }, + { + "epoch": 2.8665714913244016, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1568198204040527, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7277883291244507, + "num_tokens": 675418776.0, + "step": 26103 + }, + { + "epoch": 2.8666813090270153, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 1.965519666671753, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7448089122772217, + "num_tokens": 675452449.0, + "step": 26104 + }, + { + "epoch": 2.8667911267296287, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5908772945404053, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7285906076431274, + "num_tokens": 675472964.0, + "step": 26105 + }, + { + "epoch": 2.8669009444322424, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.394042730331421, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7289112210273743, + "num_tokens": 675499433.0, + "step": 26106 + }, + { + "epoch": 2.867010762134856, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.142392635345459, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7076532244682312, + "num_tokens": 675531158.0, + "step": 26107 + }, + { + "epoch": 2.86712057983747, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6612353324890137, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7459837198257446, + "num_tokens": 675552905.0, + "step": 26108 + }, + { + "epoch": 2.8672303975400837, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4065115451812744, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7354629039764404, + "num_tokens": 675577409.0, + "step": 26109 + }, + { + "epoch": 2.867340215242697, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.512761354446411, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7461687326431274, + "num_tokens": 675600582.0, + "step": 26110 + }, + { + "epoch": 2.8674500329453108, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.492966890335083, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7509269118309021, + "num_tokens": 675623921.0, + "step": 26111 + }, + { + "epoch": 2.8675598506479245, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.610785722732544, + "learning_rate": 1e-06, + "loss": 0.7563, + "mean_token_accuracy": 0.7593624591827393, + "num_tokens": 675644555.0, + "step": 26112 + }, + { + "epoch": 2.8676696683505383, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6203627586364746, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7334111332893372, + "num_tokens": 675667029.0, + "step": 26113 + }, + { + "epoch": 2.867779486053152, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0421228408813477, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7079167366027832, + "num_tokens": 675702160.0, + "step": 26114 + }, + { + "epoch": 2.8678893037557653, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.471831798553467, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.748710036277771, + "num_tokens": 675723715.0, + "step": 26115 + }, + { + "epoch": 2.867999121458379, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.436856269836426, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7244122624397278, + "num_tokens": 675749148.0, + "step": 26116 + }, + { + "epoch": 2.868108939160993, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.297264575958252, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7193773984909058, + "num_tokens": 675776758.0, + "step": 26117 + }, + { + "epoch": 2.868218756863606, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4813194274902344, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7245627045631409, + "num_tokens": 675800674.0, + "step": 26118 + }, + { + "epoch": 2.86832857456622, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.263301134109497, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7097818851470947, + "num_tokens": 675827345.0, + "step": 26119 + }, + { + "epoch": 2.8684383922688337, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 1.9400899410247803, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7259269952774048, + "num_tokens": 675861466.0, + "step": 26120 + }, + { + "epoch": 2.8685482099714474, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5024659633636475, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7371708750724792, + "num_tokens": 675884135.0, + "step": 26121 + }, + { + "epoch": 2.868658027674061, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.441913604736328, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.732795000076294, + "num_tokens": 675908730.0, + "step": 26122 + }, + { + "epoch": 2.8687678453766745, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4847517013549805, + "learning_rate": 1e-06, + "loss": 0.7809, + "mean_token_accuracy": 0.7527090311050415, + "num_tokens": 675932288.0, + "step": 26123 + }, + { + "epoch": 2.8688776630792883, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.60611891746521, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7449396848678589, + "num_tokens": 675954517.0, + "step": 26124 + }, + { + "epoch": 2.868987480781902, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2561278343200684, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7356114983558655, + "num_tokens": 675980211.0, + "step": 26125 + }, + { + "epoch": 2.869097298484516, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3289291858673096, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7498435378074646, + "num_tokens": 676005009.0, + "step": 26126 + }, + { + "epoch": 2.8692071161871295, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3859002590179443, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7313627600669861, + "num_tokens": 676030630.0, + "step": 26127 + }, + { + "epoch": 2.869316933889743, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3320837020874023, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7421691417694092, + "num_tokens": 676055394.0, + "step": 26128 + }, + { + "epoch": 2.8694267515923566, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2872397899627686, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7423862814903259, + "num_tokens": 676081325.0, + "step": 26129 + }, + { + "epoch": 2.8695365692949704, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.334911584854126, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7124794721603394, + "num_tokens": 676107605.0, + "step": 26130 + }, + { + "epoch": 2.869646386997584, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5274598598480225, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7339911460876465, + "num_tokens": 676130392.0, + "step": 26131 + }, + { + "epoch": 2.869756204700198, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.516272783279419, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7258431911468506, + "num_tokens": 676155618.0, + "step": 26132 + }, + { + "epoch": 2.869866022402811, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.473841905593872, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7376697063446045, + "num_tokens": 676180193.0, + "step": 26133 + }, + { + "epoch": 2.869975840105425, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3649158477783203, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.702064573764801, + "num_tokens": 676207552.0, + "step": 26134 + }, + { + "epoch": 2.8700856578080387, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5506699085235596, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7295541763305664, + "num_tokens": 676231269.0, + "step": 26135 + }, + { + "epoch": 2.8701954755106525, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7298877239227295, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7514082193374634, + "num_tokens": 676251030.0, + "step": 26136 + }, + { + "epoch": 2.870305293213266, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4218881130218506, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7504897117614746, + "num_tokens": 676277159.0, + "step": 26137 + }, + { + "epoch": 2.8704151109158795, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.488619565963745, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7344146966934204, + "num_tokens": 676302300.0, + "step": 26138 + }, + { + "epoch": 2.8705249286184933, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6011745929718018, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7464690208435059, + "num_tokens": 676322622.0, + "step": 26139 + }, + { + "epoch": 2.870634746321107, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4159047603607178, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7175168991088867, + "num_tokens": 676347935.0, + "step": 26140 + }, + { + "epoch": 2.8707445640237204, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1331536769866943, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7106306552886963, + "num_tokens": 676379047.0, + "step": 26141 + }, + { + "epoch": 2.8708543817263346, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1833572387695312, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6919916868209839, + "num_tokens": 676408422.0, + "step": 26142 + }, + { + "epoch": 2.870964199428948, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3823065757751465, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7252559661865234, + "num_tokens": 676435121.0, + "step": 26143 + }, + { + "epoch": 2.8710740171315616, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.394963026046753, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.719589352607727, + "num_tokens": 676460750.0, + "step": 26144 + }, + { + "epoch": 2.8711838348341754, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4371111392974854, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7462441921234131, + "num_tokens": 676482900.0, + "step": 26145 + }, + { + "epoch": 2.8712936525367887, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.376819372177124, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7287873029708862, + "num_tokens": 676510402.0, + "step": 26146 + }, + { + "epoch": 2.8714034702394025, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.146960496902466, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7366279363632202, + "num_tokens": 676541557.0, + "step": 26147 + }, + { + "epoch": 2.871513287942016, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.290151357650757, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7503461241722107, + "num_tokens": 676567699.0, + "step": 26148 + }, + { + "epoch": 2.87162310564463, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5184881687164307, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7365167737007141, + "num_tokens": 676590430.0, + "step": 26149 + }, + { + "epoch": 2.8717329233472437, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.40830397605896, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7264699935913086, + "num_tokens": 676616324.0, + "step": 26150 + }, + { + "epoch": 2.871842741049857, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2006895542144775, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7344805002212524, + "num_tokens": 676645348.0, + "step": 26151 + }, + { + "epoch": 2.871952558752471, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7234530448913574, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7474414110183716, + "num_tokens": 676664263.0, + "step": 26152 + }, + { + "epoch": 2.8720623764550846, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.265580654144287, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7363884449005127, + "num_tokens": 676689852.0, + "step": 26153 + }, + { + "epoch": 2.8721721941576983, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.419438362121582, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7104227542877197, + "num_tokens": 676716524.0, + "step": 26154 + }, + { + "epoch": 2.872282011860312, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3048105239868164, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7435532808303833, + "num_tokens": 676741984.0, + "step": 26155 + }, + { + "epoch": 2.8723918295629254, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.134207248687744, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7097587585449219, + "num_tokens": 676774168.0, + "step": 26156 + }, + { + "epoch": 2.872501647265539, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1283063888549805, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7164672017097473, + "num_tokens": 676806321.0, + "step": 26157 + }, + { + "epoch": 2.872611464968153, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4608314037323, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7424628734588623, + "num_tokens": 676830146.0, + "step": 26158 + }, + { + "epoch": 2.8727212826707667, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.273956537246704, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7439961433410645, + "num_tokens": 676856918.0, + "step": 26159 + }, + { + "epoch": 2.8728311003733804, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7126810550689697, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.740545392036438, + "num_tokens": 676877334.0, + "step": 26160 + }, + { + "epoch": 2.8729409180759937, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.339529275894165, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7359198331832886, + "num_tokens": 676902466.0, + "step": 26161 + }, + { + "epoch": 2.8730507357786075, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.546844005584717, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7356306314468384, + "num_tokens": 676923813.0, + "step": 26162 + }, + { + "epoch": 2.8731605534812212, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.577000141143799, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7338504791259766, + "num_tokens": 676946898.0, + "step": 26163 + }, + { + "epoch": 2.873270371183835, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.459865093231201, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7245449423789978, + "num_tokens": 676972284.0, + "step": 26164 + }, + { + "epoch": 2.8733801888864487, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5054447650909424, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7287911772727966, + "num_tokens": 676997213.0, + "step": 26165 + }, + { + "epoch": 2.873490006589062, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5548315048217773, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7231295108795166, + "num_tokens": 677020832.0, + "step": 26166 + }, + { + "epoch": 2.873599824291676, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.529860019683838, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.719812273979187, + "num_tokens": 677044474.0, + "step": 26167 + }, + { + "epoch": 2.8737096419942896, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2408533096313477, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.737799882888794, + "num_tokens": 677071226.0, + "step": 26168 + }, + { + "epoch": 2.873819459696903, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4474024772644043, + "learning_rate": 1e-06, + "loss": 0.7643, + "mean_token_accuracy": 0.7532836198806763, + "num_tokens": 677094413.0, + "step": 26169 + }, + { + "epoch": 2.8739292773995166, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 1.9924507141113281, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7177599668502808, + "num_tokens": 677128378.0, + "step": 26170 + }, + { + "epoch": 2.8740390951021304, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6003048419952393, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7232377529144287, + "num_tokens": 677151911.0, + "step": 26171 + }, + { + "epoch": 2.874148912804744, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.217639446258545, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7408934831619263, + "num_tokens": 677180932.0, + "step": 26172 + }, + { + "epoch": 2.874258730507358, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.378829002380371, + "learning_rate": 1e-06, + "loss": 0.7772, + "mean_token_accuracy": 0.7503960132598877, + "num_tokens": 677204864.0, + "step": 26173 + }, + { + "epoch": 2.8743685482099712, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3443398475646973, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7255444526672363, + "num_tokens": 677229322.0, + "step": 26174 + }, + { + "epoch": 2.874478365912585, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.634561061859131, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7318554520606995, + "num_tokens": 677251165.0, + "step": 26175 + }, + { + "epoch": 2.8745881836151987, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4251480102539062, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7301623821258545, + "num_tokens": 677276506.0, + "step": 26176 + }, + { + "epoch": 2.8746980013178125, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4070987701416016, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7161368727684021, + "num_tokens": 677303683.0, + "step": 26177 + }, + { + "epoch": 2.8748078190204263, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2483041286468506, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7472943067550659, + "num_tokens": 677329909.0, + "step": 26178 + }, + { + "epoch": 2.8749176367230396, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.545433521270752, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7287783622741699, + "num_tokens": 677352625.0, + "step": 26179 + }, + { + "epoch": 2.8750274544256533, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3696305751800537, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7339072227478027, + "num_tokens": 677377575.0, + "step": 26180 + }, + { + "epoch": 2.875137272128267, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.039339065551758, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7186552882194519, + "num_tokens": 677413143.0, + "step": 26181 + }, + { + "epoch": 2.875247089830881, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4963676929473877, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7254551649093628, + "num_tokens": 677437420.0, + "step": 26182 + }, + { + "epoch": 2.8753569075334946, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4589173793792725, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7061061859130859, + "num_tokens": 677460817.0, + "step": 26183 + }, + { + "epoch": 2.875466725236108, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2434463500976562, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.6998088359832764, + "num_tokens": 677490990.0, + "step": 26184 + }, + { + "epoch": 2.8755765429387217, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3232717514038086, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.698430061340332, + "num_tokens": 677519083.0, + "step": 26185 + }, + { + "epoch": 2.8756863606413354, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7048346996307373, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.723565936088562, + "num_tokens": 677541676.0, + "step": 26186 + }, + { + "epoch": 2.875796178343949, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.145350217819214, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7076395750045776, + "num_tokens": 677574898.0, + "step": 26187 + }, + { + "epoch": 2.875905996046563, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5371265411376953, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7241084575653076, + "num_tokens": 677597511.0, + "step": 26188 + }, + { + "epoch": 2.8760158137491763, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4561712741851807, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.735059380531311, + "num_tokens": 677621779.0, + "step": 26189 + }, + { + "epoch": 2.87612563145179, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5310299396514893, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7198014855384827, + "num_tokens": 677645561.0, + "step": 26190 + }, + { + "epoch": 2.8762354491544038, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3456504344940186, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7384885549545288, + "num_tokens": 677669700.0, + "step": 26191 + }, + { + "epoch": 2.876345266857017, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2786591053009033, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7124464511871338, + "num_tokens": 677698827.0, + "step": 26192 + }, + { + "epoch": 2.8764550845596313, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1436734199523926, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7264040112495422, + "num_tokens": 677729257.0, + "step": 26193 + }, + { + "epoch": 2.8765649022622446, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.322831153869629, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7165793180465698, + "num_tokens": 677755695.0, + "step": 26194 + }, + { + "epoch": 2.8766747199648584, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2986419200897217, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7215890884399414, + "num_tokens": 677783088.0, + "step": 26195 + }, + { + "epoch": 2.876784537667472, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1820037364959717, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7306360006332397, + "num_tokens": 677814594.0, + "step": 26196 + }, + { + "epoch": 2.8768943553700854, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.526265859603882, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7503743171691895, + "num_tokens": 677836354.0, + "step": 26197 + }, + { + "epoch": 2.877004173072699, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2131075859069824, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7427796125411987, + "num_tokens": 677865061.0, + "step": 26198 + }, + { + "epoch": 2.877113990775313, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3751065731048584, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7033907175064087, + "num_tokens": 677893485.0, + "step": 26199 + }, + { + "epoch": 2.8772238084779267, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.452610969543457, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7222859263420105, + "num_tokens": 677917057.0, + "step": 26200 + }, + { + "epoch": 2.8773336261805404, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1024816036224365, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7413213849067688, + "num_tokens": 677946716.0, + "step": 26201 + }, + { + "epoch": 2.8774434438831538, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.378246307373047, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7110124826431274, + "num_tokens": 677974654.0, + "step": 26202 + }, + { + "epoch": 2.8775532615857675, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3940742015838623, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7198423147201538, + "num_tokens": 678000009.0, + "step": 26203 + }, + { + "epoch": 2.8776630792883813, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1666476726531982, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7353293299674988, + "num_tokens": 678027105.0, + "step": 26204 + }, + { + "epoch": 2.877772896990995, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0747838020324707, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7314720153808594, + "num_tokens": 678057760.0, + "step": 26205 + }, + { + "epoch": 2.877882714693609, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3994455337524414, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7338306307792664, + "num_tokens": 678084386.0, + "step": 26206 + }, + { + "epoch": 2.877992532396222, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.314470052719116, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7225031852722168, + "num_tokens": 678110250.0, + "step": 26207 + }, + { + "epoch": 2.878102350098836, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.285079002380371, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7099778056144714, + "num_tokens": 678138519.0, + "step": 26208 + }, + { + "epoch": 2.8782121678014496, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6197681427001953, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7231889367103577, + "num_tokens": 678160459.0, + "step": 26209 + }, + { + "epoch": 2.8783219855040634, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5380961894989014, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7387491464614868, + "num_tokens": 678183192.0, + "step": 26210 + }, + { + "epoch": 2.878431803206677, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2604846954345703, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.732863187789917, + "num_tokens": 678212127.0, + "step": 26211 + }, + { + "epoch": 2.8785416209092904, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2456984519958496, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7358609437942505, + "num_tokens": 678240686.0, + "step": 26212 + }, + { + "epoch": 2.878651438611904, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.183199167251587, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7578731775283813, + "num_tokens": 678270121.0, + "step": 26213 + }, + { + "epoch": 2.878761256314518, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.144015073776245, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7013089060783386, + "num_tokens": 678300259.0, + "step": 26214 + }, + { + "epoch": 2.8788710740171317, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2452237606048584, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7225151062011719, + "num_tokens": 678325595.0, + "step": 26215 + }, + { + "epoch": 2.8789808917197455, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4399986267089844, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7245514392852783, + "num_tokens": 678351866.0, + "step": 26216 + }, + { + "epoch": 2.879090709422359, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.279975652694702, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7182843685150146, + "num_tokens": 678380016.0, + "step": 26217 + }, + { + "epoch": 2.8792005271249725, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 3.8169755935668945, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7421075701713562, + "num_tokens": 678406271.0, + "step": 26218 + }, + { + "epoch": 2.8793103448275863, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2823898792266846, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7382419109344482, + "num_tokens": 678433377.0, + "step": 26219 + }, + { + "epoch": 2.8794201625301996, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.401210069656372, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7366679906845093, + "num_tokens": 678457691.0, + "step": 26220 + }, + { + "epoch": 2.8795299802328134, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2144412994384766, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7165385484695435, + "num_tokens": 678486304.0, + "step": 26221 + }, + { + "epoch": 2.879639797935427, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.321056604385376, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7184191942214966, + "num_tokens": 678512215.0, + "step": 26222 + }, + { + "epoch": 2.879749615638041, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.290208339691162, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7330213785171509, + "num_tokens": 678537965.0, + "step": 26223 + }, + { + "epoch": 2.8798594333406546, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1426336765289307, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7368236780166626, + "num_tokens": 678567305.0, + "step": 26224 + }, + { + "epoch": 2.879969251043268, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.651219129562378, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7371419668197632, + "num_tokens": 678588741.0, + "step": 26225 + }, + { + "epoch": 2.8800790687458817, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.392817497253418, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7230573296546936, + "num_tokens": 678614744.0, + "step": 26226 + }, + { + "epoch": 2.8801888864484955, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.868851661682129, + "learning_rate": 1e-06, + "loss": 0.7441, + "mean_token_accuracy": 0.7506847977638245, + "num_tokens": 678633366.0, + "step": 26227 + }, + { + "epoch": 2.8802987041511092, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.630117893218994, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7438011169433594, + "num_tokens": 678655254.0, + "step": 26228 + }, + { + "epoch": 2.880408521853723, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4340603351593018, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.752609133720398, + "num_tokens": 678678121.0, + "step": 26229 + }, + { + "epoch": 2.8805183395563363, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.07902193069458, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7270852327346802, + "num_tokens": 678708706.0, + "step": 26230 + }, + { + "epoch": 2.88062815725895, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2696762084960938, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7156757116317749, + "num_tokens": 678736911.0, + "step": 26231 + }, + { + "epoch": 2.880737974961564, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3938486576080322, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.729501485824585, + "num_tokens": 678760187.0, + "step": 26232 + }, + { + "epoch": 2.8808477926641776, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4800100326538086, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7347747087478638, + "num_tokens": 678783427.0, + "step": 26233 + }, + { + "epoch": 2.8809576103667913, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.549435615539551, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7358219623565674, + "num_tokens": 678806767.0, + "step": 26234 + }, + { + "epoch": 2.8810674280694046, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.60526704788208, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7267754077911377, + "num_tokens": 678831097.0, + "step": 26235 + }, + { + "epoch": 2.8811772457720184, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.272557258605957, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7174803018569946, + "num_tokens": 678859565.0, + "step": 26236 + }, + { + "epoch": 2.881287063474632, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2478139400482178, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7313547730445862, + "num_tokens": 678887636.0, + "step": 26237 + }, + { + "epoch": 2.881396881177246, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3297197818756104, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7294508814811707, + "num_tokens": 678915873.0, + "step": 26238 + }, + { + "epoch": 2.8815066988798597, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3058090209960938, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7261268496513367, + "num_tokens": 678941675.0, + "step": 26239 + }, + { + "epoch": 2.881616516582473, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.508134365081787, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7370296120643616, + "num_tokens": 678965197.0, + "step": 26240 + }, + { + "epoch": 2.8817263342850867, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4586660861968994, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7095690965652466, + "num_tokens": 678990892.0, + "step": 26241 + }, + { + "epoch": 2.8818361519877005, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.512282371520996, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7447847127914429, + "num_tokens": 679012837.0, + "step": 26242 + }, + { + "epoch": 2.881945969690314, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1435937881469727, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7043365240097046, + "num_tokens": 679045032.0, + "step": 26243 + }, + { + "epoch": 2.882055787392928, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.787663459777832, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7380449771881104, + "num_tokens": 679064203.0, + "step": 26244 + }, + { + "epoch": 2.8821656050955413, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2753922939300537, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7274115681648254, + "num_tokens": 679091818.0, + "step": 26245 + }, + { + "epoch": 2.882275422798155, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.446901559829712, + "learning_rate": 1e-06, + "loss": 0.7516, + "mean_token_accuracy": 0.758294939994812, + "num_tokens": 679117480.0, + "step": 26246 + }, + { + "epoch": 2.882385240500769, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.201974868774414, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7208073735237122, + "num_tokens": 679147118.0, + "step": 26247 + }, + { + "epoch": 2.882495058203382, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.44512677192688, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.721401572227478, + "num_tokens": 679171613.0, + "step": 26248 + }, + { + "epoch": 2.882604875905996, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.268016815185547, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7296925783157349, + "num_tokens": 679200715.0, + "step": 26249 + }, + { + "epoch": 2.8827146936086097, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.153916358947754, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7392594218254089, + "num_tokens": 679229449.0, + "step": 26250 + }, + { + "epoch": 2.8828245113112234, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3679566383361816, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7066479921340942, + "num_tokens": 679255059.0, + "step": 26251 + }, + { + "epoch": 2.882934329013837, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.726534128189087, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7362927198410034, + "num_tokens": 679274261.0, + "step": 26252 + }, + { + "epoch": 2.8830441467164505, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0603067874908447, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.755150556564331, + "num_tokens": 679302410.0, + "step": 26253 + }, + { + "epoch": 2.8831539644190642, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.301102876663208, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.755018413066864, + "num_tokens": 679327577.0, + "step": 26254 + }, + { + "epoch": 2.883263782121678, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3655340671539307, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7438528537750244, + "num_tokens": 679353635.0, + "step": 26255 + }, + { + "epoch": 2.8833735998242918, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.382819652557373, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.723594605922699, + "num_tokens": 679382008.0, + "step": 26256 + }, + { + "epoch": 2.8834834175269055, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2841222286224365, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7193060517311096, + "num_tokens": 679410828.0, + "step": 26257 + }, + { + "epoch": 2.883593235229519, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0233521461486816, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7259180545806885, + "num_tokens": 679443533.0, + "step": 26258 + }, + { + "epoch": 2.8837030529321326, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.810610055923462, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7455475926399231, + "num_tokens": 679463114.0, + "step": 26259 + }, + { + "epoch": 2.8838128706347463, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2424540519714355, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7560284733772278, + "num_tokens": 679490246.0, + "step": 26260 + }, + { + "epoch": 2.88392268833736, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2065131664276123, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.734894335269928, + "num_tokens": 679517195.0, + "step": 26261 + }, + { + "epoch": 2.884032506039974, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5814692974090576, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7399218678474426, + "num_tokens": 679540984.0, + "step": 26262 + }, + { + "epoch": 2.884142323742587, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6189193725585938, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.721301257610321, + "num_tokens": 679564476.0, + "step": 26263 + }, + { + "epoch": 2.884252141445201, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.266225576400757, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7294749021530151, + "num_tokens": 679591224.0, + "step": 26264 + }, + { + "epoch": 2.8843619591478147, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2209646701812744, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7538483142852783, + "num_tokens": 679618867.0, + "step": 26265 + }, + { + "epoch": 2.8844717768504284, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2469773292541504, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7164670825004578, + "num_tokens": 679647701.0, + "step": 26266 + }, + { + "epoch": 2.884581594553042, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.8125698566436768, + "learning_rate": 1e-06, + "loss": 0.7702, + "mean_token_accuracy": 0.7504163980484009, + "num_tokens": 679666694.0, + "step": 26267 + }, + { + "epoch": 2.8846914122556555, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3948721885681152, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7436332106590271, + "num_tokens": 679690306.0, + "step": 26268 + }, + { + "epoch": 2.8848012299582693, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7905943393707275, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7267345786094666, + "num_tokens": 679710856.0, + "step": 26269 + }, + { + "epoch": 2.884911047660883, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1652841567993164, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7140122056007385, + "num_tokens": 679740890.0, + "step": 26270 + }, + { + "epoch": 2.8850208653634963, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1121985912323, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7306686639785767, + "num_tokens": 679772541.0, + "step": 26271 + }, + { + "epoch": 2.88513068306611, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6931815147399902, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7404062747955322, + "num_tokens": 679792942.0, + "step": 26272 + }, + { + "epoch": 2.885240500768724, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2497997283935547, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.719113826751709, + "num_tokens": 679821926.0, + "step": 26273 + }, + { + "epoch": 2.8853503184713376, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.357189416885376, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7426549196243286, + "num_tokens": 679846469.0, + "step": 26274 + }, + { + "epoch": 2.8854601361739514, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.114655017852783, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7149471640586853, + "num_tokens": 679876950.0, + "step": 26275 + }, + { + "epoch": 2.8855699538765647, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.208564281463623, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7132954001426697, + "num_tokens": 679907480.0, + "step": 26276 + }, + { + "epoch": 2.8856797715791784, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4245285987854004, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7378891110420227, + "num_tokens": 679932416.0, + "step": 26277 + }, + { + "epoch": 2.885789589281792, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.447065591812134, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7291436195373535, + "num_tokens": 679956498.0, + "step": 26278 + }, + { + "epoch": 2.885899406984406, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.206894874572754, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7191407680511475, + "num_tokens": 679983183.0, + "step": 26279 + }, + { + "epoch": 2.8860092246870197, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.14487624168396, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7559770345687866, + "num_tokens": 680013070.0, + "step": 26280 + }, + { + "epoch": 2.886119042389633, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4723715782165527, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7165690660476685, + "num_tokens": 680038849.0, + "step": 26281 + }, + { + "epoch": 2.8862288600922468, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.320528984069824, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7299520969390869, + "num_tokens": 680064328.0, + "step": 26282 + }, + { + "epoch": 2.8863386777948605, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.410938024520874, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7098163962364197, + "num_tokens": 680091408.0, + "step": 26283 + }, + { + "epoch": 2.8864484954974743, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1191036701202393, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7371770143508911, + "num_tokens": 680123767.0, + "step": 26284 + }, + { + "epoch": 2.886558313200088, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.84428071975708, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7255131006240845, + "num_tokens": 680143812.0, + "step": 26285 + }, + { + "epoch": 2.8866681309027014, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3099112510681152, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7541707754135132, + "num_tokens": 680169260.0, + "step": 26286 + }, + { + "epoch": 2.886777948605315, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.34609055519104, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7433106899261475, + "num_tokens": 680193558.0, + "step": 26287 + }, + { + "epoch": 2.886887766307929, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3212599754333496, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7244366407394409, + "num_tokens": 680220597.0, + "step": 26288 + }, + { + "epoch": 2.8869975840105426, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.75130033493042, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7513201832771301, + "num_tokens": 680240737.0, + "step": 26289 + }, + { + "epoch": 2.8871074017131564, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3323707580566406, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7418076992034912, + "num_tokens": 680266014.0, + "step": 26290 + }, + { + "epoch": 2.8872172194157697, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4481334686279297, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.734856903553009, + "num_tokens": 680290115.0, + "step": 26291 + }, + { + "epoch": 2.8873270371183835, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3191025257110596, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7241995334625244, + "num_tokens": 680318250.0, + "step": 26292 + }, + { + "epoch": 2.887436854820997, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4069771766662598, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7288913726806641, + "num_tokens": 680342706.0, + "step": 26293 + }, + { + "epoch": 2.887546672523611, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.234467029571533, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7184797525405884, + "num_tokens": 680373243.0, + "step": 26294 + }, + { + "epoch": 2.8876564902262247, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0167129039764404, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7263056039810181, + "num_tokens": 680404986.0, + "step": 26295 + }, + { + "epoch": 2.887766307928838, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.363030195236206, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7249518632888794, + "num_tokens": 680430597.0, + "step": 26296 + }, + { + "epoch": 2.887876125631452, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2365365028381348, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7447348833084106, + "num_tokens": 680458734.0, + "step": 26297 + }, + { + "epoch": 2.8879859433340656, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2072913646698, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7047967314720154, + "num_tokens": 680490034.0, + "step": 26298 + }, + { + "epoch": 2.888095761036679, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3644485473632812, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7453465461730957, + "num_tokens": 680515547.0, + "step": 26299 + }, + { + "epoch": 2.8882055787392926, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.487544536590576, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7354001998901367, + "num_tokens": 680538587.0, + "step": 26300 + }, + { + "epoch": 2.8883153964419064, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.329702377319336, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.723753035068512, + "num_tokens": 680565636.0, + "step": 26301 + }, + { + "epoch": 2.88842521414452, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3984007835388184, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7376676201820374, + "num_tokens": 680589704.0, + "step": 26302 + }, + { + "epoch": 2.888535031847134, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5329959392547607, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.72414231300354, + "num_tokens": 680611431.0, + "step": 26303 + }, + { + "epoch": 2.888644849549747, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5225605964660645, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7197365164756775, + "num_tokens": 680634011.0, + "step": 26304 + }, + { + "epoch": 2.888754667252361, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3868465423583984, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7488484978675842, + "num_tokens": 680657458.0, + "step": 26305 + }, + { + "epoch": 2.8888644849549747, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1727776527404785, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.7582083344459534, + "num_tokens": 680683301.0, + "step": 26306 + }, + { + "epoch": 2.8889743026575885, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4380595684051514, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7134233713150024, + "num_tokens": 680709705.0, + "step": 26307 + }, + { + "epoch": 2.8890841203602022, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3535032272338867, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7220803499221802, + "num_tokens": 680736867.0, + "step": 26308 + }, + { + "epoch": 2.8891939380628155, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.481842041015625, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.730181097984314, + "num_tokens": 680761365.0, + "step": 26309 + }, + { + "epoch": 2.8893037557654293, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5313937664031982, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7141514420509338, + "num_tokens": 680785840.0, + "step": 26310 + }, + { + "epoch": 2.889413573468043, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4648756980895996, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7303308248519897, + "num_tokens": 680811875.0, + "step": 26311 + }, + { + "epoch": 2.889523391170657, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.452955484390259, + "learning_rate": 1e-06, + "loss": 0.7549, + "mean_token_accuracy": 0.7648601531982422, + "num_tokens": 680834686.0, + "step": 26312 + }, + { + "epoch": 2.8896332088732706, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2106168270111084, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7215025424957275, + "num_tokens": 680863684.0, + "step": 26313 + }, + { + "epoch": 2.889743026575884, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4740500450134277, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7341014742851257, + "num_tokens": 680886106.0, + "step": 26314 + }, + { + "epoch": 2.8898528442784976, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2487995624542236, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7235767245292664, + "num_tokens": 680914405.0, + "step": 26315 + }, + { + "epoch": 2.8899626619811114, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1913270950317383, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7279972434043884, + "num_tokens": 680943760.0, + "step": 26316 + }, + { + "epoch": 2.890072479683725, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3210325241088867, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7399919033050537, + "num_tokens": 680971377.0, + "step": 26317 + }, + { + "epoch": 2.890182297386339, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1195833683013916, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7252092361450195, + "num_tokens": 681002931.0, + "step": 26318 + }, + { + "epoch": 2.8902921150889522, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.251563787460327, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7263638377189636, + "num_tokens": 681030516.0, + "step": 26319 + }, + { + "epoch": 2.890401932791566, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1956002712249756, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7044127583503723, + "num_tokens": 681059702.0, + "step": 26320 + }, + { + "epoch": 2.8905117504941797, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2103450298309326, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6857339143753052, + "num_tokens": 681092970.0, + "step": 26321 + }, + { + "epoch": 2.890621568196793, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5240471363067627, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7474432587623596, + "num_tokens": 681116302.0, + "step": 26322 + }, + { + "epoch": 2.8907313858994073, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.407644748687744, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7218948602676392, + "num_tokens": 681140960.0, + "step": 26323 + }, + { + "epoch": 2.8908412036020206, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.470771551132202, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7220842242240906, + "num_tokens": 681166078.0, + "step": 26324 + }, + { + "epoch": 2.8909510213046343, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4890432357788086, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7108088135719299, + "num_tokens": 681194378.0, + "step": 26325 + }, + { + "epoch": 2.891060839007248, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3450419902801514, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7280305027961731, + "num_tokens": 681220387.0, + "step": 26326 + }, + { + "epoch": 2.8911706567098614, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4827163219451904, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7227410078048706, + "num_tokens": 681246376.0, + "step": 26327 + }, + { + "epoch": 2.891280474412475, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1979970932006836, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7141342759132385, + "num_tokens": 681277320.0, + "step": 26328 + }, + { + "epoch": 2.891390292115089, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4686079025268555, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7437041401863098, + "num_tokens": 681301160.0, + "step": 26329 + }, + { + "epoch": 2.8915001098177027, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6009836196899414, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7393969297409058, + "num_tokens": 681322226.0, + "step": 26330 + }, + { + "epoch": 2.8916099275203164, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4686899185180664, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7437909841537476, + "num_tokens": 681345895.0, + "step": 26331 + }, + { + "epoch": 2.8917197452229297, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.578563928604126, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7428531646728516, + "num_tokens": 681369058.0, + "step": 26332 + }, + { + "epoch": 2.8918295629255435, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4397361278533936, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7215006947517395, + "num_tokens": 681393988.0, + "step": 26333 + }, + { + "epoch": 2.8919393806281573, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.392916202545166, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7444844245910645, + "num_tokens": 681419271.0, + "step": 26334 + }, + { + "epoch": 2.892049198330771, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4571077823638916, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7149434685707092, + "num_tokens": 681445181.0, + "step": 26335 + }, + { + "epoch": 2.8921590160333848, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0524489879608154, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7141786813735962, + "num_tokens": 681478406.0, + "step": 26336 + }, + { + "epoch": 2.892268833735998, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.460057258605957, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7278138399124146, + "num_tokens": 681502338.0, + "step": 26337 + }, + { + "epoch": 2.892378651438612, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.447955846786499, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7117292284965515, + "num_tokens": 681531407.0, + "step": 26338 + }, + { + "epoch": 2.8924884691412256, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.124919891357422, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7127879858016968, + "num_tokens": 681563462.0, + "step": 26339 + }, + { + "epoch": 2.8925982868438394, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7112112045288086, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7489864826202393, + "num_tokens": 681584214.0, + "step": 26340 + }, + { + "epoch": 2.892708104546453, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.191502571105957, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7410087585449219, + "num_tokens": 681613791.0, + "step": 26341 + }, + { + "epoch": 2.8928179222490664, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2092716693878174, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7179362773895264, + "num_tokens": 681646841.0, + "step": 26342 + }, + { + "epoch": 2.89292773995168, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3279783725738525, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7144067883491516, + "num_tokens": 681672239.0, + "step": 26343 + }, + { + "epoch": 2.893037557654294, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7137255668640137, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7252479791641235, + "num_tokens": 681692934.0, + "step": 26344 + }, + { + "epoch": 2.8931473753569077, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.389611005783081, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7509676218032837, + "num_tokens": 681716216.0, + "step": 26345 + }, + { + "epoch": 2.8932571930595214, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2136337757110596, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7246956825256348, + "num_tokens": 681745947.0, + "step": 26346 + }, + { + "epoch": 2.8933670107621348, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7407288551330566, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7452609539031982, + "num_tokens": 681764836.0, + "step": 26347 + }, + { + "epoch": 2.8934768284647485, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4414658546447754, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7206020951271057, + "num_tokens": 681790826.0, + "step": 26348 + }, + { + "epoch": 2.8935866461673623, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.225210428237915, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7377235889434814, + "num_tokens": 681818482.0, + "step": 26349 + }, + { + "epoch": 2.8936964638699756, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4725427627563477, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.71321040391922, + "num_tokens": 681843663.0, + "step": 26350 + }, + { + "epoch": 2.8938062815725893, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.353351354598999, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7291125059127808, + "num_tokens": 681870011.0, + "step": 26351 + }, + { + "epoch": 2.893916099275203, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4036872386932373, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7452307343482971, + "num_tokens": 681894347.0, + "step": 26352 + }, + { + "epoch": 2.894025916977817, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.176470994949341, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.742023229598999, + "num_tokens": 681921147.0, + "step": 26353 + }, + { + "epoch": 2.8941357346804306, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3336355686187744, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.7404066324234009, + "num_tokens": 681946221.0, + "step": 26354 + }, + { + "epoch": 2.894245552383044, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.204627752304077, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.70000159740448, + "num_tokens": 681977249.0, + "step": 26355 + }, + { + "epoch": 2.8943553700856577, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2701263427734375, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.72907555103302, + "num_tokens": 682004527.0, + "step": 26356 + }, + { + "epoch": 2.8944651877882714, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4366543292999268, + "learning_rate": 1e-06, + "loss": 0.7686, + "mean_token_accuracy": 0.7545264959335327, + "num_tokens": 682028877.0, + "step": 26357 + }, + { + "epoch": 2.894575005490885, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.291189193725586, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7420631647109985, + "num_tokens": 682055359.0, + "step": 26358 + }, + { + "epoch": 2.894684823193499, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6362364292144775, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7378529906272888, + "num_tokens": 682077733.0, + "step": 26359 + }, + { + "epoch": 2.8947946408961123, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6369643211364746, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7227348685264587, + "num_tokens": 682099772.0, + "step": 26360 + }, + { + "epoch": 2.894904458598726, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3385794162750244, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7388308048248291, + "num_tokens": 682127088.0, + "step": 26361 + }, + { + "epoch": 2.89501427630134, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6152195930480957, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.736452579498291, + "num_tokens": 682149027.0, + "step": 26362 + }, + { + "epoch": 2.8951240940039535, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5705349445343018, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7233613133430481, + "num_tokens": 682172011.0, + "step": 26363 + }, + { + "epoch": 2.8952339117065673, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.348602533340454, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7172221541404724, + "num_tokens": 682199704.0, + "step": 26364 + }, + { + "epoch": 2.8953437294091806, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.69201397895813, + "learning_rate": 1e-06, + "loss": 0.7182, + "mean_token_accuracy": 0.7706853747367859, + "num_tokens": 682219918.0, + "step": 26365 + }, + { + "epoch": 2.8954535471117944, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7235376834869385, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.744509220123291, + "num_tokens": 682241287.0, + "step": 26366 + }, + { + "epoch": 2.895563364814408, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.392725706100464, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.726199746131897, + "num_tokens": 682266925.0, + "step": 26367 + }, + { + "epoch": 2.895673182517022, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.8067495822906494, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7416400909423828, + "num_tokens": 682287503.0, + "step": 26368 + }, + { + "epoch": 2.8957830002196356, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4201972484588623, + "learning_rate": 1e-06, + "loss": 0.7826, + "mean_token_accuracy": 0.7560673356056213, + "num_tokens": 682311142.0, + "step": 26369 + }, + { + "epoch": 2.895892817922249, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 1.9931048154830933, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7151955366134644, + "num_tokens": 682346342.0, + "step": 26370 + }, + { + "epoch": 2.8960026356248627, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.476454734802246, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7272793054580688, + "num_tokens": 682370628.0, + "step": 26371 + }, + { + "epoch": 2.8961124533274765, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.625169277191162, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7326366901397705, + "num_tokens": 682393547.0, + "step": 26372 + }, + { + "epoch": 2.89622227103009, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2548835277557373, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7388420104980469, + "num_tokens": 682421741.0, + "step": 26373 + }, + { + "epoch": 2.896332088732704, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2798426151275635, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7258672118186951, + "num_tokens": 682449528.0, + "step": 26374 + }, + { + "epoch": 2.8964419064353173, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4774930477142334, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7175189256668091, + "num_tokens": 682473028.0, + "step": 26375 + }, + { + "epoch": 2.896551724137931, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.355809450149536, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7209382653236389, + "num_tokens": 682500082.0, + "step": 26376 + }, + { + "epoch": 2.896661541840545, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.27248477935791, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7198003530502319, + "num_tokens": 682530425.0, + "step": 26377 + }, + { + "epoch": 2.896771359543158, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.586683511734009, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7234785556793213, + "num_tokens": 682554091.0, + "step": 26378 + }, + { + "epoch": 2.896881177245772, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3763415813446045, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7259876728057861, + "num_tokens": 682579492.0, + "step": 26379 + }, + { + "epoch": 2.8969909949483856, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6006903648376465, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7524370551109314, + "num_tokens": 682603660.0, + "step": 26380 + }, + { + "epoch": 2.8971008126509994, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3626251220703125, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7188999652862549, + "num_tokens": 682631859.0, + "step": 26381 + }, + { + "epoch": 2.897210630353613, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.48862886428833, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7431448101997375, + "num_tokens": 682656306.0, + "step": 26382 + }, + { + "epoch": 2.8973204480562265, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.497393846511841, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7340388298034668, + "num_tokens": 682679307.0, + "step": 26383 + }, + { + "epoch": 2.89743026575884, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3193042278289795, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.736757218837738, + "num_tokens": 682704005.0, + "step": 26384 + }, + { + "epoch": 2.897540083461454, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.26658296585083, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7427881360054016, + "num_tokens": 682730369.0, + "step": 26385 + }, + { + "epoch": 2.8976499011640677, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1383979320526123, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7190647125244141, + "num_tokens": 682761701.0, + "step": 26386 + }, + { + "epoch": 2.8977597188666815, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.259789228439331, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7385444045066833, + "num_tokens": 682787398.0, + "step": 26387 + }, + { + "epoch": 2.897869536569295, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3712995052337646, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7220890522003174, + "num_tokens": 682813490.0, + "step": 26388 + }, + { + "epoch": 2.8979793542719086, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4036450386047363, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7425424456596375, + "num_tokens": 682839551.0, + "step": 26389 + }, + { + "epoch": 2.8980891719745223, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.405557870864868, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7254205942153931, + "num_tokens": 682864892.0, + "step": 26390 + }, + { + "epoch": 2.898198989677136, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3770945072174072, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7122112512588501, + "num_tokens": 682892896.0, + "step": 26391 + }, + { + "epoch": 2.89830880737975, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6168196201324463, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7277838587760925, + "num_tokens": 682915798.0, + "step": 26392 + }, + { + "epoch": 2.898418625082363, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.337311029434204, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7301340103149414, + "num_tokens": 682940538.0, + "step": 26393 + }, + { + "epoch": 2.898528442784977, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.226173162460327, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7058417797088623, + "num_tokens": 682972109.0, + "step": 26394 + }, + { + "epoch": 2.8986382604875907, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2943978309631348, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.722215473651886, + "num_tokens": 682998786.0, + "step": 26395 + }, + { + "epoch": 2.8987480781902044, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4516985416412354, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7083761096000671, + "num_tokens": 683025291.0, + "step": 26396 + }, + { + "epoch": 2.898857895892818, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.150745391845703, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7335416674613953, + "num_tokens": 683056370.0, + "step": 26397 + }, + { + "epoch": 2.8989677135954315, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.315129518508911, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7296168208122253, + "num_tokens": 683084465.0, + "step": 26398 + }, + { + "epoch": 2.8990775312980452, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.569530725479126, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7320572137832642, + "num_tokens": 683108337.0, + "step": 26399 + }, + { + "epoch": 2.899187349000659, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.421429395675659, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7247488498687744, + "num_tokens": 683134858.0, + "step": 26400 + }, + { + "epoch": 2.8992971667032723, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5553088188171387, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7289682030677795, + "num_tokens": 683158821.0, + "step": 26401 + }, + { + "epoch": 2.899406984405886, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3092775344848633, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7222604751586914, + "num_tokens": 683185864.0, + "step": 26402 + }, + { + "epoch": 2.8995168021085, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.355527877807617, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7243527173995972, + "num_tokens": 683211009.0, + "step": 26403 + }, + { + "epoch": 2.8996266198111136, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.189976453781128, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7431229948997498, + "num_tokens": 683240253.0, + "step": 26404 + }, + { + "epoch": 2.8997364375137273, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6050989627838135, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7200720906257629, + "num_tokens": 683264149.0, + "step": 26405 + }, + { + "epoch": 2.8998462552163407, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.383535146713257, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7191857695579529, + "num_tokens": 683292588.0, + "step": 26406 + }, + { + "epoch": 2.8999560729189544, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5371363162994385, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7435475587844849, + "num_tokens": 683314964.0, + "step": 26407 + }, + { + "epoch": 2.900065890621568, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.213942050933838, + "learning_rate": 1e-06, + "loss": 0.7223, + "mean_token_accuracy": 0.7676649689674377, + "num_tokens": 683340361.0, + "step": 26408 + }, + { + "epoch": 2.900175708324182, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.623493194580078, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7355499267578125, + "num_tokens": 683362288.0, + "step": 26409 + }, + { + "epoch": 2.9002855260267957, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3608241081237793, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.728066086769104, + "num_tokens": 683387158.0, + "step": 26410 + }, + { + "epoch": 2.900395343729409, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7536416053771973, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7218669652938843, + "num_tokens": 683408163.0, + "step": 26411 + }, + { + "epoch": 2.9005051614320227, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.427635669708252, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7249799966812134, + "num_tokens": 683434116.0, + "step": 26412 + }, + { + "epoch": 2.9006149791346365, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2223398685455322, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7291254997253418, + "num_tokens": 683463104.0, + "step": 26413 + }, + { + "epoch": 2.9007247968372503, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.324122428894043, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7241257429122925, + "num_tokens": 683488989.0, + "step": 26414 + }, + { + "epoch": 2.900834614539864, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.106783628463745, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7224178314208984, + "num_tokens": 683520293.0, + "step": 26415 + }, + { + "epoch": 2.9009444322424773, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.702312469482422, + "learning_rate": 1e-06, + "loss": 0.7501, + "mean_token_accuracy": 0.7533320188522339, + "num_tokens": 683542115.0, + "step": 26416 + }, + { + "epoch": 2.901054249945091, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.248854637145996, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7318207025527954, + "num_tokens": 683572409.0, + "step": 26417 + }, + { + "epoch": 2.901164067647705, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4616804122924805, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7324748039245605, + "num_tokens": 683597498.0, + "step": 26418 + }, + { + "epoch": 2.9012738853503186, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 1.9987244606018066, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7431368827819824, + "num_tokens": 683632653.0, + "step": 26419 + }, + { + "epoch": 2.9013837030529324, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1986711025238037, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7299362421035767, + "num_tokens": 683662090.0, + "step": 26420 + }, + { + "epoch": 2.9014935207555457, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5981764793395996, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.734618604183197, + "num_tokens": 683684165.0, + "step": 26421 + }, + { + "epoch": 2.9016033384581594, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.419949531555176, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7157731056213379, + "num_tokens": 683710140.0, + "step": 26422 + }, + { + "epoch": 2.901713156160773, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3633103370666504, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7120104432106018, + "num_tokens": 683738357.0, + "step": 26423 + }, + { + "epoch": 2.9018229738633865, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.201606035232544, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7531794309616089, + "num_tokens": 683765150.0, + "step": 26424 + }, + { + "epoch": 2.9019327915660007, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4265329837799072, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7094328999519348, + "num_tokens": 683791127.0, + "step": 26425 + }, + { + "epoch": 2.902042609268614, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.432116746902466, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.737311840057373, + "num_tokens": 683814724.0, + "step": 26426 + }, + { + "epoch": 2.9021524269712278, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7102763652801514, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7637591361999512, + "num_tokens": 683834075.0, + "step": 26427 + }, + { + "epoch": 2.9022622446738415, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.07702898979187, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7427436709403992, + "num_tokens": 683865794.0, + "step": 26428 + }, + { + "epoch": 2.902372062376455, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4317939281463623, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7342309951782227, + "num_tokens": 683891188.0, + "step": 26429 + }, + { + "epoch": 2.9024818800790686, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.01837420463562, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6961507797241211, + "num_tokens": 683926150.0, + "step": 26430 + }, + { + "epoch": 2.9025916977816824, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6824777126312256, + "learning_rate": 1e-06, + "loss": 0.7543, + "mean_token_accuracy": 0.7628113627433777, + "num_tokens": 683946763.0, + "step": 26431 + }, + { + "epoch": 2.902701515484296, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.358529567718506, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7351928949356079, + "num_tokens": 683972539.0, + "step": 26432 + }, + { + "epoch": 2.90281133318691, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.538393259048462, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7448059916496277, + "num_tokens": 683995041.0, + "step": 26433 + }, + { + "epoch": 2.902921150889523, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0615622997283936, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7094530463218689, + "num_tokens": 684028733.0, + "step": 26434 + }, + { + "epoch": 2.903030968592137, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3319737911224365, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7246087789535522, + "num_tokens": 684056832.0, + "step": 26435 + }, + { + "epoch": 2.9031407862947507, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.182110071182251, + "learning_rate": 1e-06, + "loss": 0.7577, + "mean_token_accuracy": 0.7564330697059631, + "num_tokens": 684085015.0, + "step": 26436 + }, + { + "epoch": 2.9032506039973645, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5827829837799072, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7472995519638062, + "num_tokens": 684109006.0, + "step": 26437 + }, + { + "epoch": 2.903360421699978, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.553093433380127, + "learning_rate": 1e-06, + "loss": 0.7797, + "mean_token_accuracy": 0.7469965219497681, + "num_tokens": 684132450.0, + "step": 26438 + }, + { + "epoch": 2.9034702394025915, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0672671794891357, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7296184301376343, + "num_tokens": 684164024.0, + "step": 26439 + }, + { + "epoch": 2.9035800571052053, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1911470890045166, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7181459069252014, + "num_tokens": 684193418.0, + "step": 26440 + }, + { + "epoch": 2.903689874807819, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5476491451263428, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7240923047065735, + "num_tokens": 684219420.0, + "step": 26441 + }, + { + "epoch": 2.903799692510433, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3814616203308105, + "learning_rate": 1e-06, + "loss": 0.6699, + "mean_token_accuracy": 0.7834527492523193, + "num_tokens": 684241687.0, + "step": 26442 + }, + { + "epoch": 2.9039095102130466, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.262404680252075, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7367916107177734, + "num_tokens": 684270672.0, + "step": 26443 + }, + { + "epoch": 2.90401932791566, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4603540897369385, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7210500240325928, + "num_tokens": 684296671.0, + "step": 26444 + }, + { + "epoch": 2.9041291456182736, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3851218223571777, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7259119749069214, + "num_tokens": 684322906.0, + "step": 26445 + }, + { + "epoch": 2.9042389633208874, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.218459129333496, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7218705415725708, + "num_tokens": 684352381.0, + "step": 26446 + }, + { + "epoch": 2.904348781023501, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4964916706085205, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7333474159240723, + "num_tokens": 684376296.0, + "step": 26447 + }, + { + "epoch": 2.904458598726115, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.460447072982788, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7325792908668518, + "num_tokens": 684400919.0, + "step": 26448 + }, + { + "epoch": 2.904568416428728, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3731741905212402, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7223663330078125, + "num_tokens": 684428314.0, + "step": 26449 + }, + { + "epoch": 2.904678234131342, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5113747119903564, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7152009010314941, + "num_tokens": 684452540.0, + "step": 26450 + }, + { + "epoch": 2.9047880518339557, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.473602771759033, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7205824851989746, + "num_tokens": 684477807.0, + "step": 26451 + }, + { + "epoch": 2.904897869536569, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.8499252796173096, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7570675611495972, + "num_tokens": 684496994.0, + "step": 26452 + }, + { + "epoch": 2.905007687239183, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.089590311050415, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.718630313873291, + "num_tokens": 684529066.0, + "step": 26453 + }, + { + "epoch": 2.9051175049417965, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5102779865264893, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7493771314620972, + "num_tokens": 684551104.0, + "step": 26454 + }, + { + "epoch": 2.9052273226444103, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.334155559539795, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7314338684082031, + "num_tokens": 684577229.0, + "step": 26455 + }, + { + "epoch": 2.905337140347024, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2115976810455322, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7355994582176208, + "num_tokens": 684605572.0, + "step": 26456 + }, + { + "epoch": 2.9054469580496374, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1547739505767822, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7356173992156982, + "num_tokens": 684636023.0, + "step": 26457 + }, + { + "epoch": 2.905556775752251, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4420831203460693, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7223209142684937, + "num_tokens": 684663762.0, + "step": 26458 + }, + { + "epoch": 2.905666593454865, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5941951274871826, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7449950575828552, + "num_tokens": 684685399.0, + "step": 26459 + }, + { + "epoch": 2.9057764111574786, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.529747724533081, + "learning_rate": 1e-06, + "loss": 0.7544, + "mean_token_accuracy": 0.7570879459381104, + "num_tokens": 684706935.0, + "step": 26460 + }, + { + "epoch": 2.9058862288600924, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2386574745178223, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7364054918289185, + "num_tokens": 684735676.0, + "step": 26461 + }, + { + "epoch": 2.9059960465627057, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.175581932067871, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7254040241241455, + "num_tokens": 684767002.0, + "step": 26462 + }, + { + "epoch": 2.9061058642653195, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3536674976348877, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7178959846496582, + "num_tokens": 684794623.0, + "step": 26463 + }, + { + "epoch": 2.9062156819679332, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3678653240203857, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7320196032524109, + "num_tokens": 684820573.0, + "step": 26464 + }, + { + "epoch": 2.906325499670547, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.130605936050415, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7273343801498413, + "num_tokens": 684848905.0, + "step": 26465 + }, + { + "epoch": 2.9064353173731607, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.443690776824951, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.758714497089386, + "num_tokens": 684872597.0, + "step": 26466 + }, + { + "epoch": 2.906545135075774, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.402667999267578, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7043077349662781, + "num_tokens": 684899741.0, + "step": 26467 + }, + { + "epoch": 2.906654952778388, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5293312072753906, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7443695068359375, + "num_tokens": 684922939.0, + "step": 26468 + }, + { + "epoch": 2.9067647704810016, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3357014656066895, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7286543846130371, + "num_tokens": 684948105.0, + "step": 26469 + }, + { + "epoch": 2.9068745881836153, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0967979431152344, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7165340185165405, + "num_tokens": 684979234.0, + "step": 26470 + }, + { + "epoch": 2.906984405886229, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.593137741088867, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7424311637878418, + "num_tokens": 685001535.0, + "step": 26471 + }, + { + "epoch": 2.9070942235888424, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5625956058502197, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7326407432556152, + "num_tokens": 685024754.0, + "step": 26472 + }, + { + "epoch": 2.907204041291456, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.216647148132324, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7327344417572021, + "num_tokens": 685054084.0, + "step": 26473 + }, + { + "epoch": 2.90731385899407, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1704602241516113, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7202274799346924, + "num_tokens": 685084931.0, + "step": 26474 + }, + { + "epoch": 2.9074236766966837, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.34653902053833, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7189124822616577, + "num_tokens": 685112701.0, + "step": 26475 + }, + { + "epoch": 2.9075334943992974, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3258581161499023, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7252864837646484, + "num_tokens": 685139446.0, + "step": 26476 + }, + { + "epoch": 2.9076433121019107, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7548868656158447, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.751009464263916, + "num_tokens": 685160733.0, + "step": 26477 + }, + { + "epoch": 2.9077531298045245, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.151372194290161, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7229757905006409, + "num_tokens": 685189134.0, + "step": 26478 + }, + { + "epoch": 2.9078629475071383, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.577604293823242, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7552289366722107, + "num_tokens": 685211228.0, + "step": 26479 + }, + { + "epoch": 2.9079727652097516, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4414966106414795, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7268310785293579, + "num_tokens": 685236995.0, + "step": 26480 + }, + { + "epoch": 2.9080825829123653, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4223077297210693, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7236024141311646, + "num_tokens": 685262502.0, + "step": 26481 + }, + { + "epoch": 2.908192400614979, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5199341773986816, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7401120066642761, + "num_tokens": 685283923.0, + "step": 26482 + }, + { + "epoch": 2.908302218317593, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2238106727600098, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7302004098892212, + "num_tokens": 685311142.0, + "step": 26483 + }, + { + "epoch": 2.9084120360202066, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5110106468200684, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7434403896331787, + "num_tokens": 685334840.0, + "step": 26484 + }, + { + "epoch": 2.90852185372282, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3620431423187256, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7271997928619385, + "num_tokens": 685362203.0, + "step": 26485 + }, + { + "epoch": 2.9086316714254337, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.482034921646118, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7257187366485596, + "num_tokens": 685385678.0, + "step": 26486 + }, + { + "epoch": 2.9087414891280474, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1612441539764404, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7151205539703369, + "num_tokens": 685416591.0, + "step": 26487 + }, + { + "epoch": 2.908851306830661, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5307705402374268, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7298987507820129, + "num_tokens": 685439488.0, + "step": 26488 + }, + { + "epoch": 2.908961124533275, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.23153018951416, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7202368974685669, + "num_tokens": 685467009.0, + "step": 26489 + }, + { + "epoch": 2.9090709422358882, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2629947662353516, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7366647720336914, + "num_tokens": 685492831.0, + "step": 26490 + }, + { + "epoch": 2.909180759938502, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.381862163543701, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7212391495704651, + "num_tokens": 685520269.0, + "step": 26491 + }, + { + "epoch": 2.9092905776411158, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.740018844604492, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7398042678833008, + "num_tokens": 685539928.0, + "step": 26492 + }, + { + "epoch": 2.9094003953437295, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6597671508789062, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7140839099884033, + "num_tokens": 685564126.0, + "step": 26493 + }, + { + "epoch": 2.9095102130463433, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.264212131500244, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7318812012672424, + "num_tokens": 685590910.0, + "step": 26494 + }, + { + "epoch": 2.9096200307489566, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3534488677978516, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7191052436828613, + "num_tokens": 685620084.0, + "step": 26495 + }, + { + "epoch": 2.9097298484515703, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2358880043029785, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7478939890861511, + "num_tokens": 685647562.0, + "step": 26496 + }, + { + "epoch": 2.909839666154184, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.189788341522217, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7218069434165955, + "num_tokens": 685675349.0, + "step": 26497 + }, + { + "epoch": 2.909949483856798, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4553537368774414, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7452035546302795, + "num_tokens": 685698449.0, + "step": 26498 + }, + { + "epoch": 2.9100593015594116, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3062758445739746, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7053613066673279, + "num_tokens": 685727170.0, + "step": 26499 + }, + { + "epoch": 2.910169119262025, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.130870819091797, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7167373895645142, + "num_tokens": 685759830.0, + "step": 26500 + }, + { + "epoch": 2.9102789369646387, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.092481851577759, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7322099804878235, + "num_tokens": 685790878.0, + "step": 26501 + }, + { + "epoch": 2.9103887546672524, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.106795310974121, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7361913323402405, + "num_tokens": 685822032.0, + "step": 26502 + }, + { + "epoch": 2.9104985723698658, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4461588859558105, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7243359088897705, + "num_tokens": 685849320.0, + "step": 26503 + }, + { + "epoch": 2.91060839007248, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4414491653442383, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7176104784011841, + "num_tokens": 685873771.0, + "step": 26504 + }, + { + "epoch": 2.9107182077750933, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.165583610534668, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7111427783966064, + "num_tokens": 685904227.0, + "step": 26505 + }, + { + "epoch": 2.910828025477707, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3111891746520996, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7358349561691284, + "num_tokens": 685931888.0, + "step": 26506 + }, + { + "epoch": 2.910937843180321, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4801509380340576, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7356904745101929, + "num_tokens": 685955824.0, + "step": 26507 + }, + { + "epoch": 2.911047660882934, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3718488216400146, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.726218044757843, + "num_tokens": 685980405.0, + "step": 26508 + }, + { + "epoch": 2.911157478585548, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.720914840698242, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7372725009918213, + "num_tokens": 686000283.0, + "step": 26509 + }, + { + "epoch": 2.9112672962881616, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4230291843414307, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7447327375411987, + "num_tokens": 686024486.0, + "step": 26510 + }, + { + "epoch": 2.9113771139907754, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2276482582092285, + "learning_rate": 1e-06, + "loss": 0.7595, + "mean_token_accuracy": 0.7564144134521484, + "num_tokens": 686049863.0, + "step": 26511 + }, + { + "epoch": 2.911486931693389, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3439652919769287, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7372903227806091, + "num_tokens": 686074858.0, + "step": 26512 + }, + { + "epoch": 2.9115967493960024, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4875972270965576, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7264858484268188, + "num_tokens": 686098909.0, + "step": 26513 + }, + { + "epoch": 2.911706567098616, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3915231227874756, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7212271690368652, + "num_tokens": 686122999.0, + "step": 26514 + }, + { + "epoch": 2.91181638480123, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6276583671569824, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7397201061248779, + "num_tokens": 686146006.0, + "step": 26515 + }, + { + "epoch": 2.9119262025038437, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0631613731384277, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7191199660301208, + "num_tokens": 686179381.0, + "step": 26516 + }, + { + "epoch": 2.9120360202064575, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2860898971557617, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7254735827445984, + "num_tokens": 686207753.0, + "step": 26517 + }, + { + "epoch": 2.912145837909071, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4106898307800293, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7216064929962158, + "num_tokens": 686233578.0, + "step": 26518 + }, + { + "epoch": 2.9122556556116845, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.439833402633667, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7417248487472534, + "num_tokens": 686258649.0, + "step": 26519 + }, + { + "epoch": 2.9123654733142983, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3258073329925537, + "learning_rate": 1e-06, + "loss": 0.719, + "mean_token_accuracy": 0.7723140716552734, + "num_tokens": 686282824.0, + "step": 26520 + }, + { + "epoch": 2.912475291016912, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0083370208740234, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7205406427383423, + "num_tokens": 686315799.0, + "step": 26521 + }, + { + "epoch": 2.912585108719526, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4933621883392334, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7441122531890869, + "num_tokens": 686339348.0, + "step": 26522 + }, + { + "epoch": 2.912694926422139, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.736860990524292, + "learning_rate": 1e-06, + "loss": 0.7739, + "mean_token_accuracy": 0.755569577217102, + "num_tokens": 686360029.0, + "step": 26523 + }, + { + "epoch": 2.912804744124753, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3762269020080566, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7099081873893738, + "num_tokens": 686387739.0, + "step": 26524 + }, + { + "epoch": 2.9129145618273666, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5430939197540283, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7352335453033447, + "num_tokens": 686410635.0, + "step": 26525 + }, + { + "epoch": 2.9130243795299804, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.425427198410034, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7418303489685059, + "num_tokens": 686436426.0, + "step": 26526 + }, + { + "epoch": 2.913134197232594, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.075749397277832, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7164064645767212, + "num_tokens": 686466815.0, + "step": 26527 + }, + { + "epoch": 2.9132440149352075, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.401517629623413, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.716380774974823, + "num_tokens": 686491711.0, + "step": 26528 + }, + { + "epoch": 2.913353832637821, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4226090908050537, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7165671586990356, + "num_tokens": 686515784.0, + "step": 26529 + }, + { + "epoch": 2.913463650340435, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6585822105407715, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7232986092567444, + "num_tokens": 686538227.0, + "step": 26530 + }, + { + "epoch": 2.9135734680430483, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2319841384887695, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7081107497215271, + "num_tokens": 686568077.0, + "step": 26531 + }, + { + "epoch": 2.913683285745662, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.22119140625, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7284141778945923, + "num_tokens": 686597018.0, + "step": 26532 + }, + { + "epoch": 2.913793103448276, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.689905881881714, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7352373600006104, + "num_tokens": 686618219.0, + "step": 26533 + }, + { + "epoch": 2.9139029211508896, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4228007793426514, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7286008596420288, + "num_tokens": 686642533.0, + "step": 26534 + }, + { + "epoch": 2.9140127388535033, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.50494647026062, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7128176689147949, + "num_tokens": 686665114.0, + "step": 26535 + }, + { + "epoch": 2.9141225565561166, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2638256549835205, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.707837700843811, + "num_tokens": 686695160.0, + "step": 26536 + }, + { + "epoch": 2.9142323742587304, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.602797508239746, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7163236141204834, + "num_tokens": 686717975.0, + "step": 26537 + }, + { + "epoch": 2.914342191961344, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3680837154388428, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7299189567565918, + "num_tokens": 686744794.0, + "step": 26538 + }, + { + "epoch": 2.914452009663958, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.357949733734131, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7109348177909851, + "num_tokens": 686770704.0, + "step": 26539 + }, + { + "epoch": 2.9145618273665717, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.101036548614502, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7387992739677429, + "num_tokens": 686800460.0, + "step": 26540 + }, + { + "epoch": 2.914671645069185, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.039189577102661, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6906934380531311, + "num_tokens": 686835550.0, + "step": 26541 + }, + { + "epoch": 2.9147814627717987, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3294811248779297, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.735138475894928, + "num_tokens": 686860874.0, + "step": 26542 + }, + { + "epoch": 2.9148912804744125, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.270784854888916, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7199978232383728, + "num_tokens": 686889331.0, + "step": 26543 + }, + { + "epoch": 2.9150010981770262, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.236231803894043, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7234057188034058, + "num_tokens": 686919202.0, + "step": 26544 + }, + { + "epoch": 2.91511091587964, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5086467266082764, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7229466438293457, + "num_tokens": 686942904.0, + "step": 26545 + }, + { + "epoch": 2.9152207335822533, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1790311336517334, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7307320237159729, + "num_tokens": 686973126.0, + "step": 26546 + }, + { + "epoch": 2.915330551284867, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4586877822875977, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7354085445404053, + "num_tokens": 686998311.0, + "step": 26547 + }, + { + "epoch": 2.915440368987481, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2518749237060547, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7244851589202881, + "num_tokens": 687025114.0, + "step": 26548 + }, + { + "epoch": 2.9155501866900946, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.340632200241089, + "learning_rate": 1e-06, + "loss": 0.706, + "mean_token_accuracy": 0.7623502016067505, + "num_tokens": 687050089.0, + "step": 26549 + }, + { + "epoch": 2.9156600043927083, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2087857723236084, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7157437205314636, + "num_tokens": 687079672.0, + "step": 26550 + }, + { + "epoch": 2.9157698220953217, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4414122104644775, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7399251461029053, + "num_tokens": 687104830.0, + "step": 26551 + }, + { + "epoch": 2.9158796397979354, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.828778028488159, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7509931921958923, + "num_tokens": 687124592.0, + "step": 26552 + }, + { + "epoch": 2.915989457500549, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.409507989883423, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7494577169418335, + "num_tokens": 687149660.0, + "step": 26553 + }, + { + "epoch": 2.9160992752031625, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.540585517883301, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.712218165397644, + "num_tokens": 687173940.0, + "step": 26554 + }, + { + "epoch": 2.9162090929057767, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.162112236022949, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.721571147441864, + "num_tokens": 687206727.0, + "step": 26555 + }, + { + "epoch": 2.91631891060839, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5864853858947754, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7409728169441223, + "num_tokens": 687227811.0, + "step": 26556 + }, + { + "epoch": 2.9164287283110037, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1857569217681885, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7295833826065063, + "num_tokens": 687256730.0, + "step": 26557 + }, + { + "epoch": 2.9165385460136175, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.294555425643921, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7469719052314758, + "num_tokens": 687281628.0, + "step": 26558 + }, + { + "epoch": 2.916648363716231, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.11614990234375, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7366273403167725, + "num_tokens": 687310682.0, + "step": 26559 + }, + { + "epoch": 2.9167581814188446, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.1855452060699463, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7118266224861145, + "num_tokens": 687340045.0, + "step": 26560 + }, + { + "epoch": 2.9168679991214583, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.283235549926758, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7312438488006592, + "num_tokens": 687365898.0, + "step": 26561 + }, + { + "epoch": 2.916977816824072, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2430737018585205, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7550938725471497, + "num_tokens": 687390425.0, + "step": 26562 + }, + { + "epoch": 2.917087634526686, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6376149654388428, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7364814281463623, + "num_tokens": 687412248.0, + "step": 26563 + }, + { + "epoch": 2.917197452229299, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.408176898956299, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7265310287475586, + "num_tokens": 687438042.0, + "step": 26564 + }, + { + "epoch": 2.917307269931913, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.349557638168335, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7161830067634583, + "num_tokens": 687464699.0, + "step": 26565 + }, + { + "epoch": 2.9174170876345267, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2706761360168457, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7368655800819397, + "num_tokens": 687490154.0, + "step": 26566 + }, + { + "epoch": 2.9175269053371404, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.560110569000244, + "learning_rate": 1e-06, + "loss": 0.7549, + "mean_token_accuracy": 0.7510443329811096, + "num_tokens": 687510151.0, + "step": 26567 + }, + { + "epoch": 2.917636723039754, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.4037086963653564, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7232635021209717, + "num_tokens": 687534651.0, + "step": 26568 + }, + { + "epoch": 2.9177465407423675, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.534665584564209, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7176969051361084, + "num_tokens": 687557368.0, + "step": 26569 + }, + { + "epoch": 2.9178563584449813, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3566646575927734, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7362291812896729, + "num_tokens": 687581983.0, + "step": 26570 + }, + { + "epoch": 2.917966176147595, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.217226028442383, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7168025970458984, + "num_tokens": 687610019.0, + "step": 26571 + }, + { + "epoch": 2.9180759938502088, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.078268527984619, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7079080939292908, + "num_tokens": 687644128.0, + "step": 26572 + }, + { + "epoch": 2.9181858115528225, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.7417476177215576, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.712486982345581, + "num_tokens": 687664735.0, + "step": 26573 + }, + { + "epoch": 2.918295629255436, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.391324520111084, + "learning_rate": 1e-06, + "loss": 0.7363, + "mean_token_accuracy": 0.760525107383728, + "num_tokens": 687687794.0, + "step": 26574 + }, + { + "epoch": 2.9184054469580496, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.493523120880127, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7211626768112183, + "num_tokens": 687711311.0, + "step": 26575 + }, + { + "epoch": 2.9185152646606634, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.6426620483398438, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7393972277641296, + "num_tokens": 687733156.0, + "step": 26576 + }, + { + "epoch": 2.918625082363277, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.583749532699585, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7373393177986145, + "num_tokens": 687755265.0, + "step": 26577 + }, + { + "epoch": 2.918734900065891, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.323115587234497, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7373867034912109, + "num_tokens": 687780821.0, + "step": 26578 + }, + { + "epoch": 2.918844717768504, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.5854499340057373, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.750010073184967, + "num_tokens": 687803688.0, + "step": 26579 + }, + { + "epoch": 2.918954535471118, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.3611366748809814, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7410606741905212, + "num_tokens": 687829995.0, + "step": 26580 + }, + { + "epoch": 2.9190643531737317, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.2274677753448486, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7241628766059875, + "num_tokens": 687859682.0, + "step": 26581 + }, + { + "epoch": 2.919174170876345, + "ewc_loss": 0.00022983551025390625, + "grad_norm": 2.0387284755706787, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7132169008255005, + "num_tokens": 687893976.0, + "step": 26582 + }, + { + "epoch": 2.9192839885789588, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.207599639892578, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7494005560874939, + "num_tokens": 687922005.0, + "step": 26583 + }, + { + "epoch": 2.9193938062815725, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.537553548812866, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7338244915008545, + "num_tokens": 687946333.0, + "step": 26584 + }, + { + "epoch": 2.9195036239841863, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6204113960266113, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.7578458786010742, + "num_tokens": 687966792.0, + "step": 26585 + }, + { + "epoch": 2.9196134416868, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3160135746002197, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7343035936355591, + "num_tokens": 687993284.0, + "step": 26586 + }, + { + "epoch": 2.9197232593894134, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.610224962234497, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7364382147789001, + "num_tokens": 688016354.0, + "step": 26587 + }, + { + "epoch": 2.919833077092027, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.317535877227783, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7465113401412964, + "num_tokens": 688042962.0, + "step": 26588 + }, + { + "epoch": 2.919942894794641, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6043453216552734, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7239668369293213, + "num_tokens": 688065254.0, + "step": 26589 + }, + { + "epoch": 2.9200527124972546, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.226977586746216, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7010923624038696, + "num_tokens": 688094167.0, + "step": 26590 + }, + { + "epoch": 2.9201625301998684, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.330068826675415, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7350108623504639, + "num_tokens": 688121059.0, + "step": 26591 + }, + { + "epoch": 2.9202723479024817, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3288865089416504, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.728913426399231, + "num_tokens": 688148228.0, + "step": 26592 + }, + { + "epoch": 2.9203821656050954, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7592504024505615, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7415623664855957, + "num_tokens": 688168188.0, + "step": 26593 + }, + { + "epoch": 2.920491983307709, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3710248470306396, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7267751097679138, + "num_tokens": 688193376.0, + "step": 26594 + }, + { + "epoch": 2.920601801010323, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1750431060791016, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7231801748275757, + "num_tokens": 688223449.0, + "step": 26595 + }, + { + "epoch": 2.9207116187129367, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3849353790283203, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7350430488586426, + "num_tokens": 688248470.0, + "step": 26596 + }, + { + "epoch": 2.92082143641555, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.240412950515747, + "learning_rate": 1e-06, + "loss": 0.7599, + "mean_token_accuracy": 0.7558760643005371, + "num_tokens": 688275671.0, + "step": 26597 + }, + { + "epoch": 2.920931254118164, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2777886390686035, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7313624620437622, + "num_tokens": 688305394.0, + "step": 26598 + }, + { + "epoch": 2.9210410718207775, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6286001205444336, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7292253971099854, + "num_tokens": 688327261.0, + "step": 26599 + }, + { + "epoch": 2.9211508895233913, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.283167839050293, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7491800785064697, + "num_tokens": 688354440.0, + "step": 26600 + }, + { + "epoch": 2.921260707226005, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.627061367034912, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7329142093658447, + "num_tokens": 688376412.0, + "step": 26601 + }, + { + "epoch": 2.9213705249286184, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2062816619873047, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.70374596118927, + "num_tokens": 688408083.0, + "step": 26602 + }, + { + "epoch": 2.921480342631232, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4638609886169434, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7532185912132263, + "num_tokens": 688430832.0, + "step": 26603 + }, + { + "epoch": 2.921590160333846, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4405710697174072, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7277130484580994, + "num_tokens": 688455208.0, + "step": 26604 + }, + { + "epoch": 2.921699978036459, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.286700487136841, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7374128103256226, + "num_tokens": 688480867.0, + "step": 26605 + }, + { + "epoch": 2.9218097957390734, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3637659549713135, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7177059054374695, + "num_tokens": 688508797.0, + "step": 26606 + }, + { + "epoch": 2.9219196134416867, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.210005044937134, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7074422836303711, + "num_tokens": 688539193.0, + "step": 26607 + }, + { + "epoch": 2.9220294311443005, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1436140537261963, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7261204123497009, + "num_tokens": 688568712.0, + "step": 26608 + }, + { + "epoch": 2.9221392488469142, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5714659690856934, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7269152402877808, + "num_tokens": 688592866.0, + "step": 26609 + }, + { + "epoch": 2.9222490665495275, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.508822202682495, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7381846308708191, + "num_tokens": 688617035.0, + "step": 26610 + }, + { + "epoch": 2.9223588842521413, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3072636127471924, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7413493990898132, + "num_tokens": 688642534.0, + "step": 26611 + }, + { + "epoch": 2.922468701954755, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5004475116729736, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7307649850845337, + "num_tokens": 688666128.0, + "step": 26612 + }, + { + "epoch": 2.922578519657369, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5363662242889404, + "learning_rate": 1e-06, + "loss": 0.7497, + "mean_token_accuracy": 0.7605196833610535, + "num_tokens": 688688505.0, + "step": 26613 + }, + { + "epoch": 2.9226883373599826, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1115808486938477, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7171230912208557, + "num_tokens": 688719595.0, + "step": 26614 + }, + { + "epoch": 2.922798155062596, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6378796100616455, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7149620652198792, + "num_tokens": 688741262.0, + "step": 26615 + }, + { + "epoch": 2.9229079727652096, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.539874792098999, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7320923805236816, + "num_tokens": 688764920.0, + "step": 26616 + }, + { + "epoch": 2.9230177904678234, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.208130359649658, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7079344987869263, + "num_tokens": 688796330.0, + "step": 26617 + }, + { + "epoch": 2.923127608170437, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5335049629211426, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7546969652175903, + "num_tokens": 688817560.0, + "step": 26618 + }, + { + "epoch": 2.923237425873051, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.456670045852661, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7399499416351318, + "num_tokens": 688840841.0, + "step": 26619 + }, + { + "epoch": 2.9233472435756642, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.614900588989258, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7344993352890015, + "num_tokens": 688863396.0, + "step": 26620 + }, + { + "epoch": 2.923457061278278, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.636866331100464, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.749765157699585, + "num_tokens": 688884373.0, + "step": 26621 + }, + { + "epoch": 2.9235668789808917, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5868096351623535, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.7607129216194153, + "num_tokens": 688906250.0, + "step": 26622 + }, + { + "epoch": 2.9236766966835055, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5665924549102783, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.738787055015564, + "num_tokens": 688927619.0, + "step": 26623 + }, + { + "epoch": 2.9237865143861193, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1880581378936768, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7504689693450928, + "num_tokens": 688955302.0, + "step": 26624 + }, + { + "epoch": 2.9238963320887326, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2109079360961914, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7165425419807434, + "num_tokens": 688985863.0, + "step": 26625 + }, + { + "epoch": 2.9240061497913463, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3166298866271973, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7460083365440369, + "num_tokens": 689013871.0, + "step": 26626 + }, + { + "epoch": 2.92411596749396, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.652247190475464, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7295556664466858, + "num_tokens": 689036331.0, + "step": 26627 + }, + { + "epoch": 2.924225785196574, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.332392692565918, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7277258634567261, + "num_tokens": 689064330.0, + "step": 26628 + }, + { + "epoch": 2.9243356028991876, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.46354079246521, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7183632850646973, + "num_tokens": 689089193.0, + "step": 26629 + }, + { + "epoch": 2.924445420601801, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2738075256347656, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7257497310638428, + "num_tokens": 689117563.0, + "step": 26630 + }, + { + "epoch": 2.9245552383044147, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4911179542541504, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7381348609924316, + "num_tokens": 689141597.0, + "step": 26631 + }, + { + "epoch": 2.9246650560070284, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.313148021697998, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7486615180969238, + "num_tokens": 689167430.0, + "step": 26632 + }, + { + "epoch": 2.9247748737096417, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1596896648406982, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7418979406356812, + "num_tokens": 689199668.0, + "step": 26633 + }, + { + "epoch": 2.924884691412256, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1957569122314453, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.729397177696228, + "num_tokens": 689232059.0, + "step": 26634 + }, + { + "epoch": 2.9249945091148692, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.318669319152832, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7151817679405212, + "num_tokens": 689259070.0, + "step": 26635 + }, + { + "epoch": 2.925104326817483, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3209125995635986, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7041956186294556, + "num_tokens": 689284076.0, + "step": 26636 + }, + { + "epoch": 2.9252141445200968, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5515730381011963, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7201298475265503, + "num_tokens": 689308431.0, + "step": 26637 + }, + { + "epoch": 2.92532396222271, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.0466530323028564, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7307063341140747, + "num_tokens": 689340239.0, + "step": 26638 + }, + { + "epoch": 2.925433779925324, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.222785234451294, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7332653999328613, + "num_tokens": 689369272.0, + "step": 26639 + }, + { + "epoch": 2.9255435976279376, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2086732387542725, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7017503976821899, + "num_tokens": 689400323.0, + "step": 26640 + }, + { + "epoch": 2.9256534153305513, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.672046184539795, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7364063858985901, + "num_tokens": 689422422.0, + "step": 26641 + }, + { + "epoch": 2.925763233033165, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2377395629882812, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7077596783638, + "num_tokens": 689451044.0, + "step": 26642 + }, + { + "epoch": 2.9258730507357784, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.540501832962036, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7439272403717041, + "num_tokens": 689473795.0, + "step": 26643 + }, + { + "epoch": 2.925982868438392, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2634294033050537, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7214726805686951, + "num_tokens": 689503380.0, + "step": 26644 + }, + { + "epoch": 2.926092686141006, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.544074296951294, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7261227369308472, + "num_tokens": 689526467.0, + "step": 26645 + }, + { + "epoch": 2.9262025038436197, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2680344581604004, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7157829999923706, + "num_tokens": 689555347.0, + "step": 26646 + }, + { + "epoch": 2.9263123215462334, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 1.9957553148269653, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.695644736289978, + "num_tokens": 689592022.0, + "step": 26647 + }, + { + "epoch": 2.9264221392488468, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5647737979888916, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7316387891769409, + "num_tokens": 689612055.0, + "step": 26648 + }, + { + "epoch": 2.9265319569514605, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4022347927093506, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7130594253540039, + "num_tokens": 689638227.0, + "step": 26649 + }, + { + "epoch": 2.9266417746540743, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2196242809295654, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7295066714286804, + "num_tokens": 689669711.0, + "step": 26650 + }, + { + "epoch": 2.926751592356688, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2416582107543945, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7166831493377686, + "num_tokens": 689699272.0, + "step": 26651 + }, + { + "epoch": 2.926861410059302, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2223081588745117, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7130350470542908, + "num_tokens": 689729717.0, + "step": 26652 + }, + { + "epoch": 2.926971227761915, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.268807888031006, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7254376411437988, + "num_tokens": 689755552.0, + "step": 26653 + }, + { + "epoch": 2.927081045464529, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5334219932556152, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7007513046264648, + "num_tokens": 689780275.0, + "step": 26654 + }, + { + "epoch": 2.9271908631671426, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.281564950942993, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7251666784286499, + "num_tokens": 689806057.0, + "step": 26655 + }, + { + "epoch": 2.9273006808697564, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4925196170806885, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7346116304397583, + "num_tokens": 689831959.0, + "step": 26656 + }, + { + "epoch": 2.92741049857237, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.095435619354248, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7337614893913269, + "num_tokens": 689863640.0, + "step": 26657 + }, + { + "epoch": 2.9275203162749834, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3717312812805176, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7199849486351013, + "num_tokens": 689892249.0, + "step": 26658 + }, + { + "epoch": 2.927630133977597, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4086382389068604, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7329607009887695, + "num_tokens": 689916482.0, + "step": 26659 + }, + { + "epoch": 2.927739951680211, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4018380641937256, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7423152923583984, + "num_tokens": 689941566.0, + "step": 26660 + }, + { + "epoch": 2.9278497693828243, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.257648468017578, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7340210676193237, + "num_tokens": 689968884.0, + "step": 26661 + }, + { + "epoch": 2.927959587085438, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.262594223022461, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.719663143157959, + "num_tokens": 689997689.0, + "step": 26662 + }, + { + "epoch": 2.9280694047880518, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5097155570983887, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7432301044464111, + "num_tokens": 690020953.0, + "step": 26663 + }, + { + "epoch": 2.9281792224906655, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2838940620422363, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7213994264602661, + "num_tokens": 690049019.0, + "step": 26664 + }, + { + "epoch": 2.9282890401932793, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3728418350219727, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7367448806762695, + "num_tokens": 690075642.0, + "step": 26665 + }, + { + "epoch": 2.9283988578958926, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.108092784881592, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7251274585723877, + "num_tokens": 690107687.0, + "step": 26666 + }, + { + "epoch": 2.9285086755985064, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.9227349758148193, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7283875942230225, + "num_tokens": 690128154.0, + "step": 26667 + }, + { + "epoch": 2.92861849330112, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6261990070343018, + "learning_rate": 1e-06, + "loss": 0.774, + "mean_token_accuracy": 0.7593899369239807, + "num_tokens": 690149424.0, + "step": 26668 + }, + { + "epoch": 2.928728311003734, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.407442331314087, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7386389970779419, + "num_tokens": 690174474.0, + "step": 26669 + }, + { + "epoch": 2.9288381287063476, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5765697956085205, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.737085223197937, + "num_tokens": 690198756.0, + "step": 26670 + }, + { + "epoch": 2.928947946408961, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.654818534851074, + "learning_rate": 1e-06, + "loss": 0.7354, + "mean_token_accuracy": 0.7649073004722595, + "num_tokens": 690219581.0, + "step": 26671 + }, + { + "epoch": 2.9290577641115747, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1669671535491943, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7171790599822998, + "num_tokens": 690249638.0, + "step": 26672 + }, + { + "epoch": 2.9291675818141885, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3703081607818604, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7146915197372437, + "num_tokens": 690275987.0, + "step": 26673 + }, + { + "epoch": 2.929277399516802, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.731201410293579, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7536952495574951, + "num_tokens": 690295474.0, + "step": 26674 + }, + { + "epoch": 2.929387217219416, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.729322671890259, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.739396333694458, + "num_tokens": 690315944.0, + "step": 26675 + }, + { + "epoch": 2.9294970349220293, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4190962314605713, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7354265451431274, + "num_tokens": 690340715.0, + "step": 26676 + }, + { + "epoch": 2.929606852624643, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.556016445159912, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7343206405639648, + "num_tokens": 690361929.0, + "step": 26677 + }, + { + "epoch": 2.929716670327257, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1669344902038574, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.717403769493103, + "num_tokens": 690391351.0, + "step": 26678 + }, + { + "epoch": 2.9298264880298706, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1520333290100098, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7230403423309326, + "num_tokens": 690421413.0, + "step": 26679 + }, + { + "epoch": 2.9299363057324843, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5155327320098877, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7071403861045837, + "num_tokens": 690446312.0, + "step": 26680 + }, + { + "epoch": 2.9300461234350976, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5282773971557617, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7033414244651794, + "num_tokens": 690471830.0, + "step": 26681 + }, + { + "epoch": 2.9301559411377114, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3886351585388184, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7117252349853516, + "num_tokens": 690501404.0, + "step": 26682 + }, + { + "epoch": 2.930265758840325, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.470395565032959, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.726828932762146, + "num_tokens": 690524742.0, + "step": 26683 + }, + { + "epoch": 2.9303755765429385, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5802485942840576, + "learning_rate": 1e-06, + "loss": 0.7402, + "mean_token_accuracy": 0.7724299430847168, + "num_tokens": 690546333.0, + "step": 26684 + }, + { + "epoch": 2.9304853942455527, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.518047332763672, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7442753314971924, + "num_tokens": 690569697.0, + "step": 26685 + }, + { + "epoch": 2.930595211948166, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3813300132751465, + "learning_rate": 1e-06, + "loss": 0.7136, + "mean_token_accuracy": 0.7719302177429199, + "num_tokens": 690593988.0, + "step": 26686 + }, + { + "epoch": 2.9307050296507797, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4313697814941406, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7229002118110657, + "num_tokens": 690618230.0, + "step": 26687 + }, + { + "epoch": 2.9308148473533935, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6580495834350586, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.739080548286438, + "num_tokens": 690639264.0, + "step": 26688 + }, + { + "epoch": 2.930924665056007, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4792885780334473, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7521664500236511, + "num_tokens": 690661980.0, + "step": 26689 + }, + { + "epoch": 2.9310344827586206, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5628321170806885, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7379788756370544, + "num_tokens": 690685725.0, + "step": 26690 + }, + { + "epoch": 2.9311443004612343, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5274808406829834, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7274676561355591, + "num_tokens": 690710122.0, + "step": 26691 + }, + { + "epoch": 2.931254118163848, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.549813985824585, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7317100763320923, + "num_tokens": 690737360.0, + "step": 26692 + }, + { + "epoch": 2.931363935866462, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6426544189453125, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7276058197021484, + "num_tokens": 690758151.0, + "step": 26693 + }, + { + "epoch": 2.931473753569075, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.273083209991455, + "learning_rate": 1e-06, + "loss": 0.7354, + "mean_token_accuracy": 0.762797474861145, + "num_tokens": 690782626.0, + "step": 26694 + }, + { + "epoch": 2.931583571271689, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6579508781433105, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7448173761367798, + "num_tokens": 690804268.0, + "step": 26695 + }, + { + "epoch": 2.9316933889743026, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.136646032333374, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7244407534599304, + "num_tokens": 690833758.0, + "step": 26696 + }, + { + "epoch": 2.9318032066769164, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.127297878265381, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7048747539520264, + "num_tokens": 690867639.0, + "step": 26697 + }, + { + "epoch": 2.93191302437953, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.0205016136169434, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7141630053520203, + "num_tokens": 690902197.0, + "step": 26698 + }, + { + "epoch": 2.9320228420821435, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2348742485046387, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7119446992874146, + "num_tokens": 690932496.0, + "step": 26699 + }, + { + "epoch": 2.9321326597847572, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7065258026123047, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7374825477600098, + "num_tokens": 690956064.0, + "step": 26700 + }, + { + "epoch": 2.932242477487371, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.36446475982666, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7209843993186951, + "num_tokens": 690980200.0, + "step": 26701 + }, + { + "epoch": 2.9323522951899847, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.416775703430176, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7357490062713623, + "num_tokens": 691003199.0, + "step": 26702 + }, + { + "epoch": 2.9324621128925985, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.564850091934204, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7326054573059082, + "num_tokens": 691025235.0, + "step": 26703 + }, + { + "epoch": 2.932571930595212, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3312079906463623, + "learning_rate": 1e-06, + "loss": 0.7648, + "mean_token_accuracy": 0.7579526901245117, + "num_tokens": 691049426.0, + "step": 26704 + }, + { + "epoch": 2.9326817482978256, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1697168350219727, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7297320365905762, + "num_tokens": 691080489.0, + "step": 26705 + }, + { + "epoch": 2.9327915660004393, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.272081136703491, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7222079038619995, + "num_tokens": 691109836.0, + "step": 26706 + }, + { + "epoch": 2.932901383703053, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.519606113433838, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7333834171295166, + "num_tokens": 691133883.0, + "step": 26707 + }, + { + "epoch": 2.933011201405667, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.368945598602295, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7310353517532349, + "num_tokens": 691159481.0, + "step": 26708 + }, + { + "epoch": 2.93312101910828, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 3.0328173637390137, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.738494873046875, + "num_tokens": 691178012.0, + "step": 26709 + }, + { + "epoch": 2.933230836810894, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6186187267303467, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7277899384498596, + "num_tokens": 691200489.0, + "step": 26710 + }, + { + "epoch": 2.9333406545135077, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.671492576599121, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.743925929069519, + "num_tokens": 691220284.0, + "step": 26711 + }, + { + "epoch": 2.933450472216121, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2517495155334473, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7579125165939331, + "num_tokens": 691247468.0, + "step": 26712 + }, + { + "epoch": 2.9335602899187347, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.378034830093384, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7272486090660095, + "num_tokens": 691274301.0, + "step": 26713 + }, + { + "epoch": 2.9336701076213485, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7832298278808594, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.743717610836029, + "num_tokens": 691294714.0, + "step": 26714 + }, + { + "epoch": 2.9337799253239623, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3026986122131348, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7130813598632812, + "num_tokens": 691320042.0, + "step": 26715 + }, + { + "epoch": 2.933889743026576, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 3.080810070037842, + "learning_rate": 1e-06, + "loss": 0.773, + "mean_token_accuracy": 0.7552406787872314, + "num_tokens": 691336829.0, + "step": 26716 + }, + { + "epoch": 2.9339995607291893, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.106827974319458, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7480920553207397, + "num_tokens": 691366283.0, + "step": 26717 + }, + { + "epoch": 2.934109378431803, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5134453773498535, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7447786927223206, + "num_tokens": 691390445.0, + "step": 26718 + }, + { + "epoch": 2.934219196134417, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3666961193084717, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7392776608467102, + "num_tokens": 691416398.0, + "step": 26719 + }, + { + "epoch": 2.9343290138370306, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 1.9890695810317993, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7111124992370605, + "num_tokens": 691451529.0, + "step": 26720 + }, + { + "epoch": 2.9344388315396444, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.442232847213745, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7391850352287292, + "num_tokens": 691476310.0, + "step": 26721 + }, + { + "epoch": 2.9345486492422577, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6462974548339844, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7447603940963745, + "num_tokens": 691499514.0, + "step": 26722 + }, + { + "epoch": 2.9346584669448714, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.641002655029297, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7280317544937134, + "num_tokens": 691521563.0, + "step": 26723 + }, + { + "epoch": 2.934768284647485, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.409222364425659, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7249274849891663, + "num_tokens": 691545425.0, + "step": 26724 + }, + { + "epoch": 2.934878102350099, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3266327381134033, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7056395411491394, + "num_tokens": 691571772.0, + "step": 26725 + }, + { + "epoch": 2.9349879200527127, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.454267740249634, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7152923345565796, + "num_tokens": 691597828.0, + "step": 26726 + }, + { + "epoch": 2.935097737755326, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.274318218231201, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7030007243156433, + "num_tokens": 691627954.0, + "step": 26727 + }, + { + "epoch": 2.9352075554579398, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3656439781188965, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7019646167755127, + "num_tokens": 691655678.0, + "step": 26728 + }, + { + "epoch": 2.9353173731605535, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.442105531692505, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7242634892463684, + "num_tokens": 691681956.0, + "step": 26729 + }, + { + "epoch": 2.9354271908631673, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.431373119354248, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7323534488677979, + "num_tokens": 691708417.0, + "step": 26730 + }, + { + "epoch": 2.935537008565781, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6609861850738525, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7292183041572571, + "num_tokens": 691728792.0, + "step": 26731 + }, + { + "epoch": 2.9356468262683943, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5068359375, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7440550923347473, + "num_tokens": 691751611.0, + "step": 26732 + }, + { + "epoch": 2.935756643971008, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.637112617492676, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7269275784492493, + "num_tokens": 691773539.0, + "step": 26733 + }, + { + "epoch": 2.935866461673622, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5099105834960938, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7344943881034851, + "num_tokens": 691796543.0, + "step": 26734 + }, + { + "epoch": 2.935976279376235, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.309666395187378, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7365099787712097, + "num_tokens": 691824616.0, + "step": 26735 + }, + { + "epoch": 2.9360860970788494, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.514775276184082, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7178177833557129, + "num_tokens": 691847901.0, + "step": 26736 + }, + { + "epoch": 2.9361959147814627, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.181809902191162, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7420144081115723, + "num_tokens": 691875459.0, + "step": 26737 + }, + { + "epoch": 2.9363057324840764, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.251721143722534, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7231778502464294, + "num_tokens": 691905817.0, + "step": 26738 + }, + { + "epoch": 2.93641555018669, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2956910133361816, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.738986074924469, + "num_tokens": 691932060.0, + "step": 26739 + }, + { + "epoch": 2.9365253678893035, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2110610008239746, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7129116058349609, + "num_tokens": 691959423.0, + "step": 26740 + }, + { + "epoch": 2.9366351855919173, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1814651489257812, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7339431047439575, + "num_tokens": 691989589.0, + "step": 26741 + }, + { + "epoch": 2.936745003294531, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.240859031677246, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.733055830001831, + "num_tokens": 692018265.0, + "step": 26742 + }, + { + "epoch": 2.936854820997145, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2915849685668945, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7288091778755188, + "num_tokens": 692044060.0, + "step": 26743 + }, + { + "epoch": 2.9369646386997585, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2856714725494385, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7292641401290894, + "num_tokens": 692071309.0, + "step": 26744 + }, + { + "epoch": 2.937074456402372, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3905744552612305, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7466192245483398, + "num_tokens": 692095403.0, + "step": 26745 + }, + { + "epoch": 2.9371842741049856, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.389662504196167, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7201424837112427, + "num_tokens": 692120477.0, + "step": 26746 + }, + { + "epoch": 2.9372940918075994, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2552249431610107, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7226184606552124, + "num_tokens": 692147261.0, + "step": 26747 + }, + { + "epoch": 2.937403909510213, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5370945930480957, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7147747278213501, + "num_tokens": 692170175.0, + "step": 26748 + }, + { + "epoch": 2.937513727212827, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 1.9415955543518066, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7181621193885803, + "num_tokens": 692207159.0, + "step": 26749 + }, + { + "epoch": 2.93762354491544, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.665598154067993, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7467168569564819, + "num_tokens": 692228771.0, + "step": 26750 + }, + { + "epoch": 2.937733362618054, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4875919818878174, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7259159088134766, + "num_tokens": 692252062.0, + "step": 26751 + }, + { + "epoch": 2.9378431803206677, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.505753755569458, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7397750616073608, + "num_tokens": 692276146.0, + "step": 26752 + }, + { + "epoch": 2.9379529980232815, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3586061000823975, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7358837127685547, + "num_tokens": 692302206.0, + "step": 26753 + }, + { + "epoch": 2.9380628157258952, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5294723510742188, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7357546091079712, + "num_tokens": 692325652.0, + "step": 26754 + }, + { + "epoch": 2.9381726334285085, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1663482189178467, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7204866409301758, + "num_tokens": 692355280.0, + "step": 26755 + }, + { + "epoch": 2.9382824511311223, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.385420083999634, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7260928153991699, + "num_tokens": 692379783.0, + "step": 26756 + }, + { + "epoch": 2.938392268833736, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5288565158843994, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7336182594299316, + "num_tokens": 692402919.0, + "step": 26757 + }, + { + "epoch": 2.93850208653635, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.360623836517334, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.739059567451477, + "num_tokens": 692429074.0, + "step": 26758 + }, + { + "epoch": 2.9386119042389636, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5926976203918457, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7292006015777588, + "num_tokens": 692452534.0, + "step": 26759 + }, + { + "epoch": 2.938721721941577, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.464539051055908, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.737044095993042, + "num_tokens": 692477274.0, + "step": 26760 + }, + { + "epoch": 2.9388315396441906, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5947346687316895, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7400583028793335, + "num_tokens": 692500243.0, + "step": 26761 + }, + { + "epoch": 2.9389413573468044, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.496286392211914, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7280945777893066, + "num_tokens": 692523281.0, + "step": 26762 + }, + { + "epoch": 2.9390511750494177, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2385013103485107, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7367444038391113, + "num_tokens": 692550813.0, + "step": 26763 + }, + { + "epoch": 2.9391609927520315, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7807061672210693, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7402260303497314, + "num_tokens": 692571744.0, + "step": 26764 + }, + { + "epoch": 2.939270810454645, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.564328193664551, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7388737201690674, + "num_tokens": 692594231.0, + "step": 26765 + }, + { + "epoch": 2.939380628157259, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.668484926223755, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7407364249229431, + "num_tokens": 692614966.0, + "step": 26766 + }, + { + "epoch": 2.9394904458598727, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.498124837875366, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7340714335441589, + "num_tokens": 692638923.0, + "step": 26767 + }, + { + "epoch": 2.939600263562486, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.407583475112915, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7174489498138428, + "num_tokens": 692664177.0, + "step": 26768 + }, + { + "epoch": 2.9397100812651, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.492499828338623, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7409979104995728, + "num_tokens": 692686427.0, + "step": 26769 + }, + { + "epoch": 2.9398198989677136, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2905781269073486, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7248657941818237, + "num_tokens": 692714060.0, + "step": 26770 + }, + { + "epoch": 2.9399297166703273, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5835630893707275, + "learning_rate": 1e-06, + "loss": 0.6957, + "mean_token_accuracy": 0.7725163698196411, + "num_tokens": 692735925.0, + "step": 26771 + }, + { + "epoch": 2.940039534372941, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7735438346862793, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7310593128204346, + "num_tokens": 692755877.0, + "step": 26772 + }, + { + "epoch": 2.9401493520755544, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5642447471618652, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7132256031036377, + "num_tokens": 692779742.0, + "step": 26773 + }, + { + "epoch": 2.940259169778168, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3325893878936768, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7202372550964355, + "num_tokens": 692807891.0, + "step": 26774 + }, + { + "epoch": 2.940368987480782, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7535483837127686, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.734641969203949, + "num_tokens": 692827979.0, + "step": 26775 + }, + { + "epoch": 2.9404788051833957, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3148245811462402, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7038933038711548, + "num_tokens": 692857900.0, + "step": 26776 + }, + { + "epoch": 2.9405886228860094, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.218050479888916, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7128815650939941, + "num_tokens": 692887200.0, + "step": 26777 + }, + { + "epoch": 2.9406984405886227, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.199547529220581, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7370814085006714, + "num_tokens": 692916475.0, + "step": 26778 + }, + { + "epoch": 2.9408082582912365, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.415543794631958, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.723088264465332, + "num_tokens": 692941391.0, + "step": 26779 + }, + { + "epoch": 2.9409180759938502, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3693089485168457, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7375918626785278, + "num_tokens": 692965948.0, + "step": 26780 + }, + { + "epoch": 2.941027893696464, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5742299556732178, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.732455849647522, + "num_tokens": 692988724.0, + "step": 26781 + }, + { + "epoch": 2.9411377113990778, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6768205165863037, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7313816547393799, + "num_tokens": 693010367.0, + "step": 26782 + }, + { + "epoch": 2.941247529101691, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3233933448791504, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7368808388710022, + "num_tokens": 693036425.0, + "step": 26783 + }, + { + "epoch": 2.941357346804305, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.9785523414611816, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7419331073760986, + "num_tokens": 693055801.0, + "step": 26784 + }, + { + "epoch": 2.9414671645069186, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.7518184185028076, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7298564910888672, + "num_tokens": 693076977.0, + "step": 26785 + }, + { + "epoch": 2.9415769822095323, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2670705318450928, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7122852802276611, + "num_tokens": 693105603.0, + "step": 26786 + }, + { + "epoch": 2.941686799912146, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4080536365509033, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7376089096069336, + "num_tokens": 693132176.0, + "step": 26787 + }, + { + "epoch": 2.9417966176147594, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3341801166534424, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7425550222396851, + "num_tokens": 693158898.0, + "step": 26788 + }, + { + "epoch": 2.941906435317373, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4873316287994385, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7442294359207153, + "num_tokens": 693181428.0, + "step": 26789 + }, + { + "epoch": 2.942016253019987, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.279054641723633, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7282360196113586, + "num_tokens": 693209373.0, + "step": 26790 + }, + { + "epoch": 2.9421260707226002, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5453622341156006, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7457840442657471, + "num_tokens": 693233344.0, + "step": 26791 + }, + { + "epoch": 2.942235888425214, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.392239809036255, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7346103191375732, + "num_tokens": 693256521.0, + "step": 26792 + }, + { + "epoch": 2.9423457061278278, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5041215419769287, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7333953380584717, + "num_tokens": 693282044.0, + "step": 26793 + }, + { + "epoch": 2.9424555238304415, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3061349391937256, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7282825708389282, + "num_tokens": 693306197.0, + "step": 26794 + }, + { + "epoch": 2.9425653415330553, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2741661071777344, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7169234752655029, + "num_tokens": 693333976.0, + "step": 26795 + }, + { + "epoch": 2.9426751592356686, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.507078170776367, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.736717939376831, + "num_tokens": 693355923.0, + "step": 26796 + }, + { + "epoch": 2.9427849769382823, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2367515563964844, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7188576459884644, + "num_tokens": 693383526.0, + "step": 26797 + }, + { + "epoch": 2.942894794640896, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3719170093536377, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7210955619812012, + "num_tokens": 693408827.0, + "step": 26798 + }, + { + "epoch": 2.94300461234351, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.415158748626709, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.734754204750061, + "num_tokens": 693434950.0, + "step": 26799 + }, + { + "epoch": 2.9431144300461236, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.3006138801574707, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7352537512779236, + "num_tokens": 693461097.0, + "step": 26800 + }, + { + "epoch": 2.943224247748737, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.451106309890747, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7234209775924683, + "num_tokens": 693486772.0, + "step": 26801 + }, + { + "epoch": 2.9433340654513507, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1676206588745117, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7251616716384888, + "num_tokens": 693514765.0, + "step": 26802 + }, + { + "epoch": 2.9434438831539644, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4863662719726562, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7197331786155701, + "num_tokens": 693538919.0, + "step": 26803 + }, + { + "epoch": 2.943553700856578, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5585379600524902, + "learning_rate": 1e-06, + "loss": 0.7702, + "mean_token_accuracy": 0.7525523900985718, + "num_tokens": 693560016.0, + "step": 26804 + }, + { + "epoch": 2.943663518559192, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.189277410507202, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7217707633972168, + "num_tokens": 693590642.0, + "step": 26805 + }, + { + "epoch": 2.9437733362618053, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.5371649265289307, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7256940603256226, + "num_tokens": 693614369.0, + "step": 26806 + }, + { + "epoch": 2.943883153964419, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.198855400085449, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7273135781288147, + "num_tokens": 693643798.0, + "step": 26807 + }, + { + "epoch": 2.9439929716670328, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.0605850219726562, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7082625031471252, + "num_tokens": 693676593.0, + "step": 26808 + }, + { + "epoch": 2.9441027893696465, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.6445748805999756, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7422235012054443, + "num_tokens": 693698578.0, + "step": 26809 + }, + { + "epoch": 2.9442126070722603, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.1632230281829834, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7313390970230103, + "num_tokens": 693728855.0, + "step": 26810 + }, + { + "epoch": 2.9443224247748736, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2361457347869873, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.702045738697052, + "num_tokens": 693757497.0, + "step": 26811 + }, + { + "epoch": 2.9444322424774874, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.524062156677246, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7411506175994873, + "num_tokens": 693780563.0, + "step": 26812 + }, + { + "epoch": 2.944542060180101, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.224360942840576, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7174609899520874, + "num_tokens": 693807523.0, + "step": 26813 + }, + { + "epoch": 2.9446518778827144, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.4676356315612793, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7246459126472473, + "num_tokens": 693831195.0, + "step": 26814 + }, + { + "epoch": 2.9447616955853286, + "ewc_loss": 0.0002307891845703125, + "grad_norm": 2.2323200702667236, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7087370157241821, + "num_tokens": 693860382.0, + "step": 26815 + }, + { + "epoch": 2.944871513287942, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3186867237091064, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7218035459518433, + "num_tokens": 693887384.0, + "step": 26816 + }, + { + "epoch": 2.9449813309905557, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4653425216674805, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7391353845596313, + "num_tokens": 693910703.0, + "step": 26817 + }, + { + "epoch": 2.9450911486931695, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4213674068450928, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7122610211372375, + "num_tokens": 693935462.0, + "step": 26818 + }, + { + "epoch": 2.9452009663957828, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.523597478866577, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.746164858341217, + "num_tokens": 693956186.0, + "step": 26819 + }, + { + "epoch": 2.9453107840983965, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.170128107070923, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7278918027877808, + "num_tokens": 693985511.0, + "step": 26820 + }, + { + "epoch": 2.9454206018010103, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4688591957092285, + "learning_rate": 1e-06, + "loss": 0.7622, + "mean_token_accuracy": 0.7515451908111572, + "num_tokens": 694008428.0, + "step": 26821 + }, + { + "epoch": 2.945530419503624, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.290853261947632, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.732189416885376, + "num_tokens": 694035928.0, + "step": 26822 + }, + { + "epoch": 2.945640237206238, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3743629455566406, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.710831880569458, + "num_tokens": 694061531.0, + "step": 26823 + }, + { + "epoch": 2.945750054908851, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.462240695953369, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7274404764175415, + "num_tokens": 694084998.0, + "step": 26824 + }, + { + "epoch": 2.945859872611465, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5758113861083984, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7387726306915283, + "num_tokens": 694106285.0, + "step": 26825 + }, + { + "epoch": 2.9459696903140786, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.237792491912842, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7225815653800964, + "num_tokens": 694135134.0, + "step": 26826 + }, + { + "epoch": 2.9460795080166924, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.393531084060669, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7415815591812134, + "num_tokens": 694160220.0, + "step": 26827 + }, + { + "epoch": 2.946189325719306, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6396429538726807, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7250209450721741, + "num_tokens": 694181966.0, + "step": 26828 + }, + { + "epoch": 2.9462991434219195, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4355709552764893, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7307515144348145, + "num_tokens": 694207810.0, + "step": 26829 + }, + { + "epoch": 2.946408961124533, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1655187606811523, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7253170013427734, + "num_tokens": 694238163.0, + "step": 26830 + }, + { + "epoch": 2.946518778827147, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.593398094177246, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7502940893173218, + "num_tokens": 694260657.0, + "step": 26831 + }, + { + "epoch": 2.9466285965297607, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1316399574279785, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7053548097610474, + "num_tokens": 694290003.0, + "step": 26832 + }, + { + "epoch": 2.9467384142323745, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3648269176483154, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7299706339836121, + "num_tokens": 694315162.0, + "step": 26833 + }, + { + "epoch": 2.946848231934988, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.269930124282837, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7288198471069336, + "num_tokens": 694342414.0, + "step": 26834 + }, + { + "epoch": 2.9469580496376016, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.382528305053711, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7495753169059753, + "num_tokens": 694367724.0, + "step": 26835 + }, + { + "epoch": 2.9470678673402153, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.63004732131958, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7537352442741394, + "num_tokens": 694388463.0, + "step": 26836 + }, + { + "epoch": 2.947177685042829, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3421342372894287, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7419630289077759, + "num_tokens": 694415520.0, + "step": 26837 + }, + { + "epoch": 2.947287502745443, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5539803504943848, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7195771932601929, + "num_tokens": 694438833.0, + "step": 26838 + }, + { + "epoch": 2.947397320448056, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2174875736236572, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7466000318527222, + "num_tokens": 694466746.0, + "step": 26839 + }, + { + "epoch": 2.94750713815067, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6260976791381836, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.7556014060974121, + "num_tokens": 694488620.0, + "step": 26840 + }, + { + "epoch": 2.9476169558532836, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.492995262145996, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7223191857337952, + "num_tokens": 694513493.0, + "step": 26841 + }, + { + "epoch": 2.947726773555897, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5981438159942627, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7400112152099609, + "num_tokens": 694535953.0, + "step": 26842 + }, + { + "epoch": 2.9478365912585107, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2737905979156494, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7411706447601318, + "num_tokens": 694563545.0, + "step": 26843 + }, + { + "epoch": 2.9479464089611245, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.260524034500122, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.737193763256073, + "num_tokens": 694592406.0, + "step": 26844 + }, + { + "epoch": 2.9480562266637382, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.58986496925354, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.717838704586029, + "num_tokens": 694615286.0, + "step": 26845 + }, + { + "epoch": 2.948166044366352, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.324550151824951, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.707362711429596, + "num_tokens": 694643591.0, + "step": 26846 + }, + { + "epoch": 2.9482758620689653, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.728585958480835, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7422518730163574, + "num_tokens": 694664116.0, + "step": 26847 + }, + { + "epoch": 2.948385679771579, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4464359283447266, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7303792238235474, + "num_tokens": 694688798.0, + "step": 26848 + }, + { + "epoch": 2.948495497474193, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4908573627471924, + "learning_rate": 1e-06, + "loss": 0.7466, + "mean_token_accuracy": 0.7555465698242188, + "num_tokens": 694710847.0, + "step": 26849 + }, + { + "epoch": 2.9486053151768066, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.8050901889801025, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7408879995346069, + "num_tokens": 694729400.0, + "step": 26850 + }, + { + "epoch": 2.9487151328794203, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1417858600616455, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7073313593864441, + "num_tokens": 694760140.0, + "step": 26851 + }, + { + "epoch": 2.9488249505820336, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5964882373809814, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7218973636627197, + "num_tokens": 694782721.0, + "step": 26852 + }, + { + "epoch": 2.9489347682846474, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3429970741271973, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7337483167648315, + "num_tokens": 694807509.0, + "step": 26853 + }, + { + "epoch": 2.949044585987261, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.45963716506958, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.730592668056488, + "num_tokens": 694833524.0, + "step": 26854 + }, + { + "epoch": 2.949154403689875, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5029687881469727, + "learning_rate": 1e-06, + "loss": 0.7836, + "mean_token_accuracy": 0.7599458694458008, + "num_tokens": 694855875.0, + "step": 26855 + }, + { + "epoch": 2.9492642213924887, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.658125162124634, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7282073497772217, + "num_tokens": 694877608.0, + "step": 26856 + }, + { + "epoch": 2.949374039095102, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.348386287689209, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7323272824287415, + "num_tokens": 694901942.0, + "step": 26857 + }, + { + "epoch": 2.9494838567977157, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4974887371063232, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7408524751663208, + "num_tokens": 694925693.0, + "step": 26858 + }, + { + "epoch": 2.9495936745003295, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.213836908340454, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7293597459793091, + "num_tokens": 694957187.0, + "step": 26859 + }, + { + "epoch": 2.9497034922029433, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4485299587249756, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7288300395011902, + "num_tokens": 694982495.0, + "step": 26860 + }, + { + "epoch": 2.949813309905557, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.43148136138916, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7382088899612427, + "num_tokens": 695006353.0, + "step": 26861 + }, + { + "epoch": 2.9499231276081703, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4096808433532715, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7405014038085938, + "num_tokens": 695030589.0, + "step": 26862 + }, + { + "epoch": 2.950032945310784, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5003788471221924, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7223684191703796, + "num_tokens": 695055152.0, + "step": 26863 + }, + { + "epoch": 2.950142763013398, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5291078090667725, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7370795011520386, + "num_tokens": 695077887.0, + "step": 26864 + }, + { + "epoch": 2.950252580716011, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1486709117889404, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7168234586715698, + "num_tokens": 695108433.0, + "step": 26865 + }, + { + "epoch": 2.9503623984186254, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.403681993484497, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.723662793636322, + "num_tokens": 695134430.0, + "step": 26866 + }, + { + "epoch": 2.9504722161212387, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.601395606994629, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7312303781509399, + "num_tokens": 695157626.0, + "step": 26867 + }, + { + "epoch": 2.9505820338238524, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2810723781585693, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7170528173446655, + "num_tokens": 695185852.0, + "step": 26868 + }, + { + "epoch": 2.950691851526466, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.332622528076172, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7245887517929077, + "num_tokens": 695213099.0, + "step": 26869 + }, + { + "epoch": 2.9508016692290795, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.474515199661255, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7131959199905396, + "num_tokens": 695239514.0, + "step": 26870 + }, + { + "epoch": 2.9509114869316933, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1607437133789062, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7210003733634949, + "num_tokens": 695267943.0, + "step": 26871 + }, + { + "epoch": 2.951021304634307, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.675419569015503, + "learning_rate": 1e-06, + "loss": 0.7417, + "mean_token_accuracy": 0.7649242281913757, + "num_tokens": 695287393.0, + "step": 26872 + }, + { + "epoch": 2.9511311223369208, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.441298007965088, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.73215651512146, + "num_tokens": 695310829.0, + "step": 26873 + }, + { + "epoch": 2.9512409400395345, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.390251874923706, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7444187998771667, + "num_tokens": 695335851.0, + "step": 26874 + }, + { + "epoch": 2.951350757742148, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.429521322250366, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7321990728378296, + "num_tokens": 695359724.0, + "step": 26875 + }, + { + "epoch": 2.9514605754447616, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.450063467025757, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7233304977416992, + "num_tokens": 695383927.0, + "step": 26876 + }, + { + "epoch": 2.9515703931473753, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.066122055053711, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7320463061332703, + "num_tokens": 695415377.0, + "step": 26877 + }, + { + "epoch": 2.951680210849989, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.375274658203125, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7090901732444763, + "num_tokens": 695444063.0, + "step": 26878 + }, + { + "epoch": 2.951790028552603, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.49672532081604, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7381089925765991, + "num_tokens": 695468197.0, + "step": 26879 + }, + { + "epoch": 2.951899846255216, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1288468837738037, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7029111981391907, + "num_tokens": 695498017.0, + "step": 26880 + }, + { + "epoch": 2.95200966395783, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3172550201416016, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7036200761795044, + "num_tokens": 695523901.0, + "step": 26881 + }, + { + "epoch": 2.9521194816604437, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2950615882873535, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.730198860168457, + "num_tokens": 695552061.0, + "step": 26882 + }, + { + "epoch": 2.9522292993630574, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.231389284133911, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.700901985168457, + "num_tokens": 695582133.0, + "step": 26883 + }, + { + "epoch": 2.952339117065671, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.219688653945923, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7258377075195312, + "num_tokens": 695609439.0, + "step": 26884 + }, + { + "epoch": 2.9524489347682845, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.553983688354492, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.728278398513794, + "num_tokens": 695634479.0, + "step": 26885 + }, + { + "epoch": 2.9525587524708983, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4003775119781494, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7196410298347473, + "num_tokens": 695659018.0, + "step": 26886 + }, + { + "epoch": 2.952668570173512, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2213563919067383, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7279282808303833, + "num_tokens": 695685941.0, + "step": 26887 + }, + { + "epoch": 2.952778387876126, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.7249302864074707, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7229328155517578, + "num_tokens": 695706360.0, + "step": 26888 + }, + { + "epoch": 2.9528882055787395, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1954903602600098, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7410253286361694, + "num_tokens": 695736570.0, + "step": 26889 + }, + { + "epoch": 2.952998023281353, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5554001331329346, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7359265089035034, + "num_tokens": 695758780.0, + "step": 26890 + }, + { + "epoch": 2.9531078409839666, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6209139823913574, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7278201580047607, + "num_tokens": 695781814.0, + "step": 26891 + }, + { + "epoch": 2.9532176586865804, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.030167579650879, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.735100269317627, + "num_tokens": 695816254.0, + "step": 26892 + }, + { + "epoch": 2.9533274763891937, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4161927700042725, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.725258469581604, + "num_tokens": 695841352.0, + "step": 26893 + }, + { + "epoch": 2.9534372940918074, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.268458127975464, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7202310562133789, + "num_tokens": 695869778.0, + "step": 26894 + }, + { + "epoch": 2.953547111794421, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4540772438049316, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7220669984817505, + "num_tokens": 695895169.0, + "step": 26895 + }, + { + "epoch": 2.953656929497035, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0313642024993896, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7403979897499084, + "num_tokens": 695929771.0, + "step": 26896 + }, + { + "epoch": 2.9537667471996487, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.519808292388916, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7383695840835571, + "num_tokens": 695952223.0, + "step": 26897 + }, + { + "epoch": 2.953876564902262, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0753300189971924, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7145082354545593, + "num_tokens": 695983239.0, + "step": 26898 + }, + { + "epoch": 2.953986382604876, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4372551441192627, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7287154197692871, + "num_tokens": 696007787.0, + "step": 26899 + }, + { + "epoch": 2.9540962003074895, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.52736496925354, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7319998741149902, + "num_tokens": 696032528.0, + "step": 26900 + }, + { + "epoch": 2.9542060180101033, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.296900510787964, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.708010196685791, + "num_tokens": 696061706.0, + "step": 26901 + }, + { + "epoch": 2.954315835712717, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.241048812866211, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7299432754516602, + "num_tokens": 696090386.0, + "step": 26902 + }, + { + "epoch": 2.9544256534153304, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.342532157897949, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7384295463562012, + "num_tokens": 696114610.0, + "step": 26903 + }, + { + "epoch": 2.954535471117944, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0551397800445557, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7376539707183838, + "num_tokens": 696147123.0, + "step": 26904 + }, + { + "epoch": 2.954645288820558, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.172060012817383, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7332520484924316, + "num_tokens": 696176382.0, + "step": 26905 + }, + { + "epoch": 2.9547551065231716, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.391303062438965, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7258755564689636, + "num_tokens": 696207007.0, + "step": 26906 + }, + { + "epoch": 2.9548649242257854, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3972227573394775, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7324044108390808, + "num_tokens": 696233648.0, + "step": 26907 + }, + { + "epoch": 2.9549747419283987, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2752151489257812, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7310937643051147, + "num_tokens": 696261850.0, + "step": 26908 + }, + { + "epoch": 2.9550845596310125, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6375715732574463, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7509278655052185, + "num_tokens": 696283083.0, + "step": 26909 + }, + { + "epoch": 2.955194377333626, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1105151176452637, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.6984724998474121, + "num_tokens": 696318250.0, + "step": 26910 + }, + { + "epoch": 2.95530419503624, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6166696548461914, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7408784031867981, + "num_tokens": 696338696.0, + "step": 26911 + }, + { + "epoch": 2.9554140127388537, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.757737398147583, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7408261299133301, + "num_tokens": 696359347.0, + "step": 26912 + }, + { + "epoch": 2.955523830441467, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6636874675750732, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.747675895690918, + "num_tokens": 696379979.0, + "step": 26913 + }, + { + "epoch": 2.955633648144081, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2467617988586426, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7146767377853394, + "num_tokens": 696410744.0, + "step": 26914 + }, + { + "epoch": 2.9557434658466946, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1442556381225586, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7227759957313538, + "num_tokens": 696442078.0, + "step": 26915 + }, + { + "epoch": 2.955853283549308, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5560781955718994, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7371214628219604, + "num_tokens": 696464422.0, + "step": 26916 + }, + { + "epoch": 2.955963101251922, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4569907188415527, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7395235896110535, + "num_tokens": 696488774.0, + "step": 26917 + }, + { + "epoch": 2.9560729189545354, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2706105709075928, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7135130167007446, + "num_tokens": 696516657.0, + "step": 26918 + }, + { + "epoch": 2.956182736657149, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.73516583442688, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.730111300945282, + "num_tokens": 696537618.0, + "step": 26919 + }, + { + "epoch": 2.956292554359763, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.337843418121338, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7514464855194092, + "num_tokens": 696562197.0, + "step": 26920 + }, + { + "epoch": 2.956402372062376, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3962597846984863, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7223377823829651, + "num_tokens": 696588380.0, + "step": 26921 + }, + { + "epoch": 2.95651218976499, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4340896606445312, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7271521091461182, + "num_tokens": 696612837.0, + "step": 26922 + }, + { + "epoch": 2.9566220074676037, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5871388912200928, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7330300211906433, + "num_tokens": 696633842.0, + "step": 26923 + }, + { + "epoch": 2.9567318251702175, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2811439037323, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7277883291244507, + "num_tokens": 696660609.0, + "step": 26924 + }, + { + "epoch": 2.9568416428728312, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.305171489715576, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7201946973800659, + "num_tokens": 696688257.0, + "step": 26925 + }, + { + "epoch": 2.9569514605754446, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4997663497924805, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7455198764801025, + "num_tokens": 696712726.0, + "step": 26926 + }, + { + "epoch": 2.9570612782780583, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0309932231903076, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7363072037696838, + "num_tokens": 696743551.0, + "step": 26927 + }, + { + "epoch": 2.957171095980672, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1379857063293457, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.7446324825286865, + "num_tokens": 696772162.0, + "step": 26928 + }, + { + "epoch": 2.957280913683286, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.336239814758301, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7391438484191895, + "num_tokens": 696796890.0, + "step": 26929 + }, + { + "epoch": 2.9573907313858996, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5173792839050293, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7277234792709351, + "num_tokens": 696821425.0, + "step": 26930 + }, + { + "epoch": 2.957500549088513, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.349680185317993, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7287132740020752, + "num_tokens": 696849118.0, + "step": 26931 + }, + { + "epoch": 2.9576103667911267, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3540773391723633, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7166157960891724, + "num_tokens": 696875342.0, + "step": 26932 + }, + { + "epoch": 2.9577201844937404, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4702389240264893, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7266114950180054, + "num_tokens": 696900139.0, + "step": 26933 + }, + { + "epoch": 2.957830002196354, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4398586750030518, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7170466184616089, + "num_tokens": 696925544.0, + "step": 26934 + }, + { + "epoch": 2.957939819898968, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.179623603820801, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7277926802635193, + "num_tokens": 696957200.0, + "step": 26935 + }, + { + "epoch": 2.9580496376015812, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.29776668548584, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7170400619506836, + "num_tokens": 696984092.0, + "step": 26936 + }, + { + "epoch": 2.958159455304195, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5367820262908936, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7435146570205688, + "num_tokens": 697009435.0, + "step": 26937 + }, + { + "epoch": 2.9582692730068088, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1026759147644043, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7368319034576416, + "num_tokens": 697040127.0, + "step": 26938 + }, + { + "epoch": 2.9583790907094225, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.1454927921295166, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7269296646118164, + "num_tokens": 697068990.0, + "step": 26939 + }, + { + "epoch": 2.9584889084120363, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.428295135498047, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7255308628082275, + "num_tokens": 697093253.0, + "step": 26940 + }, + { + "epoch": 2.9585987261146496, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.038651943206787, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7154669761657715, + "num_tokens": 697127762.0, + "step": 26941 + }, + { + "epoch": 2.9587085438172633, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.4167912006378174, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7166359424591064, + "num_tokens": 697152355.0, + "step": 26942 + }, + { + "epoch": 2.958818361519877, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3515663146972656, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7518962025642395, + "num_tokens": 697177570.0, + "step": 26943 + }, + { + "epoch": 2.9589281792224904, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2603487968444824, + "learning_rate": 1e-06, + "loss": 0.7442, + "mean_token_accuracy": 0.7609408497810364, + "num_tokens": 697203614.0, + "step": 26944 + }, + { + "epoch": 2.959037996925104, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.505781412124634, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7452460527420044, + "num_tokens": 697227800.0, + "step": 26945 + }, + { + "epoch": 2.959147814627718, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.652320623397827, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7313563823699951, + "num_tokens": 697249865.0, + "step": 26946 + }, + { + "epoch": 2.9592576323303317, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.581120014190674, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7308362126350403, + "num_tokens": 697272909.0, + "step": 26947 + }, + { + "epoch": 2.9593674500329454, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.209176778793335, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7468964457511902, + "num_tokens": 697300674.0, + "step": 26948 + }, + { + "epoch": 2.9594772677355587, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3587968349456787, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7626640796661377, + "num_tokens": 697327265.0, + "step": 26949 + }, + { + "epoch": 2.9595870854381725, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.334054946899414, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7534348964691162, + "num_tokens": 697352568.0, + "step": 26950 + }, + { + "epoch": 2.9596969031407863, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0684285163879395, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7233105897903442, + "num_tokens": 697383626.0, + "step": 26951 + }, + { + "epoch": 2.9598067208434, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.606839895248413, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7302537560462952, + "num_tokens": 697406230.0, + "step": 26952 + }, + { + "epoch": 2.9599165385460138, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.562192440032959, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7408440113067627, + "num_tokens": 697428583.0, + "step": 26953 + }, + { + "epoch": 2.960026356248627, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.6343188285827637, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7292672991752625, + "num_tokens": 697450490.0, + "step": 26954 + }, + { + "epoch": 2.960136173951241, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.5266385078430176, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7178522348403931, + "num_tokens": 697474413.0, + "step": 26955 + }, + { + "epoch": 2.9602459916538546, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3829503059387207, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7117507457733154, + "num_tokens": 697500328.0, + "step": 26956 + }, + { + "epoch": 2.9603558093564684, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.107950210571289, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7449241876602173, + "num_tokens": 697531539.0, + "step": 26957 + }, + { + "epoch": 2.960465627059082, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2997143268585205, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7115778923034668, + "num_tokens": 697557611.0, + "step": 26958 + }, + { + "epoch": 2.9605754447616954, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.336052894592285, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7253005504608154, + "num_tokens": 697584171.0, + "step": 26959 + }, + { + "epoch": 2.960685262464309, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.3208765983581543, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7306445240974426, + "num_tokens": 697610562.0, + "step": 26960 + }, + { + "epoch": 2.960795080166923, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.617307186126709, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7380269765853882, + "num_tokens": 697632479.0, + "step": 26961 + }, + { + "epoch": 2.9609048978695367, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2746007442474365, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7229524254798889, + "num_tokens": 697659337.0, + "step": 26962 + }, + { + "epoch": 2.9610147155721505, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.0830514430999756, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7327385544776917, + "num_tokens": 697692917.0, + "step": 26963 + }, + { + "epoch": 2.9611245332747638, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.169907569885254, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.718727707862854, + "num_tokens": 697722169.0, + "step": 26964 + }, + { + "epoch": 2.9612343509773775, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.2783544063568115, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7273643016815186, + "num_tokens": 697750091.0, + "step": 26965 + }, + { + "epoch": 2.9613441686799913, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.431047201156616, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7197208404541016, + "num_tokens": 697776446.0, + "step": 26966 + }, + { + "epoch": 2.961453986382605, + "ewc_loss": 0.00023174285888671875, + "grad_norm": 2.7241742610931396, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7264130115509033, + "num_tokens": 697796895.0, + "step": 26967 + }, + { + "epoch": 2.961563804085219, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.21722412109375, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7062916159629822, + "num_tokens": 697827015.0, + "step": 26968 + }, + { + "epoch": 2.961673621787832, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.729032516479492, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7600523233413696, + "num_tokens": 697849556.0, + "step": 26969 + }, + { + "epoch": 2.961783439490446, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.452698230743408, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7486568689346313, + "num_tokens": 697873031.0, + "step": 26970 + }, + { + "epoch": 2.9618932571930596, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.25685715675354, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7239869236946106, + "num_tokens": 697902202.0, + "step": 26971 + }, + { + "epoch": 2.962003074895673, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.259315252304077, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7384531497955322, + "num_tokens": 697929995.0, + "step": 26972 + }, + { + "epoch": 2.9621128925982867, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4090895652770996, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7380561232566833, + "num_tokens": 697954809.0, + "step": 26973 + }, + { + "epoch": 2.9622227103009005, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3773341178894043, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7030594348907471, + "num_tokens": 697982219.0, + "step": 26974 + }, + { + "epoch": 2.962332528003514, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3010525703430176, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.719961404800415, + "num_tokens": 698010780.0, + "step": 26975 + }, + { + "epoch": 2.962442345706128, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2536635398864746, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7396141290664673, + "num_tokens": 698036413.0, + "step": 26976 + }, + { + "epoch": 2.9625521634087413, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.29198956489563, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7428385019302368, + "num_tokens": 698062900.0, + "step": 26977 + }, + { + "epoch": 2.962661981111355, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5621702671051025, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7217333912849426, + "num_tokens": 698087121.0, + "step": 26978 + }, + { + "epoch": 2.962771798813969, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3910579681396484, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7115680575370789, + "num_tokens": 698113106.0, + "step": 26979 + }, + { + "epoch": 2.9628816165165826, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.351923704147339, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7429061532020569, + "num_tokens": 698136953.0, + "step": 26980 + }, + { + "epoch": 2.9629914342191963, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.588242530822754, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7091505527496338, + "num_tokens": 698160688.0, + "step": 26981 + }, + { + "epoch": 2.9631012519218096, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4137110710144043, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7388365268707275, + "num_tokens": 698186106.0, + "step": 26982 + }, + { + "epoch": 2.9632110696244234, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.31673526763916, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7325049638748169, + "num_tokens": 698213086.0, + "step": 26983 + }, + { + "epoch": 2.963320887327037, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.105945110321045, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7448551654815674, + "num_tokens": 698243181.0, + "step": 26984 + }, + { + "epoch": 2.963430705029651, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4808478355407715, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7398829460144043, + "num_tokens": 698267595.0, + "step": 26985 + }, + { + "epoch": 2.9635405227322646, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5283002853393555, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7476622462272644, + "num_tokens": 698291108.0, + "step": 26986 + }, + { + "epoch": 2.963650340434878, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1635587215423584, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7343230843544006, + "num_tokens": 698321436.0, + "step": 26987 + }, + { + "epoch": 2.9637601581374917, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3520383834838867, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7335705161094666, + "num_tokens": 698348173.0, + "step": 26988 + }, + { + "epoch": 2.9638699758401055, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.092973232269287, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7041321992874146, + "num_tokens": 698382618.0, + "step": 26989 + }, + { + "epoch": 2.9639797935427192, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 1.9608027935028076, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7207905650138855, + "num_tokens": 698419430.0, + "step": 26990 + }, + { + "epoch": 2.964089611245333, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.6182711124420166, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7480595111846924, + "num_tokens": 698440289.0, + "step": 26991 + }, + { + "epoch": 2.9641994289479463, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4077792167663574, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7293542623519897, + "num_tokens": 698468329.0, + "step": 26992 + }, + { + "epoch": 2.96430924665056, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2908878326416016, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7335344552993774, + "num_tokens": 698493506.0, + "step": 26993 + }, + { + "epoch": 2.964419064353174, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.715231418609619, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7505321502685547, + "num_tokens": 698512949.0, + "step": 26994 + }, + { + "epoch": 2.964528882055787, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.344266176223755, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.719653308391571, + "num_tokens": 698540485.0, + "step": 26995 + }, + { + "epoch": 2.9646386997584013, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2745296955108643, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7065338492393494, + "num_tokens": 698568590.0, + "step": 26996 + }, + { + "epoch": 2.9647485174610146, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.38356876373291, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7298500537872314, + "num_tokens": 698595857.0, + "step": 26997 + }, + { + "epoch": 2.9648583351636284, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.048276901245117, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.709998607635498, + "num_tokens": 698630604.0, + "step": 26998 + }, + { + "epoch": 2.964968152866242, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4708142280578613, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7457255721092224, + "num_tokens": 698655461.0, + "step": 26999 + }, + { + "epoch": 2.9650779705688555, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.0892889499664307, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7087652087211609, + "num_tokens": 698690583.0, + "step": 27000 + }, + { + "epoch": 2.9651877882714692, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.242577314376831, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.714641273021698, + "num_tokens": 698717539.0, + "step": 27001 + }, + { + "epoch": 2.965297605974083, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.56097674369812, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7471697330474854, + "num_tokens": 698740933.0, + "step": 27002 + }, + { + "epoch": 2.9654074236766967, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.391143560409546, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7334432601928711, + "num_tokens": 698767324.0, + "step": 27003 + }, + { + "epoch": 2.9655172413793105, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.357412815093994, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7233892679214478, + "num_tokens": 698795914.0, + "step": 27004 + }, + { + "epoch": 2.965627059081924, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.511017322540283, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.726686954498291, + "num_tokens": 698821613.0, + "step": 27005 + }, + { + "epoch": 2.9657368767845376, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.681701898574829, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7269080281257629, + "num_tokens": 698842556.0, + "step": 27006 + }, + { + "epoch": 2.9658466944871513, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3446946144104004, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.727599024772644, + "num_tokens": 698871271.0, + "step": 27007 + }, + { + "epoch": 2.965956512189765, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.317656993865967, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7109315395355225, + "num_tokens": 698897929.0, + "step": 27008 + }, + { + "epoch": 2.966066329892379, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.481060743331909, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7126621007919312, + "num_tokens": 698922956.0, + "step": 27009 + }, + { + "epoch": 2.966176147594992, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1256041526794434, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7407723665237427, + "num_tokens": 698952848.0, + "step": 27010 + }, + { + "epoch": 2.966285965297606, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4423224925994873, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7342700958251953, + "num_tokens": 698977153.0, + "step": 27011 + }, + { + "epoch": 2.9663957830002197, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.768793821334839, + "learning_rate": 1e-06, + "loss": 0.7424, + "mean_token_accuracy": 0.7657431960105896, + "num_tokens": 698996246.0, + "step": 27012 + }, + { + "epoch": 2.9665056007028334, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.592435836791992, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7364687919616699, + "num_tokens": 699018202.0, + "step": 27013 + }, + { + "epoch": 2.966615418405447, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.665457248687744, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7310667037963867, + "num_tokens": 699039269.0, + "step": 27014 + }, + { + "epoch": 2.9667252361080605, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.221264123916626, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7208510041236877, + "num_tokens": 699067947.0, + "step": 27015 + }, + { + "epoch": 2.9668350538106742, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3406460285186768, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7263937592506409, + "num_tokens": 699093164.0, + "step": 27016 + }, + { + "epoch": 2.966944871513288, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4556586742401123, + "learning_rate": 1e-06, + "loss": 0.7802, + "mean_token_accuracy": 0.7482768297195435, + "num_tokens": 699116466.0, + "step": 27017 + }, + { + "epoch": 2.9670546892159018, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.529224157333374, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7576732039451599, + "num_tokens": 699138646.0, + "step": 27018 + }, + { + "epoch": 2.9671645069185155, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4248571395874023, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7337853908538818, + "num_tokens": 699163571.0, + "step": 27019 + }, + { + "epoch": 2.967274324621129, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.175457239151001, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7176825404167175, + "num_tokens": 699194222.0, + "step": 27020 + }, + { + "epoch": 2.9673841423237426, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.080561876296997, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7262635231018066, + "num_tokens": 699225044.0, + "step": 27021 + }, + { + "epoch": 2.9674939600263563, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2062394618988037, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7124009132385254, + "num_tokens": 699254793.0, + "step": 27022 + }, + { + "epoch": 2.9676037777289697, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5554776191711426, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7300225496292114, + "num_tokens": 699276903.0, + "step": 27023 + }, + { + "epoch": 2.9677135954315834, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1374075412750244, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.7406680583953857, + "num_tokens": 699306065.0, + "step": 27024 + }, + { + "epoch": 2.967823413134197, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.7880184650421143, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7252610921859741, + "num_tokens": 699329481.0, + "step": 27025 + }, + { + "epoch": 2.967933230836811, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1052212715148926, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7122592329978943, + "num_tokens": 699361293.0, + "step": 27026 + }, + { + "epoch": 2.9680430485394247, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.348752737045288, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7403734922409058, + "num_tokens": 699387539.0, + "step": 27027 + }, + { + "epoch": 2.968152866242038, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.732577323913574, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7426728010177612, + "num_tokens": 699407609.0, + "step": 27028 + }, + { + "epoch": 2.9682626839446518, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2637264728546143, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7219492197036743, + "num_tokens": 699435534.0, + "step": 27029 + }, + { + "epoch": 2.9683725016472655, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4500699043273926, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7270200252532959, + "num_tokens": 699460880.0, + "step": 27030 + }, + { + "epoch": 2.9684823193498793, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.60493803024292, + "learning_rate": 1e-06, + "loss": 0.7673, + "mean_token_accuracy": 0.7560680508613586, + "num_tokens": 699481884.0, + "step": 27031 + }, + { + "epoch": 2.968592137052493, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4309005737304688, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.732195258140564, + "num_tokens": 699506015.0, + "step": 27032 + }, + { + "epoch": 2.9687019547551063, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3671770095825195, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7208991050720215, + "num_tokens": 699532009.0, + "step": 27033 + }, + { + "epoch": 2.96881177245772, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5481178760528564, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7333623766899109, + "num_tokens": 699553030.0, + "step": 27034 + }, + { + "epoch": 2.968921590160334, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2697856426239014, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7397359609603882, + "num_tokens": 699579827.0, + "step": 27035 + }, + { + "epoch": 2.9690314078629476, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2820677757263184, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7279759645462036, + "num_tokens": 699606097.0, + "step": 27036 + }, + { + "epoch": 2.9691412255655614, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.428462505340576, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7438765168190002, + "num_tokens": 699628139.0, + "step": 27037 + }, + { + "epoch": 2.9692510432681747, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2712113857269287, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7248755693435669, + "num_tokens": 699656894.0, + "step": 27038 + }, + { + "epoch": 2.9693608609707884, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5498976707458496, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7232903838157654, + "num_tokens": 699682476.0, + "step": 27039 + }, + { + "epoch": 2.969470678673402, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3805651664733887, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7235574126243591, + "num_tokens": 699708794.0, + "step": 27040 + }, + { + "epoch": 2.969580496376016, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5643086433410645, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7450882792472839, + "num_tokens": 699731205.0, + "step": 27041 + }, + { + "epoch": 2.9696903140786297, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.310455560684204, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7373877763748169, + "num_tokens": 699757450.0, + "step": 27042 + }, + { + "epoch": 2.969800131781243, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.594228982925415, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.7473008632659912, + "num_tokens": 699778921.0, + "step": 27043 + }, + { + "epoch": 2.969909949483857, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.222350597381592, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.730069100856781, + "num_tokens": 699808168.0, + "step": 27044 + }, + { + "epoch": 2.9700197671864705, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2232816219329834, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7271293997764587, + "num_tokens": 699838825.0, + "step": 27045 + }, + { + "epoch": 2.970129584889084, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2216269969940186, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7223663330078125, + "num_tokens": 699866348.0, + "step": 27046 + }, + { + "epoch": 2.970239402591698, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.464101552963257, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7435036897659302, + "num_tokens": 699889755.0, + "step": 27047 + }, + { + "epoch": 2.9703492202943114, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1531295776367188, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7090020179748535, + "num_tokens": 699922208.0, + "step": 27048 + }, + { + "epoch": 2.970459037996925, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.824619770050049, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7255359888076782, + "num_tokens": 699941048.0, + "step": 27049 + }, + { + "epoch": 2.970568855699539, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1627042293548584, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7146466970443726, + "num_tokens": 699969241.0, + "step": 27050 + }, + { + "epoch": 2.970678673402152, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.324739456176758, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7361093163490295, + "num_tokens": 699993630.0, + "step": 27051 + }, + { + "epoch": 2.970788491104766, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.346393585205078, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7369022369384766, + "num_tokens": 700020713.0, + "step": 27052 + }, + { + "epoch": 2.9708983088073797, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.633608818054199, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7372554540634155, + "num_tokens": 700043333.0, + "step": 27053 + }, + { + "epoch": 2.9710081265099935, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.305429220199585, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7205670475959778, + "num_tokens": 700071824.0, + "step": 27054 + }, + { + "epoch": 2.971117944212607, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3332948684692383, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7488257884979248, + "num_tokens": 700096285.0, + "step": 27055 + }, + { + "epoch": 2.9712277619152205, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3602938652038574, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7245977520942688, + "num_tokens": 700122349.0, + "step": 27056 + }, + { + "epoch": 2.9713375796178343, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1139891147613525, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.733230471611023, + "num_tokens": 700152605.0, + "step": 27057 + }, + { + "epoch": 2.971447397320448, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5430827140808105, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7517328858375549, + "num_tokens": 700173611.0, + "step": 27058 + }, + { + "epoch": 2.971557215023062, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.202003240585327, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7216890454292297, + "num_tokens": 700205386.0, + "step": 27059 + }, + { + "epoch": 2.9716670327256756, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.187873363494873, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7343044281005859, + "num_tokens": 700235046.0, + "step": 27060 + }, + { + "epoch": 2.971776850428289, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1894004344940186, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7246645092964172, + "num_tokens": 700265594.0, + "step": 27061 + }, + { + "epoch": 2.9718866681309026, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.975079298019409, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7471848726272583, + "num_tokens": 700284713.0, + "step": 27062 + }, + { + "epoch": 2.9719964858335164, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4687392711639404, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7280882596969604, + "num_tokens": 700311304.0, + "step": 27063 + }, + { + "epoch": 2.97210630353613, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5665323734283447, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7346347570419312, + "num_tokens": 700334528.0, + "step": 27064 + }, + { + "epoch": 2.972216121238744, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.165597438812256, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.718256950378418, + "num_tokens": 700365145.0, + "step": 27065 + }, + { + "epoch": 2.972325938941357, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1689939498901367, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7379094362258911, + "num_tokens": 700394174.0, + "step": 27066 + }, + { + "epoch": 2.972435756643971, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.246676206588745, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7365220785140991, + "num_tokens": 700422187.0, + "step": 27067 + }, + { + "epoch": 2.9725455743465847, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.391928195953369, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.701690673828125, + "num_tokens": 700449559.0, + "step": 27068 + }, + { + "epoch": 2.9726553920491985, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.356884002685547, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7190799713134766, + "num_tokens": 700479833.0, + "step": 27069 + }, + { + "epoch": 2.9727652097518122, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4892797470092773, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7409409880638123, + "num_tokens": 700501218.0, + "step": 27070 + }, + { + "epoch": 2.9728750274544256, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3105287551879883, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7024849653244019, + "num_tokens": 700529618.0, + "step": 27071 + }, + { + "epoch": 2.9729848451570393, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2986390590667725, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.731919527053833, + "num_tokens": 700555811.0, + "step": 27072 + }, + { + "epoch": 2.973094662859653, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.9083728790283203, + "learning_rate": 1e-06, + "loss": 0.7343, + "mean_token_accuracy": 0.7663918733596802, + "num_tokens": 700574821.0, + "step": 27073 + }, + { + "epoch": 2.9732044805622664, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.486868381500244, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7200220227241516, + "num_tokens": 700599791.0, + "step": 27074 + }, + { + "epoch": 2.97331429826488, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.330594778060913, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7350119352340698, + "num_tokens": 700627048.0, + "step": 27075 + }, + { + "epoch": 2.973424115967494, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.155162811279297, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7154280543327332, + "num_tokens": 700656950.0, + "step": 27076 + }, + { + "epoch": 2.9735339336701077, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.270960569381714, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7141916155815125, + "num_tokens": 700686909.0, + "step": 27077 + }, + { + "epoch": 2.9736437513727214, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4519550800323486, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.750823974609375, + "num_tokens": 700711077.0, + "step": 27078 + }, + { + "epoch": 2.9737535690753347, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3661253452301025, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.731569766998291, + "num_tokens": 700737323.0, + "step": 27079 + }, + { + "epoch": 2.9738633867779485, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.48714542388916, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7246513366699219, + "num_tokens": 700763459.0, + "step": 27080 + }, + { + "epoch": 2.9739732044805622, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.328277587890625, + "learning_rate": 1e-06, + "loss": 0.753, + "mean_token_accuracy": 0.761009931564331, + "num_tokens": 700790020.0, + "step": 27081 + }, + { + "epoch": 2.974083022183176, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.7126176357269287, + "learning_rate": 1e-06, + "loss": 0.7486, + "mean_token_accuracy": 0.765755295753479, + "num_tokens": 700810259.0, + "step": 27082 + }, + { + "epoch": 2.9741928398857898, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.286926031112671, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.719334602355957, + "num_tokens": 700837021.0, + "step": 27083 + }, + { + "epoch": 2.974302657588403, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.211000919342041, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7342702150344849, + "num_tokens": 700864837.0, + "step": 27084 + }, + { + "epoch": 2.974412475291017, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.241058349609375, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7315640449523926, + "num_tokens": 700893517.0, + "step": 27085 + }, + { + "epoch": 2.9745222929936306, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.422766923904419, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7184411883354187, + "num_tokens": 700918292.0, + "step": 27086 + }, + { + "epoch": 2.9746321106962443, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.386568546295166, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.73468017578125, + "num_tokens": 700942703.0, + "step": 27087 + }, + { + "epoch": 2.974741928398858, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5846049785614014, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7456841468811035, + "num_tokens": 700966432.0, + "step": 27088 + }, + { + "epoch": 2.9748517461014714, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.2521750926971436, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7296280860900879, + "num_tokens": 700994539.0, + "step": 27089 + }, + { + "epoch": 2.974961563804085, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.0723764896392822, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.718797504901886, + "num_tokens": 701028518.0, + "step": 27090 + }, + { + "epoch": 2.975071381506699, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.315725564956665, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.737250804901123, + "num_tokens": 701053782.0, + "step": 27091 + }, + { + "epoch": 2.9751811992093127, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.454563856124878, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.751815140247345, + "num_tokens": 701077793.0, + "step": 27092 + }, + { + "epoch": 2.9752910169119264, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3628931045532227, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.7075047492980957, + "num_tokens": 701104306.0, + "step": 27093 + }, + { + "epoch": 2.9754008346145397, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3155288696289062, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7402070164680481, + "num_tokens": 701129809.0, + "step": 27094 + }, + { + "epoch": 2.9755106523171535, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.303128480911255, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7405052185058594, + "num_tokens": 701157533.0, + "step": 27095 + }, + { + "epoch": 2.9756204700197673, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.1525330543518066, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7291208505630493, + "num_tokens": 701188408.0, + "step": 27096 + }, + { + "epoch": 2.9757302877223806, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.4801177978515625, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7199164032936096, + "num_tokens": 701212371.0, + "step": 27097 + }, + { + "epoch": 2.9758401054249948, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.40859317779541, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7491124868392944, + "num_tokens": 701235803.0, + "step": 27098 + }, + { + "epoch": 2.975949923127608, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3337771892547607, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7209277153015137, + "num_tokens": 701263089.0, + "step": 27099 + }, + { + "epoch": 2.976059740830222, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3717501163482666, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7203476428985596, + "num_tokens": 701288344.0, + "step": 27100 + }, + { + "epoch": 2.9761695585328356, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.348297357559204, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7253388166427612, + "num_tokens": 701314693.0, + "step": 27101 + }, + { + "epoch": 2.976279376235449, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.759294033050537, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7540286183357239, + "num_tokens": 701332861.0, + "step": 27102 + }, + { + "epoch": 2.9763891939380627, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3134493827819824, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7093181610107422, + "num_tokens": 701362650.0, + "step": 27103 + }, + { + "epoch": 2.9764990116406764, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.412015438079834, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7208459377288818, + "num_tokens": 701387886.0, + "step": 27104 + }, + { + "epoch": 2.97660882934329, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.425506591796875, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7284071445465088, + "num_tokens": 701411826.0, + "step": 27105 + }, + { + "epoch": 2.976718647045904, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.337541341781616, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7391220331192017, + "num_tokens": 701437492.0, + "step": 27106 + }, + { + "epoch": 2.9768284647485173, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.164860248565674, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7275229692459106, + "num_tokens": 701467832.0, + "step": 27107 + }, + { + "epoch": 2.976938282451131, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.3242745399475098, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7267475128173828, + "num_tokens": 701495198.0, + "step": 27108 + }, + { + "epoch": 2.9770481001537448, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.189638376235962, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7213412523269653, + "num_tokens": 701523997.0, + "step": 27109 + }, + { + "epoch": 2.9771579178563585, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.5358314514160156, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7119625806808472, + "num_tokens": 701549938.0, + "step": 27110 + }, + { + "epoch": 2.9772677355589723, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 2.315760850906372, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.725435733795166, + "num_tokens": 701576929.0, + "step": 27111 + }, + { + "epoch": 2.9773775532615856, + "ewc_loss": 0.00023365020751953125, + "grad_norm": 3.001938819885254, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7314139604568481, + "num_tokens": 701594938.0, + "step": 27112 + }, + { + "epoch": 2.9774873709641994, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3411219120025635, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7149350047111511, + "num_tokens": 701620664.0, + "step": 27113 + }, + { + "epoch": 2.977597188666813, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6271135807037354, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7493091225624084, + "num_tokens": 701642530.0, + "step": 27114 + }, + { + "epoch": 2.977707006369427, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3121564388275146, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7249599099159241, + "num_tokens": 701667475.0, + "step": 27115 + }, + { + "epoch": 2.9778168240720406, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.7644846439361572, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7265305519104004, + "num_tokens": 701689185.0, + "step": 27116 + }, + { + "epoch": 2.977926641774654, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2548911571502686, + "learning_rate": 1e-06, + "loss": 0.7657, + "mean_token_accuracy": 0.7524309158325195, + "num_tokens": 701717697.0, + "step": 27117 + }, + { + "epoch": 2.9780364594772677, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.224532127380371, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7454479336738586, + "num_tokens": 701743206.0, + "step": 27118 + }, + { + "epoch": 2.9781462771798815, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3813846111297607, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7363808155059814, + "num_tokens": 701766654.0, + "step": 27119 + }, + { + "epoch": 2.978256094882495, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5677402019500732, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7230808138847351, + "num_tokens": 701789536.0, + "step": 27120 + }, + { + "epoch": 2.978365912585109, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.495713949203491, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.734516978263855, + "num_tokens": 701812744.0, + "step": 27121 + }, + { + "epoch": 2.9784757302877223, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.429549217224121, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7456998229026794, + "num_tokens": 701836365.0, + "step": 27122 + }, + { + "epoch": 2.978585547990336, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.7810444831848145, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7366567850112915, + "num_tokens": 701857860.0, + "step": 27123 + }, + { + "epoch": 2.97869536569295, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3048343658447266, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7158077955245972, + "num_tokens": 701886383.0, + "step": 27124 + }, + { + "epoch": 2.978805183395563, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.656824827194214, + "learning_rate": 1e-06, + "loss": 0.7599, + "mean_token_accuracy": 0.7688596248626709, + "num_tokens": 701907847.0, + "step": 27125 + }, + { + "epoch": 2.978915001098177, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3521697521209717, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7334655523300171, + "num_tokens": 701936129.0, + "step": 27126 + }, + { + "epoch": 2.9790248188007906, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5303356647491455, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7291213274002075, + "num_tokens": 701959243.0, + "step": 27127 + }, + { + "epoch": 2.9791346365034044, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2639734745025635, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7343082427978516, + "num_tokens": 701985609.0, + "step": 27128 + }, + { + "epoch": 2.979244454206018, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4338552951812744, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7227203845977783, + "num_tokens": 702010856.0, + "step": 27129 + }, + { + "epoch": 2.9793542719086314, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4026777744293213, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7186374664306641, + "num_tokens": 702036322.0, + "step": 27130 + }, + { + "epoch": 2.979464089611245, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2525811195373535, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7458943128585815, + "num_tokens": 702064103.0, + "step": 27131 + }, + { + "epoch": 2.979573907313859, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.353990316390991, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7341501712799072, + "num_tokens": 702089732.0, + "step": 27132 + }, + { + "epoch": 2.9796837250164727, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5192439556121826, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.757296621799469, + "num_tokens": 702111620.0, + "step": 27133 + }, + { + "epoch": 2.9797935427190865, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.442582368850708, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7052468061447144, + "num_tokens": 702136971.0, + "step": 27134 + }, + { + "epoch": 2.9799033604217, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.127565383911133, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7272950410842896, + "num_tokens": 702166116.0, + "step": 27135 + }, + { + "epoch": 2.9800131781243135, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 3.254775047302246, + "learning_rate": 1e-06, + "loss": 0.728, + "mean_token_accuracy": 0.761914849281311, + "num_tokens": 702180819.0, + "step": 27136 + }, + { + "epoch": 2.9801229958269273, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.675339698791504, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7399928569793701, + "num_tokens": 702201448.0, + "step": 27137 + }, + { + "epoch": 2.980232813529541, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.447139263153076, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7316926717758179, + "num_tokens": 702225696.0, + "step": 27138 + }, + { + "epoch": 2.980342631232155, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4608652591705322, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7287168502807617, + "num_tokens": 702252138.0, + "step": 27139 + }, + { + "epoch": 2.980452448934768, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5469813346862793, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7428967952728271, + "num_tokens": 702275638.0, + "step": 27140 + }, + { + "epoch": 2.980562266637382, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4966111183166504, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7411932349205017, + "num_tokens": 702298503.0, + "step": 27141 + }, + { + "epoch": 2.9806720843399956, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.590864896774292, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.751229465007782, + "num_tokens": 702321521.0, + "step": 27142 + }, + { + "epoch": 2.9807819020426094, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.523608684539795, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.719505250453949, + "num_tokens": 702343966.0, + "step": 27143 + }, + { + "epoch": 2.980891719745223, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.654466390609741, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7200674414634705, + "num_tokens": 702366429.0, + "step": 27144 + }, + { + "epoch": 2.9810015374478365, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.650477170944214, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7319304943084717, + "num_tokens": 702388325.0, + "step": 27145 + }, + { + "epoch": 2.9811113551504502, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2367196083068848, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7162739634513855, + "num_tokens": 702417213.0, + "step": 27146 + }, + { + "epoch": 2.981221172853064, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.492659568786621, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7220742106437683, + "num_tokens": 702441237.0, + "step": 27147 + }, + { + "epoch": 2.9813309905556777, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0970802307128906, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7070677876472473, + "num_tokens": 702474699.0, + "step": 27148 + }, + { + "epoch": 2.9814408082582915, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2302489280700684, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.725571870803833, + "num_tokens": 702504887.0, + "step": 27149 + }, + { + "epoch": 2.981550625960905, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4800477027893066, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7133511900901794, + "num_tokens": 702528723.0, + "step": 27150 + }, + { + "epoch": 2.9816604436635186, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5197184085845947, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7488405704498291, + "num_tokens": 702551150.0, + "step": 27151 + }, + { + "epoch": 2.9817702613661323, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4884073734283447, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7514697313308716, + "num_tokens": 702574087.0, + "step": 27152 + }, + { + "epoch": 2.9818800790687456, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.688145160675049, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7207872867584229, + "num_tokens": 702601130.0, + "step": 27153 + }, + { + "epoch": 2.9819898967713594, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1700096130371094, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7285536527633667, + "num_tokens": 702633442.0, + "step": 27154 + }, + { + "epoch": 2.982099714473973, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.28564715385437, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7387071847915649, + "num_tokens": 702659621.0, + "step": 27155 + }, + { + "epoch": 2.982209532176587, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1353163719177246, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7343039512634277, + "num_tokens": 702687776.0, + "step": 27156 + }, + { + "epoch": 2.9823193498792007, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5001397132873535, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7348483800888062, + "num_tokens": 702712274.0, + "step": 27157 + }, + { + "epoch": 2.982429167581814, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.241978645324707, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7348783016204834, + "num_tokens": 702740462.0, + "step": 27158 + }, + { + "epoch": 2.9825389852844277, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.440234661102295, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7399426698684692, + "num_tokens": 702764661.0, + "step": 27159 + }, + { + "epoch": 2.9826488029870415, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3969838619232178, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7275030612945557, + "num_tokens": 702792044.0, + "step": 27160 + }, + { + "epoch": 2.9827586206896552, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.572685480117798, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7325184345245361, + "num_tokens": 702813905.0, + "step": 27161 + }, + { + "epoch": 2.982868438392269, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6799118518829346, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.733840823173523, + "num_tokens": 702834398.0, + "step": 27162 + }, + { + "epoch": 2.9829782560948823, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4361233711242676, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7195354700088501, + "num_tokens": 702860805.0, + "step": 27163 + }, + { + "epoch": 2.983088073797496, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4394071102142334, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7338833808898926, + "num_tokens": 702887110.0, + "step": 27164 + }, + { + "epoch": 2.98319789150011, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3487818241119385, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7197016477584839, + "num_tokens": 702914694.0, + "step": 27165 + }, + { + "epoch": 2.9833077092027236, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3554604053497314, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7163108587265015, + "num_tokens": 702941666.0, + "step": 27166 + }, + { + "epoch": 2.9834175269053373, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.8228719234466553, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7408658862113953, + "num_tokens": 702962865.0, + "step": 27167 + }, + { + "epoch": 2.9835273446079507, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1558401584625244, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7433447241783142, + "num_tokens": 702991618.0, + "step": 27168 + }, + { + "epoch": 2.9836371623105644, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1803441047668457, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7106243371963501, + "num_tokens": 703023298.0, + "step": 27169 + }, + { + "epoch": 2.983746980013178, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.163163900375366, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7213938236236572, + "num_tokens": 703053813.0, + "step": 27170 + }, + { + "epoch": 2.983856797715792, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.9710705280303955, + "learning_rate": 1e-06, + "loss": 0.7814, + "mean_token_accuracy": 0.7509520053863525, + "num_tokens": 703072059.0, + "step": 27171 + }, + { + "epoch": 2.9839666154184057, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5148751735687256, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.720215916633606, + "num_tokens": 703096897.0, + "step": 27172 + }, + { + "epoch": 2.984076433121019, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.690153121948242, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7267839908599854, + "num_tokens": 703117645.0, + "step": 27173 + }, + { + "epoch": 2.9841862508236328, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3520212173461914, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.7585298418998718, + "num_tokens": 703139908.0, + "step": 27174 + }, + { + "epoch": 2.9842960685262465, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.356541395187378, + "learning_rate": 1e-06, + "loss": 0.7901, + "mean_token_accuracy": 0.7521240711212158, + "num_tokens": 703163662.0, + "step": 27175 + }, + { + "epoch": 2.98440588622886, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.344367265701294, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.742050290107727, + "num_tokens": 703188222.0, + "step": 27176 + }, + { + "epoch": 2.984515703931474, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.344731569290161, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7307976484298706, + "num_tokens": 703213448.0, + "step": 27177 + }, + { + "epoch": 2.9846255216340873, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.306018590927124, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7362448573112488, + "num_tokens": 703239870.0, + "step": 27178 + }, + { + "epoch": 2.984735339336701, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5182008743286133, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7311047911643982, + "num_tokens": 703262075.0, + "step": 27179 + }, + { + "epoch": 2.984845157039315, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.483823776245117, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.729556679725647, + "num_tokens": 703286704.0, + "step": 27180 + }, + { + "epoch": 2.984954974741928, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.298159599304199, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7097135782241821, + "num_tokens": 703315082.0, + "step": 27181 + }, + { + "epoch": 2.985064792444542, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3549253940582275, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7426234483718872, + "num_tokens": 703342324.0, + "step": 27182 + }, + { + "epoch": 2.9851746101471557, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2445220947265625, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7276990413665771, + "num_tokens": 703373898.0, + "step": 27183 + }, + { + "epoch": 2.9852844278497694, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6557066440582275, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7448916435241699, + "num_tokens": 703394399.0, + "step": 27184 + }, + { + "epoch": 2.985394245552383, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5827369689941406, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7242764234542847, + "num_tokens": 703418105.0, + "step": 27185 + }, + { + "epoch": 2.9855040632549965, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.306251287460327, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7211238145828247, + "num_tokens": 703445569.0, + "step": 27186 + }, + { + "epoch": 2.9856138809576103, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.68717622756958, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7243342995643616, + "num_tokens": 703466434.0, + "step": 27187 + }, + { + "epoch": 2.985723698660224, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.353639841079712, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7343478202819824, + "num_tokens": 703493661.0, + "step": 27188 + }, + { + "epoch": 2.985833516362838, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.7137489318847656, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7365273237228394, + "num_tokens": 703515129.0, + "step": 27189 + }, + { + "epoch": 2.9859433340654515, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.9597716331481934, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7459316849708557, + "num_tokens": 703531781.0, + "step": 27190 + }, + { + "epoch": 2.986053151768065, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.087867498397827, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7137275338172913, + "num_tokens": 703564640.0, + "step": 27191 + }, + { + "epoch": 2.9861629694706786, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.293811321258545, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7301857471466064, + "num_tokens": 703590692.0, + "step": 27192 + }, + { + "epoch": 2.9862727871732924, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.481104850769043, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7268346548080444, + "num_tokens": 703614104.0, + "step": 27193 + }, + { + "epoch": 2.986382604875906, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.203387498855591, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7351185083389282, + "num_tokens": 703643107.0, + "step": 27194 + }, + { + "epoch": 2.98649242257852, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1413471698760986, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.707996666431427, + "num_tokens": 703675675.0, + "step": 27195 + }, + { + "epoch": 2.986602240281133, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.17460298538208, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7241305112838745, + "num_tokens": 703705056.0, + "step": 27196 + }, + { + "epoch": 2.986712057983747, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.39945387840271, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7560567855834961, + "num_tokens": 703728771.0, + "step": 27197 + }, + { + "epoch": 2.9868218756863607, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1941261291503906, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7341934442520142, + "num_tokens": 703757949.0, + "step": 27198 + }, + { + "epoch": 2.9869316933889745, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5324110984802246, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7385820150375366, + "num_tokens": 703781668.0, + "step": 27199 + }, + { + "epoch": 2.987041511091588, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2997522354125977, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7398136258125305, + "num_tokens": 703810232.0, + "step": 27200 + }, + { + "epoch": 2.9871513287942015, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1696693897247314, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7342251539230347, + "num_tokens": 703837505.0, + "step": 27201 + }, + { + "epoch": 2.9872611464968153, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.318864583969116, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7238088846206665, + "num_tokens": 703864601.0, + "step": 27202 + }, + { + "epoch": 2.987370964199429, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3191733360290527, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6956491470336914, + "num_tokens": 703896134.0, + "step": 27203 + }, + { + "epoch": 2.9874807819020424, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6308670043945312, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7422647476196289, + "num_tokens": 703916331.0, + "step": 27204 + }, + { + "epoch": 2.987590599604656, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.268603801727295, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.725817084312439, + "num_tokens": 703944317.0, + "step": 27205 + }, + { + "epoch": 2.98770041730727, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.600679636001587, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7343071699142456, + "num_tokens": 703965788.0, + "step": 27206 + }, + { + "epoch": 2.9878102350098836, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2964930534362793, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7318776845932007, + "num_tokens": 703991689.0, + "step": 27207 + }, + { + "epoch": 2.9879200527124974, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3217613697052, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7387394309043884, + "num_tokens": 704018020.0, + "step": 27208 + }, + { + "epoch": 2.9880298704151107, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.545879602432251, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7283989191055298, + "num_tokens": 704040366.0, + "step": 27209 + }, + { + "epoch": 2.9881396881177245, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4041388034820557, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7289832830429077, + "num_tokens": 704064833.0, + "step": 27210 + }, + { + "epoch": 2.988249505820338, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6514132022857666, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7417792677879333, + "num_tokens": 704087321.0, + "step": 27211 + }, + { + "epoch": 2.988359323522952, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2907795906066895, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7060613036155701, + "num_tokens": 704115895.0, + "step": 27212 + }, + { + "epoch": 2.9884691412255657, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0552446842193604, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7364262938499451, + "num_tokens": 704148074.0, + "step": 27213 + }, + { + "epoch": 2.988578958928179, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.350851535797119, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7221231460571289, + "num_tokens": 704175384.0, + "step": 27214 + }, + { + "epoch": 2.988688776630793, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.415670156478882, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7431633472442627, + "num_tokens": 704198495.0, + "step": 27215 + }, + { + "epoch": 2.9887985943334066, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2068052291870117, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7155207395553589, + "num_tokens": 704226538.0, + "step": 27216 + }, + { + "epoch": 2.9889084120360203, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.648974657058716, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7503530979156494, + "num_tokens": 704248083.0, + "step": 27217 + }, + { + "epoch": 2.989018229738634, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.632986307144165, + "learning_rate": 1e-06, + "loss": 0.7571, + "mean_token_accuracy": 0.7548344135284424, + "num_tokens": 704267508.0, + "step": 27218 + }, + { + "epoch": 2.9891280474412474, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.205070972442627, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7093679904937744, + "num_tokens": 704298993.0, + "step": 27219 + }, + { + "epoch": 2.989237865143861, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3803036212921143, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.741027295589447, + "num_tokens": 704323108.0, + "step": 27220 + }, + { + "epoch": 2.989347682846475, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.7618067264556885, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7336888313293457, + "num_tokens": 704343373.0, + "step": 27221 + }, + { + "epoch": 2.9894575005490887, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5138888359069824, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7204683423042297, + "num_tokens": 704366538.0, + "step": 27222 + }, + { + "epoch": 2.9895673182517024, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6783761978149414, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.723605215549469, + "num_tokens": 704390386.0, + "step": 27223 + }, + { + "epoch": 2.9896771359543157, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4062178134918213, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7177094221115112, + "num_tokens": 704415032.0, + "step": 27224 + }, + { + "epoch": 2.9897869536569295, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.294487237930298, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7475733757019043, + "num_tokens": 704442716.0, + "step": 27225 + }, + { + "epoch": 2.9898967713595432, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.370877981185913, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7423237562179565, + "num_tokens": 704467676.0, + "step": 27226 + }, + { + "epoch": 2.9900065890621566, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5919477939605713, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7394229769706726, + "num_tokens": 704488473.0, + "step": 27227 + }, + { + "epoch": 2.9901164067647708, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 1.9913724660873413, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7263615131378174, + "num_tokens": 704520933.0, + "step": 27228 + }, + { + "epoch": 2.990226224467384, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1539463996887207, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7385425567626953, + "num_tokens": 704550513.0, + "step": 27229 + }, + { + "epoch": 2.990336042169998, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.346728801727295, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7112977504730225, + "num_tokens": 704577609.0, + "step": 27230 + }, + { + "epoch": 2.9904458598726116, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3454599380493164, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7328469753265381, + "num_tokens": 704601933.0, + "step": 27231 + }, + { + "epoch": 2.990555677575225, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5388951301574707, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7308099269866943, + "num_tokens": 704624988.0, + "step": 27232 + }, + { + "epoch": 2.9906654952778386, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.248716115951538, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7308821678161621, + "num_tokens": 704653433.0, + "step": 27233 + }, + { + "epoch": 2.9907753129804524, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.140138626098633, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7360134124755859, + "num_tokens": 704681291.0, + "step": 27234 + }, + { + "epoch": 2.990885130683066, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5669655799865723, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7442759275436401, + "num_tokens": 704704250.0, + "step": 27235 + }, + { + "epoch": 2.99099494838568, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4173479080200195, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7489955425262451, + "num_tokens": 704729508.0, + "step": 27236 + }, + { + "epoch": 2.9911047660882932, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5125467777252197, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7353788018226624, + "num_tokens": 704753815.0, + "step": 27237 + }, + { + "epoch": 2.991214583790907, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.249537229537964, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7064703106880188, + "num_tokens": 704785221.0, + "step": 27238 + }, + { + "epoch": 2.9913244014935207, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.154984951019287, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7235013246536255, + "num_tokens": 704815571.0, + "step": 27239 + }, + { + "epoch": 2.9914342191961345, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.371572732925415, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7288860082626343, + "num_tokens": 704841008.0, + "step": 27240 + }, + { + "epoch": 2.9915440368987483, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3071694374084473, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7333099246025085, + "num_tokens": 704867545.0, + "step": 27241 + }, + { + "epoch": 2.9916538546013616, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6223812103271484, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7329913973808289, + "num_tokens": 704891252.0, + "step": 27242 + }, + { + "epoch": 2.9917636723039753, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.635077953338623, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7438066005706787, + "num_tokens": 704910961.0, + "step": 27243 + }, + { + "epoch": 2.991873490006589, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0594635009765625, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.707838237285614, + "num_tokens": 704943947.0, + "step": 27244 + }, + { + "epoch": 2.991983307709203, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2523624897003174, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.727357804775238, + "num_tokens": 704974192.0, + "step": 27245 + }, + { + "epoch": 2.9920931254118166, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.584923505783081, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7440176010131836, + "num_tokens": 704995669.0, + "step": 27246 + }, + { + "epoch": 2.99220294311443, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.611279249191284, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7370320558547974, + "num_tokens": 705020032.0, + "step": 27247 + }, + { + "epoch": 2.9923127608170437, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.260627508163452, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7115896344184875, + "num_tokens": 705048849.0, + "step": 27248 + }, + { + "epoch": 2.9924225785196574, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.319781541824341, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.735370397567749, + "num_tokens": 705077489.0, + "step": 27249 + }, + { + "epoch": 2.992532396222271, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1977007389068604, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7403323650360107, + "num_tokens": 705105991.0, + "step": 27250 + }, + { + "epoch": 2.992642213924885, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.677990674972534, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7371031045913696, + "num_tokens": 705126974.0, + "step": 27251 + }, + { + "epoch": 2.9927520316274983, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3210694789886475, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.70708829164505, + "num_tokens": 705151444.0, + "step": 27252 + }, + { + "epoch": 2.992861849330112, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6370716094970703, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7441680431365967, + "num_tokens": 705172830.0, + "step": 27253 + }, + { + "epoch": 2.9929716670327258, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2608835697174072, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7274342775344849, + "num_tokens": 705202731.0, + "step": 27254 + }, + { + "epoch": 2.993081484735339, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6963047981262207, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7344745993614197, + "num_tokens": 705223406.0, + "step": 27255 + }, + { + "epoch": 2.993191302437953, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.220179557800293, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.716878354549408, + "num_tokens": 705252641.0, + "step": 27256 + }, + { + "epoch": 2.9933011201405666, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0743677616119385, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7246931195259094, + "num_tokens": 705283871.0, + "step": 27257 + }, + { + "epoch": 2.9934109378431804, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3471336364746094, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7362993955612183, + "num_tokens": 705308264.0, + "step": 27258 + }, + { + "epoch": 2.993520755545794, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.515455484390259, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7450215816497803, + "num_tokens": 705331376.0, + "step": 27259 + }, + { + "epoch": 2.9936305732484074, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3339529037475586, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7014758586883545, + "num_tokens": 705359289.0, + "step": 27260 + }, + { + "epoch": 2.993740390951021, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2989532947540283, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7184673547744751, + "num_tokens": 705386707.0, + "step": 27261 + }, + { + "epoch": 2.993850208653635, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0986435413360596, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7390549778938293, + "num_tokens": 705418513.0, + "step": 27262 + }, + { + "epoch": 2.9939600263562487, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0818116664886475, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7454732656478882, + "num_tokens": 705447497.0, + "step": 27263 + }, + { + "epoch": 2.9940698440588625, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.24424147605896, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7299318313598633, + "num_tokens": 705474844.0, + "step": 27264 + }, + { + "epoch": 2.9941796617614758, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.562593698501587, + "learning_rate": 1e-06, + "loss": 0.7725, + "mean_token_accuracy": 0.7559998035430908, + "num_tokens": 705496732.0, + "step": 27265 + }, + { + "epoch": 2.9942894794640895, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.46478271484375, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.736161470413208, + "num_tokens": 705519700.0, + "step": 27266 + }, + { + "epoch": 2.9943992971667033, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2877209186553955, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7415193319320679, + "num_tokens": 705546443.0, + "step": 27267 + }, + { + "epoch": 2.994509114869317, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4812185764312744, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7389996647834778, + "num_tokens": 705569904.0, + "step": 27268 + }, + { + "epoch": 2.994618932571931, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6080191135406494, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7370220422744751, + "num_tokens": 705593737.0, + "step": 27269 + }, + { + "epoch": 2.994728750274544, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3417282104492188, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7326242923736572, + "num_tokens": 705620490.0, + "step": 27270 + }, + { + "epoch": 2.994838567977158, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.639936685562134, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7463583946228027, + "num_tokens": 705641312.0, + "step": 27271 + }, + { + "epoch": 2.9949483856797716, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.372708559036255, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7411165833473206, + "num_tokens": 705668248.0, + "step": 27272 + }, + { + "epoch": 2.9950582033823854, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.245626926422119, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7228750586509705, + "num_tokens": 705696773.0, + "step": 27273 + }, + { + "epoch": 2.995168021084999, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1866109371185303, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7193717956542969, + "num_tokens": 705728765.0, + "step": 27274 + }, + { + "epoch": 2.9952778387876124, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4198999404907227, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7117833495140076, + "num_tokens": 705755618.0, + "step": 27275 + }, + { + "epoch": 2.995387656490226, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2815544605255127, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7266693115234375, + "num_tokens": 705784998.0, + "step": 27276 + }, + { + "epoch": 2.99549747419284, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.499591588973999, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7627360224723816, + "num_tokens": 705807142.0, + "step": 27277 + }, + { + "epoch": 2.9956072918954533, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3873562812805176, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7270511388778687, + "num_tokens": 705834156.0, + "step": 27278 + }, + { + "epoch": 2.9957171095980675, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3811092376708984, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7286736965179443, + "num_tokens": 705858564.0, + "step": 27279 + }, + { + "epoch": 2.995826927300681, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.237511157989502, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7181172370910645, + "num_tokens": 705888824.0, + "step": 27280 + }, + { + "epoch": 2.9959367450032945, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5034878253936768, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7340573668479919, + "num_tokens": 705912030.0, + "step": 27281 + }, + { + "epoch": 2.9960465627059083, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4045586585998535, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7256103157997131, + "num_tokens": 705937277.0, + "step": 27282 + }, + { + "epoch": 2.9961563804085216, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.669541358947754, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.74615079164505, + "num_tokens": 705959095.0, + "step": 27283 + }, + { + "epoch": 2.9962661981111354, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6806039810180664, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7450812458992004, + "num_tokens": 705980298.0, + "step": 27284 + }, + { + "epoch": 2.996376015813749, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4161620140075684, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7575794458389282, + "num_tokens": 706002573.0, + "step": 27285 + }, + { + "epoch": 2.996485833516363, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.564610481262207, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7261509895324707, + "num_tokens": 706024834.0, + "step": 27286 + }, + { + "epoch": 2.9965956512189766, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.311908006668091, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.714268684387207, + "num_tokens": 706050593.0, + "step": 27287 + }, + { + "epoch": 2.99670546892159, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.297933340072632, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7145659923553467, + "num_tokens": 706076878.0, + "step": 27288 + }, + { + "epoch": 2.9968152866242037, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.607137680053711, + "learning_rate": 1e-06, + "loss": 0.7584, + "mean_token_accuracy": 0.758421778678894, + "num_tokens": 706098592.0, + "step": 27289 + }, + { + "epoch": 2.9969251043268175, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4167630672454834, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7534265518188477, + "num_tokens": 706122101.0, + "step": 27290 + }, + { + "epoch": 2.9970349220294312, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3336000442504883, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7315511703491211, + "num_tokens": 706150293.0, + "step": 27291 + }, + { + "epoch": 2.997144739732045, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.6234447956085205, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7551865577697754, + "num_tokens": 706172552.0, + "step": 27292 + }, + { + "epoch": 2.9972545574346583, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.206149101257324, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7257237434387207, + "num_tokens": 706202391.0, + "step": 27293 + }, + { + "epoch": 2.997364375137272, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4794681072235107, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7187203168869019, + "num_tokens": 706226644.0, + "step": 27294 + }, + { + "epoch": 2.997474192839886, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.730055332183838, + "learning_rate": 1e-06, + "loss": 0.799, + "mean_token_accuracy": 0.7468884587287903, + "num_tokens": 706245912.0, + "step": 27295 + }, + { + "epoch": 2.9975840105424996, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.7061820030212402, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7333208322525024, + "num_tokens": 706268033.0, + "step": 27296 + }, + { + "epoch": 2.9976938282451133, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.47625732421875, + "learning_rate": 1e-06, + "loss": 0.7589, + "mean_token_accuracy": 0.7548093795776367, + "num_tokens": 706290252.0, + "step": 27297 + }, + { + "epoch": 2.9978036459477266, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1638031005859375, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7079290151596069, + "num_tokens": 706320660.0, + "step": 27298 + }, + { + "epoch": 2.9979134636503404, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.609884023666382, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7249447107315063, + "num_tokens": 706342544.0, + "step": 27299 + }, + { + "epoch": 2.998023281352954, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.271122694015503, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7425130009651184, + "num_tokens": 706369294.0, + "step": 27300 + }, + { + "epoch": 2.998133099055568, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.519312858581543, + "learning_rate": 1e-06, + "loss": 0.7302, + "mean_token_accuracy": 0.7695307731628418, + "num_tokens": 706392264.0, + "step": 27301 + }, + { + "epoch": 2.9982429167581817, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.151280641555786, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7234886884689331, + "num_tokens": 706423052.0, + "step": 27302 + }, + { + "epoch": 2.998352734460795, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2654638290405273, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7545680999755859, + "num_tokens": 706448674.0, + "step": 27303 + }, + { + "epoch": 2.9984625521634087, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3583180904388428, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7359554767608643, + "num_tokens": 706475897.0, + "step": 27304 + }, + { + "epoch": 2.9985723698660225, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 1.9851692914962769, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.743537425994873, + "num_tokens": 706508259.0, + "step": 27305 + }, + { + "epoch": 2.998682187568636, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.4704196453094482, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7314842939376831, + "num_tokens": 706531279.0, + "step": 27306 + }, + { + "epoch": 2.9987920052712496, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3288791179656982, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7199000120162964, + "num_tokens": 706557018.0, + "step": 27307 + }, + { + "epoch": 2.9989018229738633, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2373030185699463, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.724303126335144, + "num_tokens": 706587542.0, + "step": 27308 + }, + { + "epoch": 2.999011640676477, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.867457151412964, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7346541881561279, + "num_tokens": 706606653.0, + "step": 27309 + }, + { + "epoch": 2.999121458379091, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.386878490447998, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7428252696990967, + "num_tokens": 706631700.0, + "step": 27310 + }, + { + "epoch": 2.999231276081704, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.3736324310302734, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7386316061019897, + "num_tokens": 706656969.0, + "step": 27311 + }, + { + "epoch": 2.999341093784318, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.5183115005493164, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7302361130714417, + "num_tokens": 706681011.0, + "step": 27312 + }, + { + "epoch": 2.9994509114869317, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.2395992279052734, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7350962162017822, + "num_tokens": 706707342.0, + "step": 27313 + }, + { + "epoch": 2.9995607291895454, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.37204647064209, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7367485761642456, + "num_tokens": 706731699.0, + "step": 27314 + }, + { + "epoch": 2.999670546892159, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.420789957046509, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7306182384490967, + "num_tokens": 706757635.0, + "step": 27315 + }, + { + "epoch": 2.9997803645947725, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.160285234451294, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7461460828781128, + "num_tokens": 706787813.0, + "step": 27316 + }, + { + "epoch": 2.9998901822973862, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.1579430103302, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7021872401237488, + "num_tokens": 706821902.0, + "step": 27317 + }, + { + "epoch": 3.0, + "ewc_loss": 0.0002346038818359375, + "grad_norm": 2.0834856033325195, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7358680963516235, + "num_tokens": 706854136.0, + "step": 27318 + }, + { + "epoch": 3.0, + "ewc_loss": 0.0002346038818359375, + "step": 27318, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.9308701857364808, + "train_runtime": 37091.194, + "train_samples_per_second": 11.784, + "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..a4709de --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf9659e73bc84b80b48118215225e31edb99f6d9891d158b2b31ab3e95d47722 +size 13393